hello everybody welcome to this AC Arturo's webinar part three why adopt a sea otter adds evidence behind the performance my name is dr. Jenny Hong and I'm a neuroradiology searching University well this is our final webinar we've come with a box of tissues each and this follows a webinar first on how we do it so that was the hell of a sea otter roads then we had webinar commonly asked questions and that was really more of the hell but also the what the where the when and even the who of a sea otter roads and now with this final webinar we're going to be talking about why why should you consider adopting a sea otter ads and continue the use or use it we are going to be reviewing some key articles here and we're going to help you interpret them and what we want you to get out of it is really for you to understand why you should be spending the extra time to change your practice and use a sea otter ads and if it is time well spent I'm going to be introducing to you - Adele other speakers we've got Franklin sorry Franklin Tesla who's the professor of radiology at the University of Alabama in Birmingham and he was the chair of the ACR tirades steering committee he continues to lead efforts on the international front and he'll tell you more about that later on we've also got dr. bill Middleton who's a professor of radiology at another Institute of radiology and he was a key member of the ACR Tyrod steering committee and it was really his data that helped us make decisions on the ACR thyroids model and then there's me I'm passionate about improving patient care directly day to day but also through research and education and I've really enjoyed being a voice on these ACR tie-red committees all three of us now that we're on our final webinar I'll have to tell you a really very passionate and generally care about everyone adopting better practice for thyroid nodules and we want you to be better with adopting AC our odds if you choose to do so and this is why we've diverted our time to putting all this together for everybody and answering questions when we get them now one might think that a webinar on the evidence might be somewhat dry but we have thought really hard about this format and rather than presenting article out of the article we've divided up the literature into just simple questions that you may wonder so I'm going to tackle the first question of wide adopt a sea otter ads then we'll have Franklin who will be talking about a sea otter it's features that are particularly challenging he'll also talk about the differences between a sea otter rose versus a TA and SIU and you'll also talked about the besides thresholds and the rationale behind that so a sea otter ads and why it's different to the others he'll touch on the question of why a sea otter it's an 80a code or could not be used simultaneously and then after that bill has a very big task of giving you some review of comparing a sea otter eyes to other rest ratifications and that's through the work that he's done as well as some literature that's come out recently and then finally we'll talk about future plans for a sea otter ads I hope that you can stay for the full hour and thank you for attending everybody this session will be recorded and made available any time afterwards as were webinars 1 & 2 these are available on the ACR tyrunt's website so first we have questions about you who are you are you our radiologists a radiology trainee physician sonographer patient or other and you can answer the questions on the GoToWebinar panel and it seems that most people are radiologists but thank you to the trainees and physicians and sonographer that have channeled in the next question is does your practice currently use ACR Tyra's yes no but we'll be in the next six months and no in the near future no not in the near future and it seems that the majority of people will be using ACR towards all have been using ACR turrets that's great and for the others maybe Wilkin can convince you okay so my section is why adopt a sea otter ads and let me start off with a sample report that I received from a colleague of a mine who was an interventional radiologist this is his sister's radiology report for her thyroid ultrasound and she's a attorney who had no palpable nodule I don't remember why she got thyroid ultrasound but it led to an ultrasound report that had this that was rather short it had a section on the technique at a section two paragraphs on an overview of the thyroid and then two nodules described in this report in two lines there was a nodule in the right lower lobe with size characteristic of seven by four millimeters and then one statement about the nodule being hypoechoic and that was the only way the nodule was characterized and then another nodule within the lower lobe but this side is not mentioned here and it is four by three millimeters and once again the echogenicity is characterized and that was all and then in the impression we have several several typos about the one centimeter nodule and then of the other four millimeter nodule and then the recommendation here was really an option it was do a six-month follow-up which is really short from our point of view and then or that the person could have an ultrasound-guided aspiration of the dominant nodule and I think these terms of dominant are particularly scary for the patient so aside from the typographical errors this report doesn't really have a full description of the features that all societies used to characterize nodules and the recommendations are rather ambiguous and also incorrect and so that leads me to this article that just came up this week it was published online and it is an article that was from Canada these authors talk 1930 sarod ultrasound reports and they answer this a very question about the quality of our radiology reports do we describe all the features to make