what okay now i'm going to share my screen and first one share and um beta i have your art id now and uh gene i have yours as well and megan i'm going to put the textbooks for professor powers in the chat okay okay so um again we're going to be doing four powerpoints um topics are going to include intrinsic tissue and intrinsic scan parameters that's going to be the first one then we're going to hit extrinsic scan parameters and pulse sequences and artifacts so four of them gonna mix in mri math i initially was just going to do a section just on math but i decided to mix it all in as we went along so uh without further ado all right so just to give you a little bit of background on um how we're going to do this like i said two sessions um one and two we'll take a break three and four and we'll finish up please ask questions if you've got them i'll answer them as posted or if i'm too busy chirping away uh and i haven't answered your question please unmute your mic and give me a shout okay um there's a 20 question practice test that's being offered uh you can get it by a separate link again i thank you for your time and attention and um i got asked multiple times the last time i do not have printable versions of this um course that are available just so you know so i do not have printable versions all right so a little bit about myself i've got 40 years of radiology experience 30 of those happen to be an mri the first 10 were in cat scan and special procedures i've got 14 years teaching experience at a local community college near my home five years as an application specialist with ge healthcare i'm the one of the instructors for pulse radiology and i also lecture for a um a private ceu company called advanced health they are out of houston texas as neil said i'm the author of two separate books mri tech to tech explanations and register review tech detect questions and answers if you are interested in those they are both available through either bonds and noble or amazon or directly through wiley blackwell publishing out of london england so without further ado um we'll keep on going so intrinsic scan parameters or intrinsic tissue characteristics are the t1 t2 relaxation properties as well as proton density and i'll go into those a little bit then in the same power point intrinsic scan parameters so don't let intrinsic confuse you so you've got intrinsic tissue parameters which means it's what the tissues do mris make noise and babies cry it's what they do all right so those are characteristics the intrinsic scan parameters trte echo train ti and flip angle after i get done doing those i'll go into what i call my um image contrast and weighting triangles something i kind of developed during my time at ge i kind of figured this out then we'll do the geometric properties the geometric parameters i'll do hit most of the major pulse sequences i'm not going to spend a lot of time doing spectroscopy or dynamic susceptibility of contrast stuff because i highly doubt the registry is going to be asking you those and then we're going to finish up just when you guys are ready to hear me be quiet i'm going to talk about artifacts all right so right off the bat we've got inherent tissue parameters uh the first one is t1 contrast that's what i'm going to get into so every tissue has its own t1 t2 relaxation time and a unique proton density now proton density is what everybody uses nowadays that's the modern term i guess you'd say but however when i was a kid learning mr a while back it was called spin density started to become proton density that time but it's also known as hydrogen density so any of those three terms are equally applicable okay so if they're asking you a proton density question they might kind of shroud that by calling it well hydrogen density is such and such you have to know the different terms just like t1 is called spin lattice and t2 is called spin spin you have to know those terms reversibly you have to know them because they love to use these old like terms for the same thing okay so t1 time of a tissue is how long it takes for 63 of the net magnetization vector to recover you've got to know that definition t1 63 is gone back to align back with b0 you've got to know that i i'll bet you a cup of coffee they're going to ask you that you're going to see something about that and i'll bet you another cup of coffee they're probably going to throw in the term spin lattice at you somewhere t1 spin lattice spin lattice t1 they're the same tomatoes tomatoes so in mri this is how i like to teach it in mri we knock them down meaning we knock down the net magnetization vector from being aligned with b0 we knock it into the transverse plane with an rf pulse so knock them down and then what are they going to do they're not staying in the transverse plane they want to go back to where they just were that is standing up so we knock them down and let them stand up they t1 so in the top right hand corner you're gonna see this little movie and if you watch it real quick there an rf pulse is applied they get knocked down into the transverse plane and over time that pink arrow is growing bigger and bigger the vector is getting larger and larger returning to the b zero okay and that's the t1 relaxation okay so watch it real quick you hit it with the rf refocus it don't worry about that right now and at t1s it's going into that large red circle t1 going on right there that's the that's the 63 of the net magnetization vector that's t1 relaxation know that there are multiple tissues that we're going to be imaging but for the most part we're going to be imaging fat and water because an mri we follow the water you've heard the the expression follow the money in mri we're following the water the body's reaction to any kind of insult be it infection tumor surgery radiation therapy is to force water to that area so we're going to be imaging the water know that water has a long t1 and a long t2 as opposed to fat which has a short t1 and t2 so fat gadden proteins are fast to t1 and water and edema are slow to t1 now we've got t2 relaxation okay that is the time for the net magnetization vector to decay it's a decay process it's going to lower itself to 37 of its original value so 37 guys so if you notice you go wait a minute t1 is 63 and t2 is 37 how convenient that adds up to 100 so if you know if you memorize one the other is true so t1 grows to 63 t2 decays to 37. memorize one and the opposite is true so a tissue is t2 time just so you know is always shorter than it's t1 time always okay so again we knock them down with an rf we let them deface they spread out that's what happens that big large red circle after we hit them with an rf they decay it spreads out in the transverse plane and then eventually goes back to b0 so t1 and t2 relaxation happen simultaneously but they are independent of each other okay got some chats here okay that's just neil that's just neil giving me kudos thank you neil okay so remember t1 and t2 simultaneous t1 and t2 independent of each other t2 another way of describing it it's loss of phase coherence okay so when they're in the transverse plane and they're all together when they start to spread out in the transverse plane that is loss of phase coherence they're in phase here all together and then they get out of phase and just so you know these are chopsticks spray painted yellow for fat and blue for water i'm probably going to be using those again at some point in time okay so t1 and t2 curves you guys are probably familiar with these and that's excellent i want you to be familiar with these um let's make that go away that's good all right so t1 curve on the left depicts the amount of relaxation occurring from two different tissues so t1 is a regrowth process so the the vectors start at zero over on the bottom left and over time they everybody will t1 relax some faster some slower so in that particular t1 relaxation curve white matter depicted by red is faster than csf depicted by well i don't know if that's gray or blue but however you've got t1 contrast t1 relaxation differences over time the csf is going to catch up with the water and they'll have equal amounts of relaxation now t2 is on the right hand side that's that depicts the protons leaving the transverse plane and going back to b0 the distance in between and here's where a lot of books and and articles don't tell you that but the distance in between those two lines that's contrast so i'm gonna i'm gonna show you a little bit better example of that so the distance in between those lines is contrast as those lines get closer together your contrast is decreasing okay so here's the t1 curve and it's just what i what i just said at number one that i have circled in red that's good t1 contrast your tr is short enough that you can see that t1 contrast if you use a tr that's too long and you start getting over towards the number two that circled what kind of contrast have you got not much because they both equally relax the same amount also remember that tissues need to t1 relax in order to give you signal if you start with a tr that's too short then nothing's relaxed and your signal is grainy and it's low signal to noise so position number two like i said is little to no t1 contrast because down the bottom of that graph you can see the tr is getting longer and longer and what happens is you lose t1 contrast same thing here this is a t2 curve one thing you might have noticed is that a t1 curve is basically the same thing as a t2 curve just flipped upside down so tissues start to t2 relax and at some point you've got good contrast between the tissues one's bright and one's dark and that's what contrast is ladies and gentlemen one's bright one's dark you have contrast if you let your te get too long over by position number two then probably the only thing that's giving you any kind of signal is water and you've got no contrast everything else is dark very very dark okay and now we're going to hit proton density so proton density is not any kind of process it's a tissue characteristic they love to say what what is proton density and they'll give you the options it's a tissue characteristic some tissues have a lot of protons in them like water and some tissues don't have as many protons in them fat so if you were to get signal from one tissue versus another you've got a difference in proton density you've got a difference in signal intensity okay so the arrow in that top right image shows csf is bright in the ventricles because it has a high proton density and the periventricular white matter is not as bright it's still got some signal to it but not as much because the white matter has a lesser proton density so you're going to have to identify curves i'll again i'll bet you're a cup of coffee you're going to have to identify a curve they'll put up a graph and they'll say what kind of curve is this and hopefully you're going to say those lines are going up so that's a regrowth process that's a t1 weighted or that's a t1 relaxation curve excellent they'll show you a t2 and they'll say what kind of curve is this and you say um that's relaxation because the lines are down going just so you know there's no such thing as a proton density curve if there is i don't know about it so when they give you your your um choices or your your possible answers one or two are probably pretty close two of them are not so whittle it down and say all right i'm not sure about this one but i know this one's wrong and that one's wrong and now if you have to guess it's a 50 50. okay so now we're done with uh intrinsic tissue parameters now we're going to jump forward to intrinsic scan parameters those intrinsic scan parameters the biggies are trt flip angle echo train and ti so remember the old rule in mri and it's been around since i started 30 years ago tr controls t1 te controls t2 and we're going to cover we're going to cover mostly tr and te right now and then we'll jump forward to flip angle and echo train so what's the definition of tr again if it's in red you're i'm again i'll i'll bet your lunch this time that you're going to see what the definition of tr is so it's the time in milliseconds from the first 90 to the next 90 and here's the key words here that excite the same slice so the way i teach it is i like to describe tr as a loop around time you hit the first slice and then you're gonna go on to two three four five and out to twenty and then you're gonna come back to t to the first slice again that's one tr period whether it's 500 milliseconds or 5 000 milliseconds it's a loop around time so it's how long before the scanner repeats itself over and over again because that's what that's what a pulse sequence is it's a very rigidly timed sequence of events and you just keep on coming back to slice one back to slice one again so if you hear a tr question you see a tr question it's often aimed at a t1 answer okay so long trs versus short trs so a long tr is going to mean that both long and short relaxing tissue for example csf or edema it's been knocked down and you let it stand up and it will be able to give you signal at t e so for example in a proton density long tr on a short te they're both going to be able to give you signal so in mri our signals have mixed we we can see signal from any one of those tissues depending on whether we use a short tr or a long tr or a short te so for example short tr is going to allow only short relaxing tissues to be able to stand up so remember tissue needs t1 relax to be ready to go for the next tr so a short tr means that short relaxing tissues have been knocked down and they've been able to stand up and can contribute signal at t e there they will be bright there's more of those short t1 relaxing tissues in the transverse plane at t e than the long ones so examples of short t1 tissues are already i've already given you fat gadon protein those are the three biggies fgp fat gadden protein long t1 relaxing tissues water aka edema csf and actually urine because it's those are both mostly water so if you want to think about tr it's a t1 relaxation filter short trs only give enough time for short t1 tissue to relax so that they can be bright so we're with a short tr we're excluding long t1 relaxing tissues which is water and edema remember we're following the water guys so here's an example uh a visual description or a visual of what tr is it's the time from the 90 to the next 90 that excites the same slice notice that the two nineties with the arrow they're the same color so they're the same slice so tr controls t1 contrast and for all different scanners there's different tr's that work best for t1 we have to find a balance between having a good t1 contrast but we also have to have good signal to noise if we go too short with our tr we're going to lose signal to noise so you want to be in the sweet spot or the goldilocks zone and that's what this image shows you right here so top right that is right out of my book and it shows that if you image with too short of a tr number one over to the left that image is kind of grainy because we're not giving enough time for tissue to t1 relax then not not enough of them are in the transverse plane to give signal then in the middle number two that's the goldilocks zone you notice we got good gray and white matter differentiation we've got good signal to noise now you're in the goldys goldilocks zone but if your tr goes too high for example 800 tr we're over where those lines are much closer together and we've lost t1 contrast so here's an example of a couple of questions that they might ask you so in a spin echo sequence short trs will make the images what so take a look at your answers take a second okay they could also ask you basically the same question but just the other way around in a spin echo sequence long trs make an image what okay so now for the answer the first one spin echo pulse sequences short trs will make your images more what they're going to be more t1 remember tr controls t1 in a short in a spin echo long trs make your image what okay remember i said earlier a tr question is probably pointing towards a t a t1 answer tr questions t1 answers okay in a spin echo long trs make an image what less t1 some of you might have said oh a long tr oh that's i need a long tr for t2 yes you do but the long tr just means less t1 contrast okay so relax t2 relaxation it's the other kind of relaxation so when we knock the protons down with the rf the protons don't just magically stand back up they start to spread out or fan out like like a fan that's loss of phase coherence so the top picture is taken for example and that's you might think i have a thing for chopsticks but they work very well so the top image there is a bunch of chopsticks all piled up on on each other and they're all together they're all in phase over time though they start to spread out they're starting to lose phase or what we call t2 relaxation the longer the te the more spreading out you get and the brighter fluids going to be because guess what remember the longer the te we're going to let fat get out of the way fat is short to t1 short to t2 so a long te is going to let fat basically get out of the way which is going to leave water to to be giving the majority of signal at t e so fat will be a little darker water is going to be a bunch brighter with a long te so t2 another name is loss of phase coherence you've got to know the definition of your te it's the time from the center of the echo to the center of the excitation pulse okay so from the 90 center of the 90 to the center of the echo so when you hear te now think t2 so again if you get a te question think it's a t2 answer or vice versa a t2 question you should start shading yourself towards ooh t2 question it's probably a te answer of some sort so what about a te question in a spin echo pulse sequence shorter tes will make your images what short tes are going to make them what pulse uh spin echo pulse sequence long tes are going to make your image what again this is basically the same question just wanting to know if you know what short and long tes do to your image contrast so the first answer is a short te is going to make your image not more t1 it's going to make it less t2 you're putting less t2 contrast into the image and spin echo pulse sequence long tes make your image what again more t2 long as soon as you hear long te you start thinking t2 okay here's another example of where they might try to trip you up and use a different name for the same thing how many of you guys knew that the word t e another name for it is two tau you don't have to raise your hand that's fine okay but that's an older name for it two tau the time from the 90 to the 180 is one tau the time from the 180 to the echo is one tau so one plus one equals two tau so if if they say oh if you use a long two tau that means a long te so just so you know that's what two tau means it's another name for te they could ask you a question something like if your te is 90 milliseconds when is your 180 degree refocusing pulse applied well if one tau and one tau equals t e so your 180 is applied at half your te the answer to that question would be 45 milliseconds half of 90 is 45. so that that's another quick little easy math that they could ask you all right now echo train length so an echo train when you hear echo train you start thinking all right fast or turbo spin echo they might say to you if your echo train is increased from two to four what's going to happen to your scan time well you've just doubled your echo train so you're gonna half your scan time so scan time is cut by a factor of the echo train that's nice and easy so the echo train or fast spin echo excuse me is a spin echo with multiple 180s sounds great multiple 180s multiple echoes you've got multiple lines of k space that you're filling per slice per tr and your scan time goes down because we all know mri it's all about time you want to scan fast what's the problem with multiple tes going into k space well the problem is if i want t1 waiting but i have a long echo train those later echoes are getting more and more t2 weighted so i'm putting t2 contrast into my t1 or supposedly t1 k space we need to manage that and that's something called the effective te which i'm not going to cover right now but when i hit pulse sequences um after our break we're going to talk about effective te and how important effective te is so just so you know just as um as an appetizer the effective te is the desired image contrast and it goes into the center of k space so if we want a uh t1 weighted sequence we're not going to put the late echoes into the center of k-space we're going to put the early ones because the early ones are less t2 okay so if you want t1 contrast you put the t1 weighted echoes in your center of k-space okay so here's a quick little example here is a fast spin echo pulse sequence there's five echoes each echo gets longer and longer so we've got a 10 millisecond te working our way out to a 50 millisecond te well the 50 millisecond te is much more t2 weighted than the 10 millisecond so if you want a t1 weighted image you're not going to put the 50s you're not going to make the 50s the effective te because that's going