a decision to use a co2 roads or any of the other risk stratification criteria so they had six features that they were looking for five of them are the descriptors used in AC arteries and then the sixth feature was a nodule size they found that of the reports that were reviewed the mean score out of six was two point five seven and then they they determined that the nodule needed to have four or more features in order to actually correctly classify it with one of the risk stratification criteria and shockingly or and even maybe not surprisingly less than 14% of reports had these features in order for someone an endocrinologist or a surgeon to classify the nodule under one of the risk stratification criteria so this sort of sets the tone of how thyroid ultrasound is done in North America so why would you not adopt a CR tirades I've heard many of these comments in participating in the ACL tears activities and doing talks it's more time-consuming to report all six features right instead of maybe two as in that previous report just the echo genocide there are too many radiologists to teach this new system too and it's time-consuming we're all busy in our day-to-day work anyway some comments I've received earth at the referrers don't like it they think the reports are too long and the referrers prefer another system such as the ATA system or they just like the way we did it in the past and then finally the stenographers don't feel confident about using ACR tyrants then why would you adopt a sea otter ads and then this really gets to the crux of my next couple of sides it gives you better patient care and that's through improved quality of the report and also the fact that ACI tourettes performs better compared to no criteria or other criteria specifically for improved quality of the report it gives you a better descriptor of the findings of the nodule so that a CEO tirades or other risk stratification criteria could be applied it makes the recommendations more standardized and it reduces a variability so let me talk about that in this paper this is a paper that was published in JCR in May so just this month it is about improved quality of the thyroid ultrasound reports after implementing ACR tyrunt's this is a study that was done with radiology partners and in this study we implemented this new structured reporting template which had the nodule number and the size and the location a descriptor of the nodule features this is a pick list so not all of this text appears it's just solid if it's solid echogenicity hypoechoic if it's hypergolic so it's not usually this busy but it lists all of the five features under ACR tyrants and then in it gives recommendation for the total points and the turrets classification it talks about a change in the nodule if there was a previous ultrasound and then it finally gives a recommendation now in this study with radiology partners the first phase was just looking at the text reports without implementing this structured reporting template the second phase was using the first half of this template so just the location and the size and the descriptor and then in the third phase after ACR tirades was published we introduced the next part of the structured reporting template so we evaluated amongst these radiologists what happened if we used a structured reporting template with the features and what happened if we implemented ACR tourettes and in each of these three phases phase one two and three we looked at the recommendations in the radiology report and there were four options that we saw that they would recommend a biopsy that the radiologist would recommend an ultrasound the follow-up ultrasound that there would be a recommendation that the nodule needed no follow-up or biopsy or that there was no nothing said about the nodule no management recommendation when I first started doing the study I really thought that we would see fewer nodules being biopsied with ACR turrets but that's not what we saw there was actually no statistically significant difference in the nodules recommended for biopsy with a CR tie rods with the template or with the final recommendations so that was surprising in the early phases but what we did see is that after implementing ACI turrets the improvement really was that there were a few far far fewer nodules with no recommendations you can see without a CRT rods and a rad system third of nodules had no recommendation of what to do and what does this lead to well it leads to the referrer and the patient being confused and maybe more likely to have biopsy or follow-up when it wasn't necessary now in the next part of the analysis we looked at the frequency of specific findings the five spinnings under ACR turrets and we saw that in phase one when there was just free Tex that the reporting particularly of shape margin echogenicity were very infrequent when we implemented the structured reporting template there were actually fields that they had to fill out so in Phase three all of these features were reported whether they were present or absent and now when we looked at the frequency of the suspicious features this was this was a revealing of why they were the same amount of CAES in phase one when it was just free text the frequency of tolls and wide suspicious margin and punctate echogenic face i were relatively low but we saw that in phase 2 and 3 when the radiologist had to create a checklist and look at these and not features in the nodule the frequency of these findings these suspicious findings were actually much higher so it shows us that in Phase one just because the radiologist didn't mention these features wasn't because they weren't present it probably was because they weren't specifically looking and didn't see them and as a result the average ACI tourettes points was higher in phase 2 and 3 compared to phase 1 and that is why in