to put the 50s in the middle so we have to be careful with what we declare as our effective te also know that as the echo train gets longer we can potentially get more blurring and a high definitely going to get a higher b1 plus root mean squared now this number three here i wanted to bring up it's very important i can remember someone emailing me and saying they asked me this question on the registry okay they said what can what can you do if the image is blurry but the patient is not moving well their answers i don't remember what the exact answers choice that they gave you but one of them was decrease the echo train length and you go oh yeah most of the time oh the pa it's blurry and the patient's moving well no there's a difference between blur and patient motion so a long echo train can cause image blurring just so you know write that down highlight it put it in sharpie long echo trains potentially are going to cause blur the other thing that long echo trains are going to do is they're going to heat the patient up a whole lot more now when you guys are scanning or you're at clinical and you go oh my my scanner start out i got to do something yeah it didn't it it when it says it's sort out it means that pulse sequence that you're trying to run has got too much rf in it the sar level is a line that the scanner doesn't want you to go above with this particular sequence you're going higher or above you're exceeding the fda limits okay that's what sar means and you're getting that by rf pulses either echo trains or sap pulses and things like that so really what you're doing when you override the tsar and you do something either increase the tr or drop your echo train or delete some sats right those are usually your first few that you go to what you're doing is you're decreasing the b1 root mean squared which is your rf exposure so they could potentially and be tricky start to use either one of those terms sar or b1 root mean squared okay they could potentially interchange those so just be mindful okay all right inversion recovery ir here's another one of your intrinsic scan parameters now as soon as you hear ir you're thinking okay inversion recovery something's getting flipped down to 180 yep exactly so as soon as you hear inversion recovery say to yourself what's my ti because that ti is going to tell you which tissue is going to be suppressed and the easy way to remember which tissue is going to be suppressed is this short ti suppresses short t1 tissue and nice and easy the opposite is true a long ti suppresses long t1 tissue so what happens is during that ti so you knock everybody down to 180 during the ti that's when tissues start to t1 relax and if their net magnetization vector is in the null point whatever tissue is in the null point when the ti is done that tissue is going to be suppressed so for example you look at this curve top left those basically look like t1 curves do they not yeah they do okay what happens is the tissue's not starting to relax from the transverse plane but it's starting to relax from the 180 down below when the tissue's relaxation curve gets to the null point if you were to start your spin echo pulse sequence then that tissue is going to be suppressed so basically an inversion recovery pole an inversion recovery pulse sequence is nothing more than spin echo but it starts out with a 180 to invert the tissue the no point you can really think of as your transverse plane if you want to they will use null point that's the term they're going to use but in the but in the back of your mind you go oh yeah the xy plane that works for me that's a little bit easier okay next we're going to move on to flip angle so as soon as you hear flip angle you say oh that's a that's a gradient echo yep you're exactly right because the gradient echo flip angle has a huge contribution to image contrast so gradient echoes they have a tr every pulse sequence has a tr you can't get away from that and a te so gradient echos and spinecos they both have trs and tes tr has a big effect on image contrast in spin echo it doesn't contribute to contrast all that much or hardly at all in gradient echoland so t e is t e and that's still for t2 weighting so the longer your te the brighter your fluids gonna be but that third option or that third parameter in gradient echoland is your flip angle and that will affect image contrast a great deal so know that the flip angle and this see it's in red so you should you should have this in the back of your minds the flip angle is how far we're pushing or knocking down the longitudinal net magnetization vector into the x y plane are we knocking it down a little are we knocking it down a lot another thing that the registry likes to use for flip angle is they'll call the nutation angle n-u-t-a-t-i-o-n notation that's nothing more than the flip angle okay again you've got to know your terminology guys okay so in grading echoes the flip angle is usually something less than 90. it seldom that high but it can be if you really wanted to so a big flip angle in gradient echo land is 60 to 70 degrees and shallow is in the 10 to 20 degrees okay so what does that mean so here's here's a visual on big flip angle or a full flip angle that's down the bottom in the middle we're knocking the we're notating i'm going to use that term we're mutating the net magnetization vector the longitudinal net magnetization vector is getting mutated down to 90 degrees it's getting flipped to 90 degrees but a partial flip angle say i'm just going to say 45 degrees which is the top right the notation angle is 45 degrees okay so another term that you guys have to remember okay so um let's see yeah all right so again tr doesn't do much to image contrast in gradient echoland it will give you more slices and or decrease tissue saturation but that's about it so tr has been replaced by flip angle for t1 contrast i'll get to that in just a second t e is t e long t e's bright fluid short t e's dark fluid flip angle in spin echo land is 90 degrees unless otherwise specified so when you knock those protons down when you mutate them into the 90 degree or transverse plane they have a lot more t wanting to do okay t1 that's a that's a stevia ism they've got to go from the transverse plane into the longitudinal so they have a lot more t wanting to do so because they got a lot more distance to go there's a much better chance that you're going to see t1 relaxation differences between them so you can think of the flip angle in gradient echoland as how much t1 you're putting into the the echo that you're going to get back that future echo is it going to have more t1 in it or is it going to have more t2 in it okay so flip high flip angle t1 think of it as your volume control for flip angle excuse me flip angle is your volume control for t1 waiting now let me this let me go back real quick when i was first learning mr i i struggled with is this t1 is this t2 what the heck is this a physicist that i was working with at the time up in boston he said gradient echoes consider them all t2 weighted until you do something about it that doing something about it is adding a making the flip angle higher put some t1 in it and shorten the te take some t2 out of it okay so it's t2 weighted until you change it and make it t1 with a higher flip angle and a shorter te okay so waiting waiting we talked about image weighting t1 t2 proton density weighting is a way to term if an image has more of one contrast either t1 t2 or proton density so for example a cup of coffee if it's got equal parts of decaf in regular it's not weighted either way i'm a starbucks strong coffee kind of guy so i use this analogy all the time so if it's half and half it's not weighted either way but if you say put in 85 regular and only 15 percent decaf well now that's decaf weighted and i'm going to go into this and equate it to mri land in just a minute so images mri images are similar in that there's only one kind of contrast that should dominate one of them is going to outweigh the other two so you got three contrasts and which one is going to dominate so for example in a t1 weighted image you've got bright white matter okay next to not so bright gray matter so the tissues are either bright or dark because of t1 relaxation differences so a good t1 contrast is very important in the brain so t1 we're going to have bright fat and dark fluid so why do i say gray matter and white matter well white matter has a high protein content to it it's got called myelin it kind of acts a little bit like fat it's got a short t1 but gray matter has a lot of water in it so it acts like water and we want to see good t1 contrast that's very important in the brain radiologists want good contrast now let's flip that the other way with long tr and the long te to make it t2 you're going to have bright csf next to grayish white matter because the white matter is a short t1 relaxing tissue so between csf and white matter you're going to get t2 relaxation differences again white matter has a lot of protein in it and it's a short t1 relaxing property so you put something bright next to something dark and that's contrast and that's what mri is all about contrast so just to review tr and te tr filters out long t1 relaxing tissues makes them dark it's going to let you see short t1 tissues because you're only letting those tissues t1 relax so here's a stevia ism and i'm full of these if you image fast you're going to see fast which means a short tr and a short te you're going to see fast t1 relaxing tissues so the opposite is is true the te is going to filter out short t2 relaxing tissue which would be fat gaddon protein so when you image slow long tr and long te you're going to see slow t2 tissues so here's my contrast triangles i developed these i don't know i didn't pattern them so um but i developed these several years ago i figured it out something popped into my mind one day when i was working for ge so you've got the three contrasts t1 t2 and proton density our inherent scan parameters that we apply to the inherent tissue parameters is going to give us triangles like this a short tr is going to maximize t1 that's triangle number one and the short te is going to minimize t2 contrast so we'll have bright fat dark fluid and proton density is just kind of process of elimination it's down the bottom triangle number two we're going to use a long tr and we're going to minimize t1 minimize put it down the bottom short so a long tr puts t1 down the bottom along t e is gonna maximize t2 contrast put it up the top and then again process of elimination proton density is down the bottom finishing up long tr is going to minimize t1 short te is going to minimize t2 so process of elimination who's going to win if you minimize two of the three the two that are minimized are going to let the one that's not minimized win okay so here's our trte combinations okay so these images here they're all the same slice they're all the same patient so top left they might say to you is something easy as what waiting is this image here well you look at it you say well bright fat dark fluid decent gray white matter differentiation i think that's t1 and you would be correct bottom left what waiting is that image okay well let's see fat's not too bright and the water is very bright so i think that's t2 and again you would be correct bottom right hmm let's see wow everything's bright here so everybody's in the transverse plane at t e so let's see if everybody's bright i think that's proton density and you would be correct now the fourth and the black sheep of our contrast combinations is top right yeah that's a little emoji throwing up yep blah that is a short tr to maximize t1 and a long te to maximize t2 so what happens is you get something oops you get something that looks like this i'll go back guys don't worry slides got out of water so here's the long tr uh a short tr to maximize t1 and a long te to maximize t2 you maximize two of them nobody wins that's pretty terrible okay you hopefully you don't hand in images like that that would be you'll get the phone call from the radiologist saying what the heck are you doing over there okay so when i say good t1 uh good t1s and with good gray white matter differentiation over on the left hand side it's t1 fluids dark that's why gray matter is dark it's got a high water content it's t1 weighted myelin in the white matter acts kind of like fat it's got a short t1 it's nice and bright so that left hand image is beautiful t1 weighted contrast on the right hand side long tr long t e so we've got a bright gray matter because it's got a high contrast it's got a high water content so it's bright and the white matter because it's good it acts a little bit like fat it's not fat but it acts a little bit like fat because it's protein is going to be darker with the long te so let's see we got a chat here let's see so yeah any questions anybody i'm not seeing any questions so i'm just going to keep on going either you're either i'm totally making sense or you guys are completely confused and um either way that's what i'm here for if you've got questions guys let me know that's what i'm here for that's what you paid the money for okay one last thing let's put this all together we'll bring in proton density long tr you minimize t1 short te minimize t2 and what happens is tissues that have a high proton density of bright and tissues with a little bit lower proton density are a little are darker okay all good so far good we'll take that okay all right so we already went through this okay so once again here's three images t1 t2 and proton density let me ask you this guys just think about it for a second which one of these images of the three t1 t2 proton density which one has the highest overall signal to noise all factors remaining the same except for tr and d e which one has the highest overall signal to noise just think about it for a second if you want to chime in with a chat please do if you don't that's okay too which one is the higher signal to noise okay well uh here's a chat here's a brave soul i think yes jane excellent proton density why well why is it at the higher signal to noise because everybody's giving you signal everybody's in the transverse plane at t e just some people are not given as much signal as others but everybody's giving it t1 you're excluding water for example t2 you're kind of excluding fat so you are making some tissues not give signal so t1 and t2 are high contrast images but not high signal to noise whereas proton density has great signal to noise good answer jane okay and we're going to finish up right here with my image quality triangle so what makes a good image three things contrast similar to noise and resolution the most important one is contrast if you don't get something bright and something dark on your image if they're all the same signal intensity you've got no contrast and the rad can't see what tissue is which you also need signal to noise if you don't have signal you can't see your contrast and then last is resolution you you've got to be able to see the edges of one structure to an adjacent of another so you need those three to have a good image quality all right so let's real quick let's review so all tissues have a unique t1 t2 and proton density our inherent scan parameters trte flip angle etc when they're applied to the tissues their inherent tissue parameters you're going to get different contrasts t1 t2 proton density are going to be made more apparent with our selective tr's and tes so the inherent scan parameters are going to maximize or minimize one of the three contrasts remember tr is the time from the first excitation to the next excitation that excites the same slice te is the definition from the center of the echo center of the excitation pulse to the center of the echo tr controls t1 and te controls t2 okay so question is with an extremity or spine image or abdomen for example yep just look and know how do you look to know if it's t1 or t2 well the at the old adage is when you look at an image for the most part and in general if you see bright fluid it's going to be t2 if it's got dark fluid it's t1 now one little caveat for that is when you start talking about stir now you've got something that's not t2 but you got bright fluid so just in the back of your mind know that all images with bright fluid are not necessarily t2 so they would have to tell you on this stir image or on this inversion recovery image you know what you know what fluids are going to be bright on a short time inversion recovery well fluids are going to be bright so you need to know your factors there but in general look for fluid and if you see bright fluid it's t2 okay where were we here all right so um tr controls t1 te controls t2 tr and gradient echoland doesn't have a whole lot to do with contrast that's mostly flip angle higher flip angles are going to increase t1 contrast remember flip angle is the volume control for t1 contrast te is still for t2 inversion recovery or the ti is what suppresses the tissue that have different t1 relaxations so short suppresses short and long suppresses long in when you're doing inversion recovery echo train length you know it's a series of 180s each causing an echo and your scan time goes down by a factor of your echo train length we're to we're going to do that mathematically when i talk about pulse sequences because that's where we're going to do scan time math as echo train goes up you get more t2 contrast which is why you want to use a short echo train for t1 weighting because you want to keep t2 out of your k space echo train goes up rf exposure increases and you potentially get blurring okay let's see and just the quick review of my image triangle short tr short te is going to maximize t1 okay image short c short image long c long so a long tr is going to minimize t1 and a long t e is going to minimize is going to maximize t2 okay so if you if you can get a visual on these it's going to help with what tr's and what tes so when you start you might say to yourself oh my god what if they ask me what tr i should use for t1 i really don't think they're going to ask you a specific tr they might ask you a range at 1.5 tesla they might say well the what range of trs would be best suited for t1 well hopefully you already know around 400 to 750 is a good range for tr for t1 at 1.5 at 3 tesla it's going to go up a little bit i probably doubt they're going to ask you very many specific questions like that they might ask you hey at 3 tesla does you know true or false at 3 tesla does t1 relaxation get longer yes it does so if it gets longer another question they could ask would be will your tr get should you use a longer tr for t1 at three tesla then that answer would be yes okay does the ti change going from 1.5 to 3t yes it does we're going to talk about that when we talk about pulse sequences okay so that's the end of this one i am going to keep on going i'm going to jump into my next one any questions guys don't be afraid i do sure what's your question maybe a little choppy from the connection okay yeah gene said audio not working um jane you can type it in the chat that's also another option but anybody else have any other questions that was a good topic steve i didn't know you were still here i thought you were of course oh my god the ghost the ghost of christmas present okay sorry i'll put in chat okay why is it so hard to get t1 versus t2 all right well it's really not you can you can get are you talking a particular field strength are you talking uh t1 uh at 3t or are you talking t2 at 3t any particular field strength okay um well i'm i'm thinking that you're asking why is it hard to get t all right and now i re now i understand the question it is a little bit more difficult to get good t1 contrast because you've got a narrower window to get that contrast okay so think of it this way t2 okay hold on okay um all right so let me answer jane's first so um you've got a narrower window of opportunity for t1 you've got a narrow window for tr and you have to keep a short te if you start getting too generous you lose t1 contrast whereas in t2 land you can use if you want to run an 8 000 millisecond tr for a t2 knock yourself out all you're going to do is make that image less t1 right and you can use a long te to give you more t2 so as opposed to using an 80 or a 90 te which is going to give you bright fluid if you bumped up to like 120 okay cool you can do that so it's a lot easier in theory to get a better a good t2 than a t1 because p1 you got a pretty narrow window of opportunity okay um all right so let's see does the ernst angle apply to gradient echo sequences yes it does that's a that's an old trick um that's been around for a while especially on so if you're talking um 2d time-of-flight mra sequences for example mra sequences are all gradient echoes correct so if you up your flip angle so suppose you've got a short tr and a short te but uh let's say a 55 degree flip angle on your t2 time of flights your 2d time of flights in the carotids and you're saying hmm i wish i could get that background tissue a little bit darker no problem bump up your flip angle basically you're going to get to the ernst angle and then beyond and what's going to happen is your stationary tissue is going to lose signal to noise and that's going to increase your image contrast does that does that make sense allison is that does that work for you is that what you mean okay um also know that the ernst angle it applies to all gradient echoes not necessarily time of flights if you keep upping your flip angle in gradient echo land you are going to lose signal to noise because the ernst angle is that flip angle for a given tr that flip angle that is going to get maximum signal to noise okay if you go past your ernst angle signal starts to drop that's what your ernst angle is so it applies to everything okay in echo land um okay so a question from nikki um proton density from what i've studied is good for cartilage correct yes i work on a 1.2 hitachi and omni protocol is almost always pd and pd fatsat yes almost all your msk work there's going to be a stir or a t2 fat sat because remember we're following the fluid right but you just want to see proton density which tissues have a lot of protons so in a t1 right what fluid is dark on t1 i mean excuse me what tissue is dark on t1 fluid okay so if you've got a meniscal tear the meniscus is dark it's black it doesn't give you any signal so if you've got something black on something black okay that's my phone and that's my remote control if you've got black on black what have you got you don't got no contrast okay proton density though is going to show black meniscus but because a tear in the meniscus is going to have protons in it right and you're going to do something that shows you signal from a lot of protons now you're going to have a tear that's showing up in your meniscus that's why almost all your msk stuff has got proton hope that i hope that answers the question okay so i am going to okay cool all right i'm going to close that and i'm going to move on to x uh intrinsic gonna go to extrinsic parameters and then we'll take a quick break guys so off we go hello all right lost my mouse come on mouse don't fail me now hey steve i just gotta ask uh let's see who am i looking for here it's alison zorns uh if you can just private message me on the chat your a-r-r-t id number okay it's for your certificate okay oh so you're talking to alice not me yep allison just allison and your arit id number okay because you've got mine already so yep what we have here is a failure to communicate okay so all right guys we're going to keep on going extrinsic or geometric parameters so we're going to talk about field of view slice thickness matrix scan matrix receiver bandwidth which is a lot of people's um you know it's the bane of a lot of mri techs if you don't get it you are not alone okay then we're going to talk about necks and concatenations all right so the geometric parameters are i mean the extrinsic parameters are geometric we're not talking about tr's tes and flip angles we're talking about the geometry what are we doing to our pixel size okay so that geometry those guys are going to change signal to noise and resolution everything else we just talked about for the last hour was contrast remember that contrast triangle i mean the image quality triangle contrast is at the top that's the most important if you don't have good contrast you're going to hear it now we're talking about geometry your pixel size okay so first thing i want to make clear here is i've used this term signal to noise so it's the words signal to noise snr or just s slash n signal to noise it's a term that we use daily in mri and it's nothing more than a simple ratio so it's signal to noise how much signal are we generating in our sequence versus how much noise is in the system during the sequence run all right and that signal to noise we're going to talk about that a lot when we talk about necks or signal averages or acquisitions whatever your system happens to use so just know that we get signal from the voxels okay and consider signal a constant we're only going to generate so much signal during from these voxels during this sequence run so we generate a finite amount of signal what we don't want to do is generate noise because if we generate a lot of noise then our ratios get screwed up so bottom left we got a big s so we're getting a lot of signal and we want to have only a little bit of noise that's beautiful that's ideal but over on the right hand side something went wrong we're not generating a lot of signal oh but our noise just came up oh man this image is lousy i gotta do it again so if your signature noise is good thumbs up if your signal to noise is bad you're probably running it again or maybe there's a system issue okay so we talked about contrast signal to noise and resolution so our field of view is going to be a huge contributor to signal to noise it's a big contributor or a big killer of signal to noise our field of view also determines pixel size because we'll do this math shortly field of view divided by your scan matrix is going to give you your pixel size and i guarantee they're going to be asking you that they're going to give you this big word problem this tr this te and the head coil and the patient's claustrophobic and they're going to put in all kinds of you know what and then they're going to say what's your pixel size and you've got to go in and tear apart this word problem and find your field of view and find your matrix and do a quick math equation which we will do in just a minute okay all right so field of view they might say to you where does your field of view come from well it comes from your frequency encoding gradient at the end of the sequence so you've excited you've phase encoded you've refocused and then the echoes forming you're going to when the echo forms you're gonna turn on your frequency encoding gradient also known as your readout gradient they love to use that term read out because that gradient is turned on while you are reading the gradient excuse me excuse me stephen we can't see the screen you can't see the screen oh dear what did i do wrong here okay hold on ah you know what i didn't do i didn't share i've been i didn't share i'm sorry thanks thanks because i was just talking to myself then wasn't i interested in there all right there we go all right sorry about that dudes okay so let's see so um all right here's the second slide so we want to generate signal we get it from our pixels or our voxels we want we get a finite amount of signal for that sequence we want to have low noise if we can help it so signal is what we generate noise is what lives in all electronic equipment so we want to have a good ratio of signal to not much noise that's the left hand see big s little n that's going to be good image quality we're going to have good signal to noise the reverse of that is when we don't generate a whole lot of signal for various reasons but our noise comes up okay so our signal-to-noise ratio is very important so signal is what we generate it's basically considered a constant for that sequence okay and we are going to have noise in the background it's what all electronic equipment does so we want high signal and low noise ideally okay so then all right so your field of view is a big killer or contributor to signal to noise and i'm going to show you some math in just a minute so the field of view is what also determines your pixel size your field of view divided by a matrix so again they might give you this big word problem with all kinds of stuff in it and then at the end say what's your pixel size just to try to get confuse you so your field of view divided by your scan matrix phase and frequency so here we go so now at the end of it when you turn on your your field of view it comes from your frequency encoding gradient so you excite you phase encode you refocus and then when the echo is forming you turn on your frequency encoding gradient also known as like i just said you are read out gradient because that gradient is turned on while you're reading the echo they love to use that term so the size of your field of view comes from the amplitude of your frequent of your frequency encoding gradient so a shallow gradient a is going to give you a big field of view for example 400 like it says right there you're doing an abdomen you're turning on a big field of view so your frequency encoding gradient does not have to work that hard but oh now i'm doing a brain i'm doing a much smaller field of view i'm doing 200 field of view that gradient has to turn on stronger in order to make that 200 millimeter field of view okay the shallow gradient is probably going to let you get a shorter te because the gradient doesn't have to work that hard it doesn't take that long to come up to strength the steeper gradient is going to take a little bit extra work to turn on that gradient to the size of the field of view okay steep gradient small field of view takes a little bit more time so why did i say that the field of view is a huge contributor to singleton uh a huge contributor or detractor of signal to noise so let's think of it this way if you double your field of view and change nothing else you're going to four times your signal to noise so you double your field of view you're going to get four times the signal to noise and vice versa if you have your field of view you go from 400 to 200 you're going to quarter your signal to noise because both of those changes you are you change the field of view in two directions so you double it two by two equals four your pixels are four times bigger or half it they're a quarter the size that they were so your signal to noise goes down because you are having pixels and they represent less protons per pixel okay so um let's just do look wrong way come on stephen why is this changing there we go all right so bigger pixels equals lower resolution so that's where field of view comes in as far as resolution signal to noise goes bigger pixels equals increased signal to noise because it's all about putting pixels in the pathology you want small pixels in your pathology while maintaining a decent signal to noise the other problem with having big pixels is you get something called partial volume and this is something else that they like to ask you what the definition is okay fields of view large field of view big anatomical coverage you get less chance of artifacts aka rap so big pixels fewer pixels in the pathology bigger pixels have more hydrogen so you got more signal to noise but does that skip i think it skipped anyway um let me tell you the definition i thought i had the definition of partial volume maybe it's not another slide partial volume means the signal intensity of a particular pixel is the average of all the signal intensities of all the tissues inside so what i mean by that is you got a pixel that's got some bright and some dark tissue in it that pixel is going to look gray that's going to happen when you've got big pixels whereas if you go to a smaller pixel it's going to be a truer signal representation that's what partial volume it's an averaging of all the signal intensities of all the tissues in that pixel or voxel okay questions nope i think that's good okay so let's go on to slice thickness now there's two methods of slice thickness adjustment this slice select gradient amplitude and there's transmitter bandwidth the most common and the most time efficient is slice select gradient change okay so it's basically the same as field of view so if you have a shallow slice select gradient you're going to get a thicker slice the tissues along that slice select gradient are all going to change in processional frequencies so the the processional frequencies are not packed in so tight they're not all crammed in with a shallow or wide gradient application okay so steep gradient is going to be a narrow slice thickness so two millimeters for example all those same processional frequencies are all crammed into a smaller area so the most common and most time efficient way to change your slice thickness is to change the amplitude of the slice select gradient you don't know which one your scanner is going to use but i guarantee you it's probably going to change the amplitude of the slice select graded it's more time efficient and i'll show you that in just a second all right come on you can do this come on there we go okay so here is a transmitted bandwidth of 63 megahertz plus or minus five that's a that's a transmitted bandwidth so you're going to excite a certain area of tissue a different area of tissue with a steep or shallow gradient so in yellow we have a shallow slice select gradient that range of frequencies along the slice gradient is in a bigger area than not so crammed in whereas if you use the same amount of rf the same rf bandwidth but increase the amplitude of your slice select gradient you're going to get a thinner slice that's the most common and that's the most time efficient now the reason why i say time efficient is when you want a thicker slice okay if you use the same amplitude slice select gradient but you want a thicker slice you can do that but you're going to put in more rf into that rf pulse you're going to put in more frequencies into the rf pulse and it takes longer time to apply all that extra or wider transmitter bandwidth because the rf pulse is an oscillating it's a sine wave it's going to go above and below that that dark line that's measured zero it's going to go above and below above and below over and over and over and over and over again that's called crossing zero the more rf the more time it has to cross that one it needs to cross that line it's going to take more milliseconds to do it so as a time saving thing or time efficiency you're going to change more than likely your scanner is going to change your amplitude of your slice select gradient you're going to cram more frequencies more tissue into a narrow into a shorter or smaller slice thickness so trade-offs for slice thickness recall a bandwidth is stated as plus or minus megahertz so 63 plus or minus 5 means you're going to put in an rf pulse between 58 and 68 that's going to be your slice that's your bandwidth if you change your amplitude you're going to put that same amount of frequencies in a smaller or larger area the trade-off with that is you're going to increase your signal to noise with a thicker slice there's more protons you're going to decrease resolution you're going to increase partial volume here's that partial volume that i was looking for earlier earlier a thin slice you're going to it's all the opposites a thinner slice so five millimeters versus three three is gonna have a lower signal to noise you're gonna have higher resolution we all know that you're gonna be able to see smaller things but you and your partial volume is going to go down and partial volume means that the signal intensity of the pixels is the average of all the tissues contained within a bright tissue and a dark tissue in the same voxel is going to equal a gray pixel or voxel so here's your slice select gradient here's how when you turn on so that is a profile of abraham lincoln he's my favorite president that's not going to be on the test so when you turn on the slice select gradient you are going to change the processional frequencies of those tissues along that slice select gradient okay so when the gradient's applied in whatever whatever plane you are doing sagittal coronal axial it doesn't matter you are going to change that processional frequency in that direction an rf pulse is applied to the tissues in the presence of that slice select gradient so let's just say for example you turn on that gradient and you put in an rf pulse that's just just for yucks and just for ease if you turn on that gradient and you apply a 61 megahertz rf pulse his chin is going to resonate the nose eyeball forehead they're not because you didn't call their name and they're all at their own different processional frequency and the only thing that's going to respond is going to be at the chin at 61 megahertz however if you turn on that gradient the exact same gradient and put in an rf pulse of 65 megahertz then his forehead is going to resonate so you've got different processional frequencies along the slice select gradient okay i've used the word pixel a lot what's a pixel and what's a voxel they love to ask that question what's the difference between a pixel and a voxel a picture is a pixel element and it has two dimensions height and width no thickness pixel pixels are measured in millimeters by millimeters height times width so when you're scanning it's important to have as many pixels in the pathology as possible which i've already said but you still need to maintain reasonable scan time and signal to noise so you've got to have balance i had a student ask me one time how much signal do you need and i think they were just seeing what my answer was going to be and i said enough you need enough signal to noise to that goes along with scan time and resolution so you need enough signal to get a good picture so here's here's an example of three different fields of view with three different matrix which one is going to have the higher resolution well hopefully you set the 16 by 16 because you are going to have smaller pixels the one in the middle you're going to have higher signal to noise but less resolution and over on the far end what your signal to noise would be out of that you would have to do some math the scanner would do the math for you but i threw in that 8 by 16 for a particular reason we're going to talk about that in a minute so the matrix is how many steps or how many times you divide your field of view in each direction not being phase or frequency so pixel size is field of view divided by your matrix so for example if they said your field of view is 250 and your matrix is 224 squared what's your pixel size all you got to do is divide 250 by 224 and that'll come up to 1.1 so your pixel size is 1.1 by 1.1 higher matrix numbers means your pixels are smaller which gives you less signal to noise but better resolution higher matrix gives you better resolution that's the one on the right but lower signal to noise because there's less hydrogen represented in each pixel it's also going to mean a longer scan time and we're going to go over scan time after we take our break but our basic scan time formula is tr times phase times next so the more we divide up our field of view the more phase steps the more us scan time goes up so i put in that oddly shaped pixel size earlier because they may ask you what is isotropic versus anisotropic they can ask you that a couple of different ways so if a pixel has equal dimensions of height and width okay left to right so on the left image that's equal top and bottom it's isotropic iso means the same and anisotropic means not the same and means not so you've got not isotropic or not square pixels over on the right hand side so that would be the example of a phase of a scan matrix of 192 by 256 for example or 256 by 320. your pixels are not going to be square if you don't divide them up the same number of times so here's a question that i heard was on the registry a while back and this is going back years but they said a sequence with a 225 field of view and a 256 matrix will give you what kind of pixels so as soon as you say hmm okay well they're not going to be square so that means they're going to be anisotropic so they might ask you a question like that do you know what and versus iso means another question that was on the registry was which matrix combination below will give you isotropic pixels so you need to look at those and go hmm which one is which okay so that would be the 160 by 160. even though you're probably never going to scan a 160 by 160 matrix it's going to get it's going to answer the question that they asked you even though the answer doesn't seem right to you and i would never scan a 160 by 160. what planet are they on it doesn't matter what planet they're on tell them what they want to hear okay so the number of phase steps directly impacts scan time resolution and signal to noise more phase steps means a longer scan time because every phase step just in regular old spin echo imaging one phase step equals one tr you're filling in one line per tr so tr times phase times next so we're going to talk about that more after we come back from our break so when we as technologists look at a scan time and you go oh five and a half minutes i don't think so tim what's the first thing you're gonna do if you're saying to yourself i'm going to drop the phase matrix yeah you and every other tech on in the planet is going to do that first okay um but what they are going to be asking you questions about is they want to know if you know what changes are going to do what is going to be the consequences of your action they have been known to say what are the effects on signal to noise and resolution and scan time if you decrease the phase matrix do you know the consequences of your actions and you have to know that guys so when you drop the phase matrix what's going to happen to signal to noise it's going to go up what's going to happen to resolution it's going to go down and like i just said when you drop your phase your scan time goes down so they can ask you these three for one questions okay just be told you need to know the consequences of your actions all right so what's the difference between a pixel and a voxel well i said earlier a pixel is basically a dot it's it it's got height and width but there's no volume there's no slice thickness okay so when you hear voxel you automatically think oh my 3d um circular willis sequence oh that those are the voxels yes they do so does 2d when you do your 2ds in the brain you've got pixels right and those pixels have slice thickness so a 2d sequence also has voxels don't get locked into 2d does not have voxels because they do okay nowadays the modern scanners now they're going to tell you voxel dimensions they'll give you eight by eight by three great okay what does that mean well hey 0.8 by 0.8 those are pretty small pixels yes they are what what's your voxel dimension okay what are your voxel dimensions they might say to you what are you've got dimensions for a 180 field of view and a 256 matrix with four millimeters thick okay now you're gonna do some math so 180 divided by 256 that comes out to 0.7 i already did that on my calculator when i was making up this slide so your your voxel dimensions would be 0.7 by point seven by four pretty easy math but you gotta know that these are probably coming at you and i would suggest you maybe practice these a little bit another question that they like to ask is what's your voxel volume uh oh again simple math you're going to have a calculator 0.7 by 0.7 by 4. so multiply those that comes out to 1.96 cubic cubic three directions cubic centimeters so be prepared to do this math guys it's pretty easy once you hear it once you see it it's like oh yeah okay that's easy but if you don't know that these are coming at you you're going to sit there and you're going to start to get stressed and you don't want to stress when you're sitting for this test any more than you have to okay so this voxel right here what if they showed you this pixel and there's this image and they say is that isotropic or not is that an isotropic voxel well it tells you that it's not isotropic but if all you got was one by one by four is that isotropic voxel no it is not okay isotropic means the same so that would have to be one by one by one or four by four by four that's isotropic so iso means the same so they they could potentially give you a visual and ask you is that isotropic or not it's an isotropic pixel but it's not an isotropic voxel they've been known to ask what kind of voxels are ideal for 3d reformatting well isotropic voxels are what you want okay so when you do your circle of willis or anything that's 3d ideally it's let's just say one by one by one or 0.6 by 0.6 by 0.6 now you're saying yourself wow those are pretty small pixels and voxels yes they are so how can a 3d with one millimeter thick slice have good signal to noise well the answer is when you put in yours when you put in your excitation pulse for a 3d sequence you're exciting the whole volume suppose your volume 60 millimeters that's a great big thick slice you got signal to noise to burn but then 3d comes from an additional slice select gradient application that makes your va your raw partitions so yes a three millimeter slice thickness is going to have high signal to noise but if you drop down a one you're not going to have the signal to noise especially if you're doing 2d but 3d all factors remaining the same you're going to have higher signal to noise with 3d because 3d is not acquired the same way as 2d okay so for example when you're doing extremely high resolution images you're looking at the iacs you want to do a cysternogram they want to take a look at the iacs you're doing small fields of view small slice thicknesses high matrix okay so you want to use isotropic voxels that's what rendered that semicircular canals and the iacs over on the bottom right isotropic are ideal for reformatting they've been known to ask you that okay so here's another question which set of scan factors will yield the most key term here is most underlined right which one is going to yield the most isotropic voxel okay you so you look at these four answers and you go all right all right which one is outright not going to be isotropic in any way shape or form well hopefully you said a because the matrix is not square okay so now your choices are down a b c or d okay you have to do the math here which one is going to give you say 2 by 2 by 2 or 1 by 1 by one you have to do the math here on your calculator and figure out which one and let me see here i want to say it was c all right if you want to take the time and do the math knock yourself out but 225 divided by 224 that's going to come out basically to 0.9 or maybe 0.99 or something and a 1 millimeter slice thickness so that's going to give you an isotropic voxel i hope that makes sense okay receiver receiver bandwidth and frequency matrix where we're we're almost done with this one guys we'll be taking a break in a few minutes so when you set up your matrix and you make your frequency matrix a certain number that's telling the receiver how many times to sample that echo so down below those are your blue dots so that's the work that the receiver is going to do it's got to measure x number of times i think down below i think that's like there's 33 dots there so how fast is that receiver going to sample those 33 dots okay that's your matrix a high matrix means there's more dots you've got to sample you've got more work to do more samples equals more work more work means a longer minimum te so the receiver tells how many frequencies are going to be mapped how many times they're going to map that across the field of view that's hertz per pixel the receiver bandwidth tells you tells the scanner how many frequencies are going to be across your entire field of view all right so now receive a bandwidth think of it this way it's a sampling rate are you going to sample all those blue dots at 60 miles an hour and be done real fast or you're going to sample it at 20 miles an hour and not be done so fast so the same thing comes across are you going to sample 256 times at 60 miles an hour or 320 at 60 or 512 times at 60 miles an hour so your sampling rate your receiver bandwidth is either the term that they hopefully should use because every manufacturer has a different terminology for their receiver bandwidth they should be using the terms wide and narrow to make it to make it um fair okay so if you're going to sample the echo 256 or however many times at a fast rate or a slow rate you're going to be done with that echo sooner so sampling at a high sampling a high matrix slower is going to mean your minimum te is going to come up so a narrow bandwidth sampling a high matrix means you're going to get better single to noise but your minimum te is going to bump out and your chemical shift is going to get worse so again the registry is more than likely going to use wide and narrow so don't get crazy trying to learn all three of them for the registry i would hope to keep it fair they're going to use wide and narrow so let's so let's look at this visually which is wide and which is narrow so a wide receiver bandwidth as i just said is a is a fast sampling rate it's also going to sample more noise because it's sampling basically the entire echo it's sampling where the echo is starting to form and there's not much amplitude versus noise so that black line running across that echo is noise so over on the left hand side circled in purple on the top we're sampling that echo where the signal to noise is pretty lousy then as the echo forms we're getting the good stuff we're getting the sweet spot and then that echo is going to fade away and we're going to sample that not so good stuff so you sample faster but your signal to noise is a little bit lower on a wide transmitter bandwidth your minimum te comes down which is a good thing but your signal to noise is overall lower now the bottom on the left hand side circled in purple that's where you start sampling with a narrow bandwidth well by there the amplitude of the echo has come up some it's higher than the noise so your signal to noise is a little bit higher so overall what a narrower receiver bandwidth is going to give you an overall higher signal to noise and they like to ask that kind of stuff what will a narrow bandwidth do for you as far as signal to noise artifact and te so you need to know those so if you remember one of them the opposite is true and i got a little graph here okay so we're commonly going to use a wide receiver bandwidth on t1s and proton densities because they have higher signal to noise all factors remaining the same than a t2 and you'll use a narrower bandwidth on a sequence like a t2 or a fat sat where your overall signal to noise is lesser okay what you want to do guys when you're scanning is consider fat sat sequences and t2s signal stop because you're taking out fat's contribution to overall signal to noise so all factors remaining the same your signal to noise is going to be lower on fat satin stirs and if you do something that hurts your signal to noise you're gonna you're gonna pay the price at the end okay so um over on the right hand side in the middle that graph with all the blue arrows you wanna you want to memorize one of those it doesn't matter pick one so if you say well i like narrow bandwidth because everything's going up yeah cool great know that a wide receiver bandwidth is going to do the exact opposite so your signal to noise is going to go down with a wide bandwidth your chemical shift is going to go down with a wide bandwidth minimum te is going to go down or minimize with a wide bandwidth and in actuality motion is going to decrease just a little bit although i wouldn't use this i wouldn't widen my bandwidth to try to decrease motion but because the te is so is shorter motion will be minimized just a little bit not a lot okay so memorize one of these directions and then know that the opposite is true okay um next acquisitions in nsa these all mean the same thing they're just different acronyms so nexus ge acquisitions is siemens nsa is um phillips hitachi and toshiba so know that when you increase the number of necks you're filling the k space more than once but you're also filling it on successive trs so each line of k space is going to you know if you say i want to do three next great you're going to fill each line of k space three times so you're going to fill line one with tr number one you want to fill line one again with number with um tr number two and then gonna fill line one again with tr number three so it's line one line one line one and then you're done you go to move up and you go to line number two then that's going to be tr number four five and six you're going to fill them on successive tr's you're not going to fill line one and come back 500 milliseconds later and fill line one again and 500 milliseconds later and scan line and fill line one again your motion is going to be absolutely god-awful so it's successive lines just know that okay now when i go up in my single when i need some more signal to noise and i go up in my acquisitions scan time is going to double so if you go from one to two acquisition scan time doubles but however your signal to noise does not it goes up by a ratio or this not the ratio excuse me the square root of the number of acquisitions so over on the right hand side the square root of 1 is 1. that's easy square root of 2 is 1.4 so if you go from one acquisition to two acquisitions you gain forty percent signal to noise at the cost of doubling your scan time taking that even further if you go to three acquisitions you're going to gain go from one to three you're going to gain 70 percent signal to noise that's great but i'm sitting here three times longer and the patient's got three times longer to move the only way you can double your signal to noise by using next or acquisitions is to go to four and that's a very inefficient way to double your signal to noise okay so why doesn't it go up hell i spent twice the time why don't i get twice the signal to noise because remember signal to noise is a constant noise is a variable it can go up so at two acquisitions or two next you might have low signal to noise on the first one and high signal to noise on the second one and you average those two out so it's not a direct relationship and finishing up with concatenation so what's a concatenation well i know what a concatenation is in mri land but webster's dictionary defines a concatenation as to link together or as in a chain or a series so in mri we use concatenations to keep image contrast and maybe you've done this during your clinical time or maybe you're doing it right now if you need for example 40 slices through i don't know the forearm okay i need 40 slices of coverage that's great but uh my ti is is over a thousand that's not t1 how am i going to do that well you can do a couple of things you can split up your stack and do 20 and 20 or you can tell the scanner do two concatenations so do half the slices at half that tr and then go do the other half of the slices at the other half of the tr so you're still getting your 40 slices your tr comes down to um 525 and contrast is maintained okay so you can manually do two concatenations you can even go up to three if you want scanners will let you go three four five okay but mostly concatenations are used to keep t1 contrast that also works very well for fat sat images too because don't forget fat sat is going to add to your tr because it's extra work extra work means extra tr everything we do in mri is related to how much work everything has to get done within your tr period okay all right so let's real quickly go over those scan time formulas that i just those different formulas that i went over pixel size is your field of view divided by your matrix your voxel size is your pixel dimensions times your thickness okay so 0.7 times 1.4 times 2 for example it's a 1.6 cubic millimeter also know that signal to noise in a 3d versus 2d it's the square root of the number of partitions so have you ever noticed when you're doing a 3d sequence as you go up in partitions oh i need a little bit more coverage on this circle of willis you go up a couple of lokes your scan time goes up but so does your signal to noise it's a square root relationship okay i'm going to skip past this signal to noise and partitions right now i'm gonna i'm gonna go past that because we'll talk about that later um square root i mean the single noise of a sequence as far as next our acquisition goes is a square root relationship so the square root of one is equal to one square root of two is one point four and right on down the line so to double your signal to noise you need to do four times the acquisitions and that is going to mean four times the scan time and um larmor frequency just to throw that in at you you are going to be asked that i'll bet uh a cup of coffee so wobble is equal to b zero times y so wobble or processional frequency is equal to the field strength the main magnetic field in tesla times 42.57 megahertz megahertz that 42.57 is the gyromagnetic ratio of hydrogen at one tesla that is a constant and they ask it all the time what is the gyromagnetic ratio of hydrogen at one tesla okay they can also ask you the dreaded gauss per um the gauss conversions okay it's actually rather easy if you think about it and you need to think of terms of a thousand okay we use that terms of thousands every day milli means a thousand so a millisecond is one thousandth of a second okay a milli tesla is 1 000 of a of a gauss the common one that they ask all the time is our pacemaker right when you say a pacemaker can't go past the five gauss line five gauss is equal to 0.5 milli tesla so that's just something you need to remember so if you know point if you know 5 is equal to 0.5 then if they said 1 gauss is 0.1 milli tesla it's a nice easy conversion okay and these i've kind of beat to death so i'm just going to keep on going right through them um anybody have any questions just feel free to hit me up i'm going to um pause this right now and i'm going to let you guys take a break for the bathroom or a drink of water and we'll come back here and let's see what time is it it is it is um 12 52 so i'm pretty much on time so let's give this about um about 10 minutes guys for a break and we'll come back okay all right so i'm going to stop sharing and i am going to pause recording and we'll come back and if you have any questions guys just um you know put them up on the chat board okay now i can share my screen okay so all right as far as the earth as far as ernst angle goes okay um ernst angle so for a given tr all right so constant tr as you change your flip angle as you increase your flip angle you're going to get more signal okay so flip angle is running this way i know i'm a terrible artist so just bear with me um so for a constant tr let's just say 500 milliseconds if you increase your flip angle say to 10 degrees okay you're going to get a certain amount of signal if you go to 20 degrees you're going to get a certain amount of signal and then 30 degrees and then 40 degrees and then let's say 50 degrees well what's going to happen is so let's just say this is 50 degrees right here if you go up to 60 what happens is your signal starts to drop and then 70 you know it keeps on going do you see a pattern here so for example for a given tissue 50 degrees is your ernst angle signal to noise increases signal to noise increases as flip angle increases to a certain point and then all of a sudden it is going to start to drop so anything or any flip angle over 50 degrees is gonna you're going to start losing signal to noise because the tissue doesn't have time to hello the tissue doesn't have time to okay come on we can do this tissue doesn't have time to t1 relax tissue doesn't have time to t1 it's got time to t1 relax here it's got time to t1 relax here but by right here it doesn't have time to t1 relax in the 500 milliseconds so flip angles higher than 50 for example you're going to lose signal to noise does that does that make sense is that a little bit better explanation hopefully it is okay so let's minimize that okay and we're going to stop share all right so hopefully that works for you guys um see two chats yes okay thank you all right cool all right so now we're gonna go back now we're gonna talk about pulse sequences here sequences there we go all right share my screen again there we go all right so we want to talk about pulse sequences and we're not going to talk about every single one but we're going to hit the high points so first question is how many kind of pulse sequences are there well if you guys start saying oh there's stir this flare there's t2 there's fat sat there's all this other kind of stuff well no there's only two real kinds of pulp sequences okay they're either spin echo based or gradient echo based everything else is just a variation on any one of those two so conventional spin echo fast spin echo stir flare dwi gradient echo base then you got gradient echo which is the conventional that you do say in the brain for looking for hemorrhage time of flight things like that so there's really only two kind of pulse sequences so one registry question might be they'll just show you this line diagram or this pulse sequence design and say well what sequence is that and you would say don't say t1 or t2 or whatever because you don't have trs and tes to back that up but if you see a 90 and a 180 ding ding ding ding it is spin echo based that's all look for the 180 and if it's there it's spin echo based it might it might be as simple as that if they show you a pulse sequence and there's no 180 then guess what gradient echo so don't let all this slice select phase and frequency encoding stuff get you confused right now we'll we'll cover some of that so if there's a 180 is a 90 180 complex it's spin echo if it's doesn't have a 180 it's gradient echo okay so here's your basic gradient echo pulse sequence look what's missing i got it circled in red there is no 180. so if you were to just see something as basic as this no 180 that's gradient echo so pulse sequences like i said earlier are a very rigidly timed series of rf pulses gradient applications and then reading the echo very very rigidly timed so each one of these things has its own job to do so for example number one that's on well it says it's on the rf line so that's easy so that's the x x excitation pulse simultaneously there is a slice select gradient applied whether that's whether that is axial coronal sagittal you don't know but it's just the slice select gradient after you excite you're going to put in a phase encoding that's number three notice how it shows gradations here because the phase encoding gradient is changed every tr because that's what's going to put this eventual echo in a different line of k space and then after you phase encode you're going to frequency encode you're going to turn on your gradient to um to frequency encode and that echo that's why it's called read so the frequency encoding gradient is on during the echo formation also called read so you're going to apply that frequency encoding gradient and measure the frequencies during that echo formation okay so this image here happens to be courtesy of siemens medical and this is more like what you're going to see those other ones that that you just saw those are mine that i did up on powerpoint they're a little bit crude but siemens has graphic designers and they can make it much better um so what you see here you know what pulse sequence is this well it says it gradient echo but what do you see that there's not one of a 180. so one of the things that they might ask you is a gradient echo is sensitive to or is affected by what three things so that's line four coming down because there's no 180 a gradient echo is susceptible to what i call the big three main magnetic field in homogeneities local field and homogeneities and magnetic susceptibility differences those three things help drop or affect lower the signal to noise whereas a spin echo with the 180 corrects for those so all factors remaining the same a gradient echo versus a spin echo gradient echo has a lower overall signature noise because there's no 180 that cleans up for those three so the question here would be what are the three things that will affect signal to noise in a gradient echo main local and magnetic susceptibility differences so let's go through those real quick so again i call them in my classes i call them the big three okay these terms here they should not be new to you you should have heard these at least a couple of dozen times before so main magnetic field in homogeneities is b0 is the main magnet that you're putting your patient into how good is it how perfect the manufacturers try to make it perfect but they just really can't i mean come on there's no things in this world that are perfect and then they're not perfect to begin with and then guess what we put the patient in there and they're they're the biggest cause of field in homogeneities going they've got air in them they've got a heart that's beating they got flowing blood um maybe they've got a mouth full of uh braces or something that's going to affect the field homogeneity so that's the main magnetic field the local field could be something as much as babe they've got a knee replacement well there goes your field homogeneity if you're scanning the knee that knee is going that knee replacement is going to mess up the field and home the field homogeneity okay they got dental work surgical clips spinal hardware that's going to affect the field right where you're imaging if you happen to be imaging that area okay um you know things that can really make the magnetic field really bad is oh say a stray oxygen bottle yeah hopefully that's never happened to anybody that'll ruin your day and then there's magnetic susceptibility differences and this is this gets worse at 3t the susceptibility differences but susceptibility is of tissues ability to be magnetized basically does it like to be magnetized or does it not like to be magnetized so water for example is considered diamagnetic i'm going to go into these diamagnetic actually repels the magnetic field just a little bit whereas ferromagnetic is going to is going to attract the field a little bit that's one of the reasons why an oxygen bottle will go flying and then there's diamagnet um then there's paramagnetic which is what gadolinium is and it only brings the field in a little bit i got some graphics on that so if you've got two tissues that are different in susceptibility you're going to mess up those are going to mess up the processional frequencies right there and you're going to get basically a little black hole a little small area of signal loss so here's an example of a lady i did from the um from the er away he's back and she was not having a good day this is a gradient echo pulse sequence i know you can't see the factors all that well but however the gradient echo doesn't have a 180 so it doesn't clean up for those little areas of hemorrhage that are in her brain so why do these show up as dark spots well because there's no 180 and you've got iron you've got blood that has hemoglobin hemoglobin contains iron which is a which is a metal so the little magnetic fields right around where the where that blood is is changed and when you change the magnetic field what do you always do you change the processional frequency so you get these little areas of signal loss so you cause increased dephasing because the iron changes the field and it makes the tissues dephase quickly so let's go into flip angle a little bit more okay or actually a review we already went over flip angle quite a bit in the other section but when we're talking we talked a little bit about spin echo already and now we're going into gradient echo so all sequences like you know always have a tr and a te gradient echo has a third parameter flip angle and it's how far will the net magnetization vector be tipped into the x y plane or mutated so the flip angle it can be associated with that signal like a funny looking a right here okay that just means it's a flip angle it's not denoting that it's 10 or 20 or 45 degrees it's just a little a that means flip angle so as you see something like that you automatically know ding ding ding ding this is a gradient echo and there's no 180. so what's the they might ask you what's the flip angle in a spin echo pulse sequence and hopefully you go hmm i don't know what it is well it's 90 degrees because we've said 90 degree excitation pulse how many times in our life so the flip angle for a spin echo is 90 degrees gradient echo it's usually something less than 90. remember if the flip angle in gradient echo land is basically the t1 how much t1 are you putting into those protons are you putting a lot with a big flip angle one not so much with a lower flip angle so you've basically seen this slide before as well so a flip angle is how far you bring things out of the transverse plane into the x y okay out of out of the longitudinal into the x y so that's t one so right here a full flip angle means the tissues have more distance to t1 relax go back to zero or they have more spin lattice to do so that's going to be more t1 weighted versus this guy right over here where you only partially flip them into the transverse plane they don't have that much t1 to do so the chances that you're going to see t1 contract going to see t1 relaxation differences between them is less okay so this would be t2 or t2 star weighted so remember earlier i said that a gradient echo is t2 well actually t2 star weighted until you do something to make it more t1 and more t1 means a higher flip angle more rf going into the patient and a short te because we all remember te controls what t2 long t e more bright fluid short t e less bright fluid okay so box a is an example of a high flip angle where there's going to get more t1 and box b over on the right hand side less t1 is being put in so you're not going to get t1 contrast so remember what you put in you get out okay so um gradient echo once again i i know i'm beating this to death but it's real important to understand the concept it was it was earth shattering when it was explained to me that way so their t2 waited until they're not till you make them t1 with flip angle and t e so you can theoretically make a proton density gradient echo we never do i mean that's that there's really no call for that but theoretically it's possible okay now we've done spin echo we're done with gradient echo and now we're going to go into fast spin echo or turbo spin echo the original name for fast spin echo was rare r-a-r-e like how do you like your steak cooked sir i like it cooked rare that means rapid acquisition with relaxation enhancement that was the old original name so it's basically a spin echo with multiple 180s multiple 180s is your echo train length okay so this particular pulse sequence right here is courtesy of siemens medical so i don't want to i don't want you to think that i did i made that up no i'm borrowing it from them so fast spin echo is plain and simple just a faster spin echo more 180s more rf more echoes more lines of case space filled in your tr so in this particular case here we've got a three echo train we're filling three lines of k space per slice per tr and for example a six-minute pulse sequence in spin echo land here is going to go down to two minutes you're cutting your scan time by by your echo train length by factor of your echo train length okay okay so now turbo of fast spin echo waiting okay short or long tr and te concepts they still apply so short tr short t e and a short echo train is going to give you t1 okay so short short and short the same thing is true for t2 long tr long t e and a long echo train is going to give you more t2 weighting but when you start adding in all these echoes now we've got multiple contrasts right because the as the te gets a lot longer fluid's going to get brighter so you can't put t2 contrast into a t1 k space you're going to start causing contrast differences okay so we're getting multiple echoes per tr we need to manage those contrasts i got that highlighted in yellow we need to manage those multiple contrasts that are coming in per tr so we need to selectively place them in k space for our desired contrast all right the t e is in fast spin echo is is now going to be called effective te so you got three echoes do you want your first one in the center k space or do you want your last one if you look at this example do you want your first one put into your center of case space or do you want your last one put into your center of k-space that's going to change your image contrast remember in k-space the center lines of k-space give you mostly image contrast and signal to noise while the outer edges give you edge detail and not much signal to noise so look at this example right here here's a fast spin echo there's a four echo train okay so great so our scan time just got cut down by cut down to a quarter you're filling multiple lines of k space per slice per tr each one of those echoes gets a little bit more t2 so how do we put those echoes where we want them we're going to change the amplitude of the phase encoding gradient so we did slice select and we phase encoded this echo then we 180 it with the same slice select gradient and then we frequency encoded that echo now because of the amplitude of this phase encoding gradient this echo right here is going to go to the edges high amplitude phase encoding gradients cause the echoes to go to the edges of k-space notice as we continue down the line here the phase encoding gradient is getting shallower and shallower it's less steep that means that these echoes that get the weak phase encoding gradient amplitude so this echo here is going to go into the center of k space note that all the gradient amplitudes are all exactly the same slice and frequency are exactly the same height and length the only one that changes is your phase encoding gradient every echo gets a different phase applied to it so that goes into a different line of k space so from what i'm just saying right here what i just told you this last echo right here was declared our effective te this guy because of the weak amplitude phase encoding gradient is going into the center of k space so why am i taking that much time to go over this well because you have to be careful with your echo train lengths you can't use a ver an extremely long echo train length in t1 land because we need to put the good stuff in the middle and if we don't have a whole lot of good stuff if we got greedy with our echo train length now the wrong contrast is going to start crowding the good stuff in the middle so just remember in k-space center gives you contrast cc kind of like cc and coke okay when you go to a buy you order a cc and coke it's coca-cola and i think it's crown royal the whiskey whiskey and coke remember cc center gives you contrast the outer edges the outer lines gives you edge detail where we put our echoes is huge in fast spin echo land okay so for example here's that line diagram again depending on what we picked for our effective te that's what's going to go into the center of k space okay so we've got a man we've got to manage these multiple tes by managing them i mean we're going to apply specific phase encoding gradients to these echoes all right so here's the here's the slide i was hoping to get to or i wanted to get to quicker so we've got a three echo train like you just saw right here three echo train over on the left-hand bottom side three echoes which echo goes in k space is going to be determined by the phase encoding gradient so we've got three echoes and i said to the scanner my effective te is going to be the 10 millisecond i want the first one to go into the center of k-space and the scanner says yep so it's going to start off with the weak amplitude phase encoding gradients and the gradient amplitude is going to get a little bit higher as you go through the echo train and then once you're done with your three echoes you move on to the next slice next time you come back to slice one it's going to be the same thing weak amplitudes first stronger amplitudes to the the last couple of echoes and that's going to put the tens in the middle of k space so your image looks like a 10 millisecond te if i said no i don't want the 10 in the middle i want a 30 millisecond t e scanner will say sure steve no problem and it will apply the strong amplitude gradients to the first echo and the weak to the last echo the 30. so it's going to put the 30s in the middle and i'm going to show you a picture right here okay exact same pulse sequence exact same slice exact same patient i declared the effective te is 10 milliseconds on the left my 10 tes went into the middle and i get a nice t1 weighted axial dark csf decent i think pretty good gray white matter differentiation okay good t1 oops i made a mistake and i said i want a 30 e here in the scanner i said sure steve no problem and it put the 30s in the middle now look at those two images figure five seven and five six which is the better t1 weighted hopefully you said 5 6 because the csf is darker with the 30 te on the right hand side i let too much t2 contrast come in and it's starting the csf is starting to brighten up so that's how we manage our echo train okay by declaring an effective te apply the weak amplitude phase encoding gradients to the desired te and put the less desirable stuff on the edges to fill in those lines of k space okay inversion recovery inversion recovery sequence will be timing an rf an inversion 180 pulse and exploit the t1 relaxation differences of fluids or even fat okay so we're gonna we can make fluids dark on purpose we can make fat bro dark on purpose so as soon as you hear ir you start thinking to yourself something's getting suppressed most of the time it's csf in the brain and a lot of times for msk and in the spine we're doing fat okay so recall that tissues have a distinct t1 and t2 relaxation okay they're actually opposite so fat is fast the first two letters of fat is the first two letters of fast fat is fast to t1 and t2 water is slow to t1 and t2 so they're opposites so an ir sequence it can and will suppress other tissues other tissues with similar relaxation times to fat or similar relaxation times to water are going to be suppressed so an ir sequence it kind of suppresses everybody who's got a fast t1 or it suppresses everybody with a slow t1 it's not selective fat sat itself is very selective but i'm going to go there in just a minute so what are some examples of fast t1 tissue well i i said them the very first hour fat gadden protein are fast t1 relaxing tissues so if you do a stir with a short t1 you're going to suppress fat gadon protein if you use a long ti in say flare you're going to suppress csf and urine if you wanted to but no one does flare in the pelvis so you don't have to worry about that so we're going to have three variations of inversion recovery we got stir short time or short tau inversion recovery we've got flare fluid attenuated inversion recovery which can be t1 or t2 weighted and we'll we'll cover those in just a minute okay so let's talk about a basic inversion recovery line diagram so it starts off with a 180 and then turns into a spin echo so basically if you see a 180 90 180 ding ding ding ding ding that is inversion recovery you don't know whether it's stir or flare because they're not telling they're not going to tell you what the ti is but if you see 180 90 180 it's inversion recovery they can ask you what happens during the ti no the answer there is t1 relaxation is happening we're letting the net magnetization vector of those protons reach the null point or not okay so the ti lets tissues t1 relax remember fat t1 relaxes faster than water which is which is huge so the 180 degree rf pulse knocks everybody down so before the 180 they're all pointing up at b0 they're all pointing with b0 after the 180 they all get inverted they're all knocked down they're all pointing the opposite way so during the ti tissues t1 relax we're going to use ti's that are short or ti's that are long so we can selectively suppress either fat or csf so over on the right hand side is my version of an invert of my version of an inversion recovery pulse sequence it's a little crude but it gets the job done over on the left hand side that might be more what you're probably going to see i don't know if they'll take it right from siemens medical but it'll be a lot prettier than mine okay so 180 90 180 okay bingo that's inversion recovery you don't know if it's stir or flare because there's no ti stated don't go don't get that particular about it now the first one that i'm going to talk about is stir short time inversion recovery or they can also call it short tau inversion recovery tau i think is greek for time so that's why it can it can be called short tau inversion recovery classic look for a stir is dark fat bright fluid very similar to a t2 fat sat fluids that we're going to be suppressing are going to be csf or that we're going to be seeing signal from in a stir is csf or edema remember i said earlier in mr we're following the water so on stir we're going to see bright fluid bright edema so the tr's are are very important believe it or not in stir and in flare but especially in stir so if you ever do a stirring you're looking you go boy that's awful long scan time wow that's almost five minutes how can i save some time and you say oh look at all that extra tr i've got i'm gonna cut that down yep and you knock it down till i don't know 1400 milliseconds and your images come up and they're terrible that's because you shortened the tr too much and you are saturating fluids which is what you're looking for stir you want to see fluid tumor infection radiation necrosis you want to see the water so don't shorten up that tr don't fall into that trap because remember the body's reaction is to force fluids to some to some area of insult the te in stir is usually in the 30 to 60 millisecond range just to help the fluids be a little bit brighter you go too long you're going to start hurting your overall signal to noise field strength is also going to change what ti you use because don't forget field strength changes the t1 relaxation of tissues and during the ti will let in tissues do what hopefully you said t1 relax okay so that's stir okay short time inversion recovery okay usually the ti for stir is about 150 milliseconds at 1.5 and it's about 190 to 200 at 3 tesla are they going to ask you that probably not but i'm going to give you some math that we can figure it out if you want because the ti's are going to change from field strength to field strength flare fluid attenuated inversion recovery of which there are two kinds there's t2 flare and t1 flare so t2 flare just remember long long and long long tr is classic in all t2 weighted images long te classic for t2 weighted images and a long ti that's going to help suppress csf so t2 flare long long and long t2 flare the long tr is gonna do what suppress t1 contrast the long te is going to maximize t2 contrast and the ti suppresses signal from csf now how can i maximize t2 for bright fluid right steve you've been saying that since we started this but however i want to make fluid be dark yeah you're going to make fluid be dark with the ti if you have the right ti you'll suppress csf but the long te is going to let you get gray white matter differentiation that's exactly the same as a regular old spin echo t2 so if you look at a t2 flare the gray white matter differentiation is the same as t2 but the ti is what makes the csf dark okay next time you do a flare in the brain you'll see it i'm pretty sure i got some images coming up now t1 flare well long tr because we have to accommodate that long ti and but we're going to use a short t e so that doesn't change go back to spin echo land short t e takes away 2 t 2 contrast so we're going to get a better t1 okay so t1 flare is long short and short don't be like this guy banging his head on on the uh on the screen don't do that okay so just remember when asked remember your basic ranges or that basic principle okay break down a big problem into a smaller problem because i just gave you all kinds of variables here right make a big problem into a smaller problem okay basics short and short is t1 so we're going to have a short a short t e and a shorter ti t1 you're going to have long and long that's going to be t2 another easy way to remember this is a short ti suppresses short t1 tissues long ti suppresses long t1 tissues okay ti times for fat are going to be around 150 to 170 milliseconds that's common will they ask you that probably not but you should know that we get the tis for whatever tissue that we want to suppress by multiplying its t1 relaxation time by 0.69 so just just roughly uh the t1 relaxation time for fat at 1.5 i think is about i think it's about uh about 240 250 or something like that if you multiply that by 0.69 you'll come up with a number right around 150. so that's how those numbers get devised okay there is no definitive ti so for them to start asking you definitive numbers is not fair they're not going to do it they might say is approximately but they're not going to okay ti's change also by radiologist preference some guys want a little bit longer ti some guys don't that's radiologist preference okay they may ask you what ti will be used to suppress fat well go wait a minute hold on fat is fast so i need a short ti so short suppress is short and long what do you do for water suppression you need a long ti because water is a long t1 relaxing tissue very easy principle okay so again where do those numbers come from you multiply the t1 relaxation time of a tissue no matter what tissue it is multiply it by .69 and that number you get is going to be the ti that will suppress it so here's that little formula that i just used so the t1 of fat at 1.5 is approximately 220 milliseconds you multiply 220 by 0.69 it comes up to 150 so if you make your ti 150 you'll suppress short t1 relaxing tissues at 1.5 same thing for 3t about 290 is the t1 time for fat multiply that by 0.69 and you come up with about 190 or so again there's no definitive ti it's just simple math so you saw these curves before so they might show you this curve and say what curve is that well you go women hold on it's a t1 but wait a minute it's starting way down at the 180. oh 180 that's inversion recovery so you know this stuff okay you're just not used to answering the questions so right is a t1 relaxation curve and it shows two different tissues fat and water we've got a fat being fast it gets through the null point quicker so if we apply our oops if we apply our 90. so we apply to 180 and we wait the ti when fat gets here if we turn on and start our spin echo pulse sequence right here we're going to suppress fat if we invert everything and wait and wait and wait and wait then when water gets to the null point this is when the 90 degree pulse comes in this is where the spin echo sequence starts okay so as long as tissue has met and gotten to the null point we will suppress it with the correct ti now let's talk about fat sap fat sat is sometimes called chemical shift everybody when they're sitting there scanning they all say fat fat because it's easier to say than chemical shift imaging but guess what they're those two um terms are interchangeable they can throw well when you do chemical shift imaging blah blah blah chemical shift means fat sat what's fat sat fat sat will put in a series of selective rf pulses that are tuned specifically to fat so the rf pulse that gets put in excuse me is at the processional frequency of fat so fat gets excited multiple times water does not because remember they're at different processional frequencies so fat sat uses a selective rf pulse whereas inversion recovery uses a non-selective everybody gets knocked down everybody gets a participation of what everyone gets knocked down with stir and flare the reason why we can do fat sat or specifically do fat sac is because of something called chemical shift now you know there's a chemical shift artifact which i'm going to talk about but what happens is when you put the patient in the magnet fat and water go to their own little selective processional frequencies fat is a fast t1 relaxing tissue but it's slow it's slower than water because it's a big molecule it doesn't rotate as fast so when i say selective i mean with we're specifically targeting fat with a rf frequency that's at the law more frequency of fat down below that's inversion recovery because you can see that it says it says ti right here okay why did i call the bottom sequence inversion recovery and not stir and flare because there's no ti stated so if you saw this diagram right here and they said what kind of sequence is that you're going to know that it's inversion recovery okay so why does chemical shift imaging work well like i just said we're taking advantage of the chemical shift or processional frequency difference between two chemicals fat and water let's just call them chemicals okay it's nice and easy so processional frequency is the processional frequencies of tissues is directly proportional to magnetic field strength so recall the law more frequency from a couple hours ago again you're going to need to know that equation whoa boy nice and easy whoa boy wobble is equal to b0 times i believe that's called lambda and it's the number is 42.57 i bet you any money they're going to ask you that but anyway processional frequency of processional frequency of hydrogen at one tesla is 42.57 at 1.5 63 and at 3 it's 127. so as water speeds up so does fat but just not at the same rate they can ask you this question right over here fat precesses what is the processional frequency difference between fat and water very open-ended very generic question well if they ask you that you just say it's 3.5 parts per million because note that i didn't state a field strength 3.5 parts per million is the difference between fat and water as and you've got to remember that but as soon as they throw in a field strength that's when you need to start doing some math okay so no field strength noted 3.5 parts per million if they stayed a field strength that's when things change so let's let's put this um into a little bit of perspective so if you put the human body into a magnet and you speed water up to a million hertz one million revolutions a minute fat is going to follow suit but it's going to be a little bit slower just ever so slightly slower so look 999 9997.5 if you add 3.5 to that you're going to get 1 million now speed up increase that field strength a little bit and speed water up to 2 million well fine fat is going to go 1 million 999 nine hundred 900 ninety three that's seven revolutions less or three point five times two do you see a pattern developing okay so then if we say all right well we're gonna put this patient into a 1.5 tesla magnet well water speeds up to 63.85 million hertz cool multiply that by 3.5 parts per million and you're going to come out with 223 223.47 hertz just call it 223. the millions on either side of that equation cancel out so the processional frequency difference between fat and water is 223. that's by my math right here most people most texts that have been around for a while are going to say at 220 close enough yeah that's fine 220. what if the registry asks you that question and they don't say 223 they say 220. that's the next closest answer give it to them tell them what they want to hear okay the answers the or the um variables the potential answers that they give you may not be exactly what's in your head you might say well steve powers said 223 but the closest answer they gave you was 220 yeah fine close enough give it to him tell them what they want to hear okay so for all you ge users out there i took this right off a ge screen and here's your chemical shift you got two chemicals fat and water fat with the blue arrow excuse me fat is the yellow arrow and water is the blue arrow that's the processional frequency difference at 1.5 because fat is a big slow molecule water is a small fast molecule if that was done at a three tesla scanner that those peaks would get even further apart because processional frequency is directly proportional to field strength we can do fat sap because now we've separated the processional frequency sufficiently enough we can put in an rf pulse that targets fat and fat only with a selective rf pulse okay that's how we do fat sat that's chemical shift imaging not chemical shift artifact but chemical shift imaging okay so let's move on to time of flight mra time of flight another nice word for is flow related enhancement those two words those two terms are equally um interchangeable so what time of flight means is it's the time that blood flies into the slice before it flies out so we're imaging extremely quickly we're getting signal from blood before it goes away okay flow related enhancement time of flight same thing excuse me i'm getting a little bit dry here guys so with flow related enhancement we're going to be using trs and tes that are shorter than the t1 and t2 times of the tissues so basically we're going to be setting up a steady state in the stationary tissue keyword there was stationary so the stationary tissue is going to be basically suppressed it's going to be saturated with rf it's not going to give us hardly any signal because we're not giving it time to t1 relax due to the flow we've got fresh spins that come into that slice they get excited with the rf and they have a short te so we can grab their signal before they go away now i couldn't separate all these little jpegs but we're imaging that slice we're exciting the blood we're collecting the signal from the blood before it goes away the background tissue we don't care about we want it to be dark we want to be saturated so you get bright vessels dark background that's what the very short tr and te do for us okay so again the tr and te that we use a shorter than the t1 relaxation time of stationary tissues we don't see tr's and tes this shot with spin echo do we you're not getting 15 to 20 millisecond tr's in spin echo land you're not getting five to seven millisecond tes in spin echo land because you just can't because of the presence of that 180 that 180 burns a lot of time so with those really short trs and tes the stationary tissue is basically stuck at the flip angle it's in a steady state and the signal is going to be low it's going to be dark background tissue bright vessels what's happening is the blood is flowing in and flowing out notice these these fresh spins are coming in and out in the slice acquire the signal move away whereas the imaging slice is going to be dark okay so registries like to alternate these names i already said that to you flow related enhancement equals time of flight i really like this um this slide that i've got here because it shows the spins very nicely so you got the imaging slice that's being hit constantly with rf and blood flows in gives it signal goes away new blood comes in gives it signal goes away and on and on okay then when we're done with that slice we move on to the next one so you have dark background tissue bright blood vessels now you we always have biphasic flow what do i mean by that well there's arterial flow and there's venous flow how are we going to make one side of that flow dark so we want to do an arteriogram we're doing that nice and easy just think carotids you guys have probably done hundreds of carotids blood flow is going north it's going into the brain so from the heart up well at the same time we've got venous flow coming down right that's the way it is that's the way it works babies crying mris make noise and blood flows in two directions so how are we going to make this only one-sided hopefully you just say to yourself wait a minute i know what i can do i can add a sat pulse that's right so sap pulse is going to be placed in the on the side of the blood you don't want to see so in this particular case we're doing carotids we're doing axials in the carotids blood is going up the neck into the brain venous blood is coming down through the sap pulse so it's getting hit with an rf before it even gets into the slice and then the slice gives it a double whammy and it gets another rf pulse and it turns dark because now it's saturated even though it's flowing it got pre-excited that's why sometimes the sap pulses are called presats so you're going to put the rf pulse on the side of the blood you don't want to see that's nice and easy in the carotids okay but they like to ask where should you put the sap pulse for doing whatever venus or arterial flow they like to ask this where should the sap bolts be placed when image veins below the heart uh oh hold on wait a minute veins below the heart all right so nice and easy just think nice and easy think in the popliteal think the knee so venous blood is coming up the leg towards the heart but you want to see the arteries in the knee where do you put the sap pulse hopefully you said below your slices because you want to take out venous blood they like to ask this kind of question okay right here okay so there's the heart beating away arterial blood is going to the head arterial blood is going to the feet you want to do a venous sap to show the artery so just got to think logically guys which way is my blood flowing okay i bet you any money they're gonna they're gonna ask this question because they did on my registry and that was a long time ago okay all right now dwi diffusion weighted imaging it's it's done basically on every brain that you ever do it's an indispensable tool looking for strokes and it's getting into um being used for tumors as well it's very useful in detecting ischemic areas in the brain because what we're going to be doing is we're going to be able to see differences in contrast because of the diffusion in the brain tissue the brain doesn't have an outright blood supply everything all the blood that it gets or all the nutrients and water and other such things it gets through diffusion blood cell blood brain cells do not have their an outright blood supply okay so just like in t1 weighted images where we get t1 contrast from t1 differences in the brain dw images we get our contrast when tissues are having differences in diffusion so dwi exploits random molecular motion called brownian or sometimes called the sodium pump different tissues if they have their blood supply or they don't have their blood supply they're going to have different amounts of diffusion going on the extent of capillary tissue movement of stuff in the cell okay is going to depend on whether the cell membranes is intact and if there's a blood supply so brownian motion maybe that's a new term for you some people call it the sodium pump but brownie in motion if you look it up in webster's is a random motion of molecules in a medium liquid or a gas the pattern of motion usually consists of random fluctuations in a particle's position within the fluid so basically it's going wherever it wants okay because there's a blood supply to the brain cells it can move stuff the brain cells can move stuff in and out at will but when there's no blood supply you have something called restricted diffusion now here comes the the gradient part of it when a gradient is applied and this goes for all mri when you turn on a gradient stationary tissue is going to pick up phase all right it always does but moving tissue for example stuff that's moving in and out of the cells is going to be when it's moving it's going to pick up even more phase so when in pots phase to everything just some get more than others and that's the moving that's the brownian motion within the cells so picking up phase because they're moving is going to mean a loss of signal that's a rule an mri when you turn on a gradient things get dephased and they lose signal so stroking tissue doesn't have a blood supply it's not diffusion there's no brownian motion and when you turn on those gradients there's no de-phasing happening in those cells so they are going to not lose signal they're going to stay bright the unhealthy tissue does not have stuff moving it's going to stay bright unaffected healthy tissue with the blood supply it's moving stuff it loses phase so you're going to have dark tissue next to bright tissue and that's contrast and that's diffusion contrast because you've getting different amounts of signal from different tissues what does it look like what's a dwi look like well right there you go oh that's a spin echo yep you're right it's a spin echo 90 180 if you said that excellent double thumbs up but what about those extra gradients on either side of the 180 heck are those those are the diffusion weighted gradients that are being turned on in order to make tissue either pick up phase the good stuff and the stuff that's not moving is not going to pick up face so something's going to be bright something's going to be dark and then at the end where you see all those echoes you'll notice that there's a bunch of gradient reversals that's that's textbook what a epi sequence looks like echo planar imaging so we're going to excite we're going to phase encode we're going to dwi it and then we're going to turn on that frequency encoding gradient dozens and dozens of times very quickly and get echoes from each reversal so that's a spin echo dwi you see those extra gradients that i've got circled in green so what's this diffusion thing well i've already gone through most of it okay cells want to be nice and how nice and healthy they want to be able to move in what they need and move out that they what they don't so they want to bring in nutrients and they want to remove waste on a cellular level so stuff's moving when stuff moves we turn on a gradient they get dephased they lose signal sodium pump stuff is not working when it loses its blood supply so you get restricted diffusion turn on a gradient stuff that's not moving doesn't pick up phase now what do i mean by phase so here's a real quick down and dirty dwi graph okay the dg or diffusion weighted gradients on either side of the 180 and there's usually three directions so we are going to dwi it in the slice direction we're going to dwi it in the phase we're going to dwi in the frequency direction too you always do three directions just so you know there's three pairs of dw gradients but let's just keep it simple with one okay so you turn on that first diffusion weighted gradient right here so you have some protons that were that are in phase then you turn on the gradient and they dephase so their vectors say go down this way well the green is moving that's going to go down even further it's going to pick up more phase down here down the bottom the blue tissue is not moving that much it's restricted so it doesn't de-phase as much okay and then you turn on another gradient and you're going to refocus them and they're gonna come above this line and then when you dwi them again they're gonna go back to where they were but one is got a lot more phase okay that is the restricted diffusion that's the blue that's the water so you de-phase it and you you de-phase it on purpose then you rephase it and you bring it back in and at t e something's going to be bright and something's going to be dark whether it's moving or not moving that's your dw contrast so you're going to get low signal from normal tissue you're going to get bright signal from abnormal tissue you've all seen those dwis where someone's had a stroke here's another down and dirty dwi pulse sequence you prob i don't know how many of these you've seen but it's a spin echo based dwi how do i know it's dwi there's extra gradients right here this happens to be from the x gradient extra dw gradients that aren't you're normally there so this is a dwi sequence and it's epi because we're doing all these gradient reversals causing echoes okay now what's this b0 thing that i've said 18 times the b value measures the degree of diffusion weighting being applied it's an indication of the amplitude of your dwi gradients okay how strong these guys are how long are they applied for okay because time runs from your left to the right how long do we turn them on for and what's the time duration between the two of them a larger b value is going to increase have increased amplitudes be applied longer and a wider time so all factors remaining the same it's combined to all factors all these factors combine to diffuse the tissues more or less a high b value is more sensitive to a stroke than a low b value typical b value is around a thousand to fifteen hundred will they ask you that probably not but they might ask you is a 500 b value more sensitive to a stroke than a 1000 i think that's a fair question so the 1000 is more sensitive to a stroke so now so let's um let's identify some pulse sequences here okay nope all right so top left what do you see well you see a 180 so you go 90 180 that's spin echo absolutely what about top right you see a 180 90 180 what do you think that is hopefully you're saying it's inversion recovery you don't know if it's stir a flare but it's inversion recovery what about bottom left gradient echo why is it gradient echo limit there's a 90 there isn't there well don't let that confuse you that that um arrow in the line that means less than or equal to 90. that means it's a gradient echo there's no 180. and then over on the bottom right that's a fast spin echo so you might see these so you need to be familiar with your uh identifying your pulse sequency pulse sequence line graphs over on the left hand side what kind of graph is that hopefully you guys said t2 in the middle what kind of relaxation curve is that t1 and last but not least over here on the right hand side what kind of pulse sequence is that well you might you might have said it's uh epi which you would be correct because you got all these frequency encoding gradient oscillations but you see these two guys right here that makes it dwi so it's a dwi epi okay so let's do some scan time formulas uh a lot of the texts that i know are kind of weak at this and this is something i want you to practice you need to do this multiple times before you go sit for your registry so the basic line diagram uh excuse me the basic um scan time formula is tr times phase times next and that's going to give you a great big number in milliseconds great easy math tr times phase times next that big number in milliseconds you need to convert to minutes and seconds so then you're going to divide those milliseconds by 60 000. 60 000 is the number of milliseconds in one minute so tr times phase times next equals your scan time that you need to divide by 60 000 to give you minutes now here's something that i want to say before we continue sometimes they will say what is your scan time in milliseconds easy peasy tr times face times next you don't divide by 60 000. they didn't ask you for it minutes and seconds they ask you for a scan time in milliseconds okay slow down read the question what are they asking me they didn't ask me to divide it and make it into minutes and seconds they just ask for the scan time in milliseconds so you're nervous you're stressing oh my god i don't slow down read the question what are they asking me they're going to give you the answer you just need to figure it out because they're going to give you four possible answers one or two are absolutely wrong and the other two are probably pretty close you've got to get your the best answer okay so that being said 500 times 256 times 2 equals 256 000 milliseconds all right if that's all they ask you for then that's what you're going to give them but if they say what is the scan time in minutes well now divide the 256k by 60k and that's going to come out to 4.26 minutes is that 4 minutes and 26 seconds no you got to do one more step so 60 is how many seconds you're going to multiply that by 0.26 that comes out to 15.6 so the scan time here is 4 minutes and 16 seconds pretty easy math but you got to know what they're asking you okay you need to practice that like i said okay here's one more so let's throw a t2 at you 3456 tr times 192 times one 663 thousand five five two milliseconds divide it by 60 000 and that comes out to 1105 60 times .05 is three seconds your scan time here is 11 minutes and three seconds pretty easy math once you practice it a few times okay change of change of pace here okay um okay so allison said okay would it say 192 by 256 um do you only worry about the 192 yes because phase 99.9 of the time the phase number is going to be lower than your frequency number it's easy if they said 192 by 192. you don't have to worry but you're right 192 would the lower number is going to be your phase steps and if you've got the time and you've only got a limited amount of time here guys um be careful maybe if you've got time do it both ways use the free if you're not sure which one is which okay if you're not sure which one is phase and which one is frequency do the math twice and if your numbers are completely out of control you know that you did the frequency wrong okay yeah that's yeah that's what you were wondering so 99.999 percent of the time phase is going to be less than frequency if they're nice guys they'll tell you which one is which okay but you can't always count on them being nice so you do this math you come up with um seventy three thousand seven hundred and twenty seven okay that sixty thousand where is that coming from well when you do when you multiply milliseconds by a number by a number you come up with a scan time in milliseconds okay now we need to make this milliseconds into minutes 60 000 is the number of milliseconds in one minute so if you've got a great big time in milliseconds that you need to make into minutes you have to divide by 60 000. it's the number of milliseconds in one minute so yeah there's a bunch of little math here but you've got to know the processes there's only three basic scan time formulas we just did regular old let's just say spin echo now we're going to do fast spin echo same basic formula tr times phase times next okay all right so tr times phase times next is going to give you your scan time in milliseconds now fast spin echo is the key term here you need to divide your time in milliseconds by your echo train okay so third was that 13 271. divide that by your echo train length 18 22 6 whatever the echo train they're going to have to state the echo train divide that by 60 000 and you're going to come up with your time in minutes and seconds okay you need to try this guys you need to practice this don't just say yeah i got this no don't do that to yourselves practice it okay all right so that's so that's your fast spin echo scan time formula tr times phase times next equals milliseconds divide that by your echo train divide that by 60 000 and you're going to come up with your scan time in minutes okay all right now i i just skipped that one real quick for time all right now 2d or 3d sequential sequences that's your that's the other white meat okay so they can say in this 3d time-of-flight mra and you're saying okay no problem what lay the factors on me tr times your phase times your next now you're going to multiply it by your partitions some places call it locations okay they're going to tell you that 3d or 2d they have to tell you the number of slices because that's going to change your scan time so tr times phase times next time 64 is your scan time okay then divide it by 60 000 just like you've done three or four or five times whatever how many times we've done it and that's going to give you your scan time in minutes and seconds again it's easy math but you've got to practice this stuff okay because practice makes perfect okay any questions guys we need to forge on here um yeah i've still got a little bit of time all right so we're going to for john so i'm going to stop sharing and i am going to go to my other powerpoint okay good we still got it there all right okay artifacts there we go all right share my screen and artifacts there it is all right guys we're rounding third heading for home i know you're tired guess what so am i okay so don't forget lay the questions on me if you've got them so i'm not going to do every artifact known to man i'm going to hit the biggies okay so what are what's an artifact if they ask you what's an artifact it's false information it's something that's not really there or it's something that shouldn't be there there are three basic kinds of artifacts or three sources of artifacts physiological which is patient moving inherent which is the physics and then there's hardware which is coils and other electronics i'm not going to break down each one of these but i'm just going to run the list so what you really need to do guys is you need to have maybe a little a little list and you need to know which artifact goes in which direction for example so if you're looking for you know if you've got an artifact and you're not sure what it is and you're saying well wait a minute that's in the phase direction well guess what if it's in the phase direction it's not chemical shift or it's not an effect so it's a process of elimination so know the directions of your different artifacts which way they go there's a couple that go in both directions and then you know okay it just helps you with the process of elimination okay so we've all seen images like this okay motion motion is always in the phase direction and it's considered a data sampling artifact you can get it from obviously from patient breathing which is what that one is you can get it from pulsatile flow or peristalsis again it's all concerning the patient's moving okay so here's it so they might just show you an image like this and say what artifact is demonstrated here okay either one of those well i'm hopefully you just went oh yeah that's motion yeah because it's um an evenly spaced rhythmic kind of motion in the phase direction they won't probably show you which way phase is running but you can tell damn that's motion what happens is or why you get motion only in the phase direction is because there's a lag time from the time that you put a phase encoding gradient on and the time you acquire the echo because remember i said the phase encoded gradients applied for the future echo remember i said that i don't know a long time ago a lot of words ago okay what happens is you phase encode the tissue and it's at place a great okay 10 12 20 milliseconds later you acquire that echo that tissue's not there anymore it's in place b and there's this smearing so it kind of brought its phase with it it smeared it across the bagel as it were you know you smear the cream cheese you smear the motion in the phase direction because of that time lag between phase encoding and measuring that echo so they might ask you what are some of the ways to decrease motion well you can't scan dead people even though you might see dead people you can't scan dead people insurance doesn't pay for that well you want to make sure they're comfortable coach them up hey listen i'm going to need you to hold your breath when i say it or listen it's going to be loud don't let that bother you you need to immobilize them pat them strap them keep them still scan fast and you can also use motion suppression techniques so here's an example of hey what artifact is demonstrated here well you can see my red arrows it's pulsatile flow because the blood was moving through the gradient it picked up phase and it kind of dropped it off in the phase direction okay it's evenly spaced repetitively um repetitively spaced artifact that's classic pulsatile that happens to be stir and why is it why is it so bad on stir you guys have probably noticed man i always see this motion on stir always well first of all inversion recovery is just terrible for motion just because and in general and also long tr and long te sequences are more prone to motion they're also more prone to pulsatile so a lot of these t well a lot of your stirs and flares and things like that you should have flow comp on to help make that go away a little bit so the image on the right no flowcom real bad pulsatile flow the image on the left well flow comp is on it's a little bit better what's flow comp flow comp sometimes sometimes it's called gradient motion nulling is a way to lessen the amount of phase shifts that are measured at t e so it uses a pair uh paired bipolar gradients after excitation but before signal readout so what what's the consequence of turning on flow comp all factors remaining the same flow comp extra gradients steve just said that extra gradients means more work right it's going to bump out your minimum te so it's a pair of unequal and opposite polarity gradients that are applied and it compensates for phase shifts induced by blood flowing through the slice select gradient kind of sounds like dwi doesn't it you turn on a gradient and you're gonna make phase shifts happen so what we're going to do is we're going to if you turn on one gradient and it causes phase shifts if you turn on a pair of gradients after the fact you can correct for those phase shifts so that's exactly what we did here it's not perfect but it works a little bit better so here's what it looks like for example in a quickie little line diagram so you've got blood flowing okay you've got stationary tissue which is blue you've got moving tissue the blood which is red you turn on this gradient well they were in phase but guess what after the gradient now they're out of phase ding ding ding oh i'm gonna get that pulsatile flow okay i don't want the pulsatile flow steve just turned on flow comp okay so you're going to turn on bipolar opposite polarity gradients so that this guy is this one here is going to over correct and then the plus one is going to recorrect and at t e those those protons are going to be in phase if you do the math you got a plus one a minus two and a plus one you do that math it comes out to zero so you don't have theoretically you have no phase shifts at t e extra work it lengthens your minimum te okay what's another common artifact okay this happens at my place all the time and it's not that we have a bad faraday cage it's just that the techs at my place think nothing of just blasting into the into my scan room while i'm scanning that's a pet peeve of mine but anyway so all mr rooms just like x-ray rooms they're shielded they're rf shielded it's not lead it's brass or copper and it keeps rrf in and it keeps somebody else's rf out when other people's rf gets in sometimes our coils pick it up and over on the right hand side you can see that zipper kind of artifact and that's rf that's classic rf sometimes it's just a single like down the bottom sometimes it's multiple like up the top if there's multiple it's a wide bandwidth of rf that got into your room notice its position now granted phase and frequency is swapped on this but however rf or zipper artifact is it's in the phase direction but its location up or down or in this case right or left it comes from what the frequency was so it's not always in the exact same spot so it's an rf artifact that's seen in the phase direction guys it's one of those that's one of the only ones that's kind of opposite you go oh it's an rf why is it in the frequency direction because it's not its location is frequency dependent in the phase direction a leasing wrap or fold over um don't know which term they're going to need excuse me which term they're going to use okay um so unfortunately you have to know all three of these um i'm trying to i there's no way i remember what they used on mine but i'm hearing that they use aliasing but who knows know all three of these so rap comes from having tissue outside the field of view in the phase direction so the scanner will fully sample the tissue inside the field of view you put this you you put your little box up there over your brain and it's going to sample everything in that field of view it's going to sample it very nice everything gets an rf even the stuff that might be outside the field of view even though you don't want it to sorry we're doing a square field of view and rf is round okay just because it's outside the field of view it's going to get an rf pulse so this tissue outside the field of view is going to get under sampled you don't look at it enough to know exactly where it should go under sampling it or not sampling it enough does not satisfy something called the nyquist theorem the nyquist theorem says you need to sample an analog wave at twice its frequency to get an accurate picture of its true form okay so you see that little sine wave that's moving from left to right if i sample it enough and then connect those dots yeah i got a pretty accurate depiction of what it looks like so i know where it should go in k space it's pretty easy okay sampling it at twice it's the minimum of twice here i sampled it five times so i really got a good a good picture of it so the nyquist theorem is satisfied there the scanner knows where to put that echo now if i don't sample it enough same same sine wave i don't sample it enough and then i connect those dots that high frequency sine wave looks like a low frequency sine wave and the scanner is going to say yep okay it goes over here it doesn't really belong over there wherever there is okay so it's not you didn't satisfy the nyquist theorem you didn't sample enough this is what happens when your tissue is outside the field of view and doesn't get sampled actually it doesn't get sampled at all because you're only telling the scanner just look at my field of view there's stuff outside it get excited it's going to give signal whether you want it to or not okay so the nyquist theorem is not satisfied this right here is going to give you rap you've seen rap a million times it's one of the most common thing they might just show you a picture and say what artifact is demonstrated here okay now the nose over on the the bottom right you know that's pretty obvious okay the one at the top ah yeah that's pretty obvious too the shoulders are wrapping in and the lower leg that's you know once you get it pointed out to you oh yeah that's rap that's wrapped from the uh lateral side of the calf getting wrapped in so again that might just show you pictures and say what artifact okay wrap 3d sequences can wrap wrap is not only um limited to the phase direction you can get rap in the frequent uh in the slice select direction you get that in 3d slices 3d um data sets so that's called slice wrap so the top image here you can see very nicely the foramen magnum that i've got circled in red but what's all that other stuff on the outside that's the top of the head so you got the bottom wrapping into the top and when you went to the top you could you know you get wraps so the top wraps into the bottom the bottom wraps into the top and that's called slice wrap and once again that's because tissue is outside the field of view and it gets mismapped and gets put in the wrong spot chemical shift artifact now we're not talking chemical shift fat sat we're talking chemical shift artifact now what happens is you got a difference in processional frequencies chemical shift means a difference in processional frequencies between fat and water so we took advantage of that to do fat sap we've already been through that but we also are a victim of chemical shift when you've got differences in processional frequencies between fat and water so one time we take advantage of it for good another time we're a victim and unfortunately it can be kind of a bad thing so chemical shift three factors that affect chemical shift or the severity of chemical shift is your field strength are you at one tesla versus three higher fields you get worse chemical shift artifact because processional frequency is proportionally is directly proportional to field strength so those two chemicals start to process at higher and higher different frequencies so far so good so field strength is a major contributor to chemical shift artifact your field of view if you have a small field of view the artifact is going to get worse because what you're doing is you're trying to put more frequencies into a pixel a pixel can only hold so much or so many frequencies okay so it's the you know it's the five pound ten pound bag thing you can't put ten pounds of potatoes into a five pound bag okay so what happens is you can only put ten pounds of signal into a five pound pixel what happens is it oozes out and because you're trying to put more frequencies in there it oozes out in the frequency direction so chemical shift is always in the frequency direction another big contributor to receiver is receiver bandwidth so receive a bandwidth i said this earlier it maps how many frequencies are put across your field of view so if you try to put too many frequencies in your field of view or into each pixel you're going to get you're going to get chemical shift artifact okay so each manufacturer and this is where i was going with this earlier as far as receiver bandwidth every manufacturer uses a different term to describe their receiver bandwidth um ge uses kilohertz siemens uses hertz per pixel and phillips uses fat water shift they're all they're all the same but they're different okay so in order to make that easy i would hope and think that the registry is going to use wide and narrow and i believe that's what they're doing so anyway so we've got fat and water peaks here okay that's the chemical shift okay so water is at your center frequency and fat is halfway to the minus over on the right hand side as field strength increases these two peaks get further apart well if you're trying to fit all those frequencies into one pixel it may not fit so when we're doing fat sat we're exploiting or taking advantage of that fat of that chemical shift but in chemical shift artifact when we're not doing fat sat we're going to be a victim of that chemical shift and just so you know there's two kinds of chemical shifts chemical shift of the first kind and second kind like that last line says i'm gonna i'm gonna cover both of those okay so here's chemical shift of the first kind we're trying to put so that middle pixel that that black dashed box that's your pixel we're trying to put too many frequencies into it what's going to happen well it's going to ooze out into the next pixel so then you're going to start getting fat and water cancelling each other out so on that kidney where there's a fat water interface okay the the red the red lines are pointing to the artifact you see this chemical shift at fat water interfaces so the kidney is a lot of water and it's surrounded by fat so it's classic to see it around the kidneys so you get a black line on one side of your image and a white line on the other that's classic chemical shift because you're trying to put too many frequencies in one pixel maybe the way to fix that is going to be to widen you receive a bandwidth and that will lessen the artifact okay so here's the kidney again exact same picture that you just saw okay the artifact is seen at fat water interfaces okay right here black line white line you also see it it's kind of subtle but i bet you any money you don't even notice it on your spines okay here is a spine at 3t you can see the black line here and the white line over here that's classic chemical shift and it's seen in the frequency direction they might say to you what artifact is demonstrated here you know question whatever oh that's chemical shift yep good two points how do you fix that artifact uh-oh the easiest quickest way to do it is to widen you receive a bandwidth use a smaller field of view or go to a 1.5 tesla scanner but that's not practical if you saw that artifact you're not going to say i'm going to get them off my 3t and put them on the one fight that don't happen artifact is in the frequency direction at fat and water interfaces you can fix it with receiver bandwidth that's the quick cheap and easy way to do it okay now chemical shift of the first kind up top you only see chemical shift on spin echo and fast spin echo sequences okay you can only see it on those two okay that's chemical shift of the first kind it's because you've got frequency differences not phase differences chemical shift of the second kind is down below hopefully you looked at that you went oh that's an out of phase te look at the black line around everything ding ding ding ding ding i owe you a cup of coffee that's exactly right that's chemical shift of the second kind also known as out of phase te what's happening is at that specific te in that gradient echo sequence you've got fat pointing let's just say right and water pointing left they cancel each other out it's out of phase or phase cancellation artifact couple of names for chemical shift of the second kind out of phase or phase cancellation you only see that out of phase te on gradient echoes you never see it anywhere on anything else ever ever and forever why let you stew on that for a second while i drink my tea all right you never see the phase the phase cancellation or out of phase t e on on spin echo because the 180 degree refocusing pulse puts everything in phase at t e so even if you pick the right te you're not they're not going to be out of phase the 180 degrees doesn't let that happen whereas gradient echo without a 180 it can happen if you pick the te correctly or if the te is right okay all right so that's chemical shift of the second kind also called out of phase or phase cancellation artifact crosstalk and cross-excitation now almost everybody uses these two terms interchangeably sorry i don't maybe i'm being pedantic but i don't teach them that way so these two terms like i said are interchanged almost everybody says crosstalk i don't teach it that way right so here's crosstalk crosstalk comes when there's rf in common between slices so on the left figure 1220 okay slice one and slice two have a little bit of rf in common okay look at those numbers okay so 64 to 66 and slice one and then slice two is 66 to 68. well what's in common 66. so when you excite slice one you're you're exciting just a little bit of slice two so the overall signal to noise of slice 2 is going to be just a little bit lower is it oh my god look at that crosstalk no but it happens look at slice number three what's in common the 68 so when you excite number two you're also exciting just a little sliver of number three because we use we think that our are that our slices are square and we'd like them to be and that's what we think but guess what we can't get we're not going to get something square out of round rf so notice the blue rf it's round we get we want a square slice and we almost kind of sort of get it but we don't get it particular you know exactly we have something called full width at half maximum this is where we get our slice thickness so that big dome that big pyramid that big pointy thing that's your rf pulse halfway up and half and full width halfway up that's our slice thickness notice that the blue area over here this is crosstalk so we are exciting slice one but a little of one's rf gets into slice number two because rf is not square that's why we get crosstalk so your signal to noise drops a little bit in slice number two okay how do we avoid that can we make that go away yes we can we can we can minimize that a little bit we can put sorry guys come on work with me here come on computer i know you're tired why is it doing this okay we can put a gap in between slice one and two there's already a gap there that's that's how it shows but we can gap or we can interleave would be a way to get rid of crosstalk skip some tissue in between slice one and two and not let the overlap of one get into two move two over a little bit okay now cross-excitation you guys have seen this artifact before if you've done a lumbar spine you've seen this artifact before everyone calls it crosstalk and that's fine but i call it cross-excitation because tissue from the yellow slab is common to the next slab above it okay so here's l5 s1 here's l45 this tissue in common here is what's getting cross-excited so when you excited l5 s1 and then moved up to l45 this tissue hasn't had time to t1 relax so it's getting a second hit of rf and it's going to be dark okay when you do these multi-slice multi-angles you're going to get that especially at l5 s1 and l45 it's not uncommon and it's worse when you do t1 weighted because t1 has a short tr so that tissue doesn't have a chance to t1 relax so if they showed you this image in 1224 and they said what artifact is demonstrated there and they said to you cross-excitation cross-excitation rap motion and chem shift cross excitation is the answer you're going to pick if they said cross talk you're going to give them crosstalk again tell them what they want to hear gradient walk is another artifact that comes when you do great big fields of you i'm sure you guys have seen this one as well the reason being is because gradients have to have three qualities they need to be linear constant and reproducible the problem with gradient warp is the frequency in the the frequency field of view is getting too big for the for the gradient to do the uh to be applied linearly linear means straight or no bending and no curving so when you turn on a gradient we assume it's perfectly straight over a distance it's constant meaning when you turn it on it stays on and doesn't like falter in strength and reproducible every time you turn it on it comes back on the same strength so linearity is what hap you're losing linearity when you do gradient when you get gradient warp okay so here's a depiction of gradient walk so you turn on the field gradient you turn on the frequency field of view and it's great for the vast majority of it but at the ends circled here in purple what's happening is the gradient just can't do it anymore you're asking too much of the gradient and the gradient starts to fall back to zero so you're losing linearity and the processional frequencies in this purple area here is starting to is starting to alter we assume that our gradient when we turn it on is this dashed line that would be perfect it's it's infinite but they're not they're only good for a certain distance when you use a big field of view that's when it's going to happen like sagittal t-spines sometimes in the abdomen long bone studies okay so here's that same image that you just saw in the middle here is a great big field of view of a great big person and these are t2 fat sac coronals or they stir i don't remember but anyway i put that circle there because ideally when we are imaging we want to be we want to make um keep our field of view within something about the size of an oversized beach ball you don't want to be imaging way off close to the bore top of the bore or the sides you want to keep it in the middle so when the field of view is too big you start to lose linearity on the edges that's what this artifact is way out here sagittal this is where you see it very often in your t-spines when you do the counting field of view do a great big field of view to count and you start to lose linearity the field of view is just too big for what the gradient can do so this white dashed line is what you want to cover but it really can't do it so you lose or warp the field at the ends when you walk the processional frequency when you walk that gradient the processional frequencies start to fail and alter and your image starts to it looks like something from outer space okay i know you guys are getting tired we're we're rounding third heading for home magnetic susceptibility that is the ability of a material to be magnetized i went over this a little bit when i started talking about gradient echoes so top left the b0 is perfect everybody's nice nice there's no there's no little waivers the the gradient field is is very homogeneous okay the middle where you see diamagnetic what's happening is diamagnetic like water lightly or very slightly repels the field okay that's that's fine but then you've got paramagnetic and ferrous materials that will bring the field into itself so you've got someone pushing it away and you get somebody pulling it in when those two guys are right next to each other you get those little black holes like you just like you saw earlier okay so you've got dire magnetic paramagnetic and ferromagnetic now when i say ferromagnetic i'm not necessarily talking about something that's going to get sucked out of your pocket when you get too close to the magnet all metals even aluminum is considered ferromagnetic it does draw the field into itself a little bit okay so diamagnetic materials intrinsically weakly repel like water air hydrogen water paramagnetic intrinsically weakly attract the field they bring it in just a little bit that's our gadolinium and then ferrous material um strongly attract they bring that that magnetic field those flux lines into each other and when you've got differences in susceptibility so diamagnetic has its own susceptibility para and ferromagnet have its own susceptibility so you've got susceptibility differences and that's when you get those little black holes like we saw in that lady's brain okay the human body is mostly the human body has air in the lungs and sinuses we're mostly water we've got magnetic susceptibilities all over the place so you've got opposing forces between dire and paramagnetic okay so that's when you get these little holes the iron in the blood in this lady's brain is warping the field a little bit it's differences in susceptibility between the brain and the iron that's in the hemoglobin that's the top image down the bottom that's a young lady who had a mouth full of braces and i i knew that this you know she said oh i got braces real big braces and you know i said let me see and i'm looking i'm going oh my god these are going to be terrible so you can't just say go away you got to try it so i ran the study we didn't give gad because they were looking for a pituitary but anyway so the braces just completely messed up the field right here and you can't image that so there's not a lot you can do except maybe get the braces removed but that's not our call so different sequences are going to react to this magnetic susceptibility worse okay these two happen to be gradient echo this is a gradient echo in the brain this is just a scout view these are the worst for seeing the magnetic um susceptibility differences because of the sequence so the way to get around fixing this susceptibility is to do something called a mars technique mars technique is short for metal artifact reduction sequence and basically you're going to use a sequence you're going to widen the receiver bandwidth that's going to help shorten your minimal te and you're going to increase your echo train length because if a gradient echo is terrible for mo for metal then because it doesn't have a 180 then a fast spin echo with multiple 180s is going to clean that up so your basic mars technique is going to help clean up the metal so over on that uh oblique coronal of the shoulder this person has had a rotator cuff repair and there's some metal in there there's not much you can do about it except to try the mars artifact okay try to stay away from gradient echo and dwis if you can and fat sats that's going to really mess up the fat sat and dwi so metal is really not much more than a bad susceptibility artifact okay how can you tell the difference between metal and just straight susceptibility here's system susceptibility in the brain top right get a visual on that okay now you go down a couple now that's susceptibility but that susceptibility is coming from metal so if they showed you this artifact would you say it's susceptibility yep would you say it's metal i would notice the little black halo here and the halo here that's classic metal artifact and that's called pile up just so you know when you're scanning and you see this right here and you've seen it worse times i'm sure that's pile up and that's in the frequency direction and that comes from metal so the way to get around that is with the mars technique the reason why we don't see good fat sacs near metal is because and these numbers are just for example here okay away from the metal everybody's processing at 63 megahertz that's great as you get close to the metal because the metal brings in oops because the metal brings in the magnetic field it strengthens it so maybe this tissue here is at 73 megahertz versus this guy that's at 63. and then real close to the metal 83 megahertz if you're at a 1.5 tesla scanner nothing is processing at 83 megahertz at least not on purpose the metal can do that so this tissue that's really close here is not being resonated by your rf pulse so metal screws up the magnetic field to a point where you can't image it it the tissue is there but it's not at this it's not at the right processional frequency so once again over on the bottom left excuse me bottom right here is a classic metal artifact you can see the pile up in the in the frequency direction and this is a an ap scout view from a cat skin of the exact same patient so here's a little uh when he was a little kid he got shot in the hip or in the pelvis by a bb gun corduroy artifact now we're getting into um artifacts that come from our equipment corduroy artifact is just nothing more than bad data points or spikes in your case base okay these spikes are external interfering signals that get placed into k space so they're bad spikes and it's going to screw up the entire image so these bad data points can come from electromagnetic spikes from gradients coil not being plugged in well power supplies are starting to fail so when you see it it looks like this okay i did this just the other day and i said oh this is a perfect example of herringbone and my anterior coil wasn't plugged in well enough so i scouted them it was terrible i went whoo okay and i went and i replugged the coil in and noticed that it's in the entire image even though it's worse in the front it's the entire image that gets affected i plugged in the coil properly better i unplugged it plug it back in and it was fine so here's some corduroy artifact notice that it fills the entire image it's not just a single band of artifact like it is in in the zipper okay rf artifact from dad from bad data points and an effect we see this occasionally usually on the big fields of you maybe you guys have seen this before yourself and it comes from nothing more than having too many coils turned on so in this particular case less is more so over on the far right there's our field of view on the lumbar spine and i purposely turned on these extra coils just for the visual so i could take the picture so you could get this artifact and a fact from having an extra coil on that's way outside your field of view so when you see something like this there's nothing you can do about it except repeat it and turn off your turn off your extra coils over sampling also can help but over sampling does what it increases time so be more mindful of what coils you have turned on again more is not always better don't just say i'm going to cover the whole damn thing no you can't do that okay so if you guys have any questions i'll take them now if you are tired of listening to me talk thank you because i'm getting tired of talking but i am certainly more than happy to answer your questions if you have any i'm going to stop sharing and if you guys have any questions lay them on me great job stephen awesome i'll uh i'll i'll take that um silence with questions as okay powers you can be quiet now oh oh oh wait a minute oh hold on yes thank you for your presentation oh okay um i can keep on talking i have more power points you just gotta let me get a fresh tea no i didn't think so all right guys well anyway thank you very much um good luck on your registry and um almost all of this material that i covered is all in either of both of my books um and i would very much like to hear from you when you take your registry and how you did um because neil and i both care a great deal on how you guys do we take great pride in what we do and we only want to try to make this better so please do the survey and uh let let me know how i did well done well done stephen um i agree with what stephen said you know we're here to make sure that you know you're sitting confidently for the registry uh you know steven and i have both been there it can be quite nerve-wracking and long number one um so it's all about sort of getting those reps understanding the full like foundation of the content and um but yeah like you said we're i'm gonna i have them scheduled to come out at 3 30 so everyone should have an email in the email like i said is a 20 question uh practice quiz you can take three attempts we have the certificate in there as well and then of course that survey the survey is important to us i know listen we've all been to restaurants and stuff it's like oh do our survey yep if you could please dress every now i feel like you never know what they feel like because it kind of tells us how you know we do so it's it's just important um especially when it's sort of a small sample size um you know i think we're going to probably move this to like quarterly sessions um and then uh you know kind of roll it out it's you know we're you know i guess it's fair to say that this is a pilot and i think that we've had some really amazing results our last class uh that you know was in our last um register review they all passed and i think that you know yes the post program is great but being able to tie it in with what professor powers brings to the table after clinical i think is vital so um we appreciate you all coming and staying attentive and the questions and more importantly thank you so much professor powers because uh you know we know it's we know it's a gauntlet and i'm sure that you know uh maybe that tea will have a little bit of extra in it tonight i just i just have one more thing one more word of advice okay you guys need to practice answering questions so if you don't have registry review books that are questions get a couple and practice answering the questions and then grade it and say oh you know something i'm okay with fast spin echo and dwi but i don't get gradient echoes okay go back and study gradient echoes work on your weaknesses not your strength but basically practice answering questions don't not answer questions and say i know stuff and go in for your registry because you're going to be unpleasantly surprised practice practice practice and with that i will stop talking thank you very much for your time guys you got and um you know for those that are pulse students you have the mock registry that's 220 questions right there uh and you get three attempts to take that as well so i you know we use it as a i guess a like a like a uh red light green light sort of mentality where if you take it we see that it's a plus to minus five percent passing rate so let's just say you score an 85 you should score somewhere between 80 and 90 but if you take it and you're in the 70s let's just say say you're at 71 or 70 you're gonna likely fall 65 to 75 which is a little risky so for those pulse students feel free to take that mock registry i think it's very important and then of course any other questions you want to kind of any other materials go for it um but that's included so you don't really purchase anything so just a reminder a reminder there all right professor powers thank you so much and um you know i guess maybe we'll we'll chat a little later and you know maybe send that video over and i'll get it going all right thank you very much everybody for your time appreciate it thanks everyone bye peace bye