phase 3 the same number of biopsies were recommended as because radiologist saw more nodules with these suspicious features that would appropriately be recommended for the biopsy so the take-home point from this study is that a structural reporting template serves as a checklist and that a CI tirades results in more definitive recommendations a next study that talks about reducing variability there's a study that we did that looked at in terms of variability and Franklin is going to talk about this article in more detail so I'm gonna skim over this but what I want to point out is that we looked at various sonographic features and we saw that margin and primatech Aegina close I had the porous agreement it was a fair and that an overall recommendation if radiologists weren't given a CR Tyra's was fair verging on slide but with a CI turrets no matter how radiologist varied in their interpretation it actually improved the recommendations being more consistent it was now having a capillary 0.51 which put it into the moderate agreement range so in this study we had high variability interpreting several ACR tyrunt's findings but applying ACR turrets reduces the variability and then the final paper that I'll talk about is one that was published in radiology in April of this year this study took eight radiologists who interpreted a hundred nodules the same nodules and these radiologists were asked to provide recommendations for biopsy as well as interprets the features of the nodules in addition we had three experts that interpreted the same 100 nozzles as the gold standard in interpretation and this study really talks about the reduction in thyroid nodule biopsies and improved accuracy when radiologists were asked to categorize the nodules from benign to highly suspicious before knowing about ACR tyrunt's this is in the pink color most radiologists thought nodules were mildly suspicious now I think this is a really good sign of radiologists being unsure and hedging being in the middle ground after we use their nodule interpretation and plate AC our turrets we can see that many more nodules were moderately or highly suspicious and this makes a lot of sense because these group of nodules were not rules they were all biopsied so they were more suspicious than the everyday diagnostic ultrasound thyroid nodule now despite more radiologists saying that nodules were mildly suspicious we found that compared to radiologists own judgment ACI tourettes applied to their own interpretation would have reduced the number of thyroid nodules biopsies from 80 to 57 now with fewer biopsies how did this affect the sensitivity and specificity this was a sensitivity and specificity an accuracy on average for the eight radiologists and this was the sensitivity specificity and accuracy after applying ACR turrets to their interpretations you can see sensitivity doesn't change significantly but specificity certainly does it improves from twenty to forty four percent and accuracy markedly improves as well we then use the expert interpretation and compared it to other risk stratification criteria and we found that ACR Tyra's has the same sensitivity is other categorization guidelines out there but the specificity was definitely better and the accuracy was better and Bill has a larger study that looks into this so I'm going to move on I do want to emphasize that what does it mean to miss a malignancy it is doesn't mean that the malignancy was missed because of an incorrect interpretation but there are certainly even high suspicion nodules that don't meet the category for biopsy because they're too small and we definitely covered this in the past webinars that were aiming to biopsy nodules that are suspicious but also result in a significant outcome in that if they're untreated though leads to harm for the patient and as we know thyroid cancer is very indolent and there are many cancers that don't progress and can be safely observed but these this table shows the five not viral malignancies that had a recommendation not to biopsy by one or more observer and we see here that there are two radiologists in particular that did not recommend follow-up for malignancies and therefore the these malignancies wouldn't be observed in their annual follow-up and this represents just two to two malignancies here and only three out of a hundred and twenty malignancy ugly houses and so 2.5% of malignancies and here the problem was a misinterpreting a solid nodule with some sister components as spongiform or mixes stick solid so in this study we identified definitely composition as an area to focus on so the take-home point from this article is a sea otter introduces a number of thyroid nodules biopsies when compared to radiologists existing practice patents and other criteria and that applying a CR tie rods improves the specificity and accuracy G's to reduced number of benign biopsies so to conclude my section using a CI turrets does provide better patient care it improves the quality of the report even if we're not looking at outcomes of whether it's a module is malignant or benign it gives us a better descriptor of the findings if we use a structured reporting template it makes the recommendations more standardized that is radiologists are more consistent with each other in their recommendations even if they report the nodule differently and then it reduces the variability in the recommendations and then the other advantage is that a CI terra's does have a better performance compared to no criteria or compared to other criteria and a bill will be reviewing his studies as well as some of the studies performed in Europe since ACO Tyros was published and so I am going to hand it over to Franklin now to do his section Thank You Jenny and as soon as I have control I'll go on to the next slide to start off with since publishing a CR tie rads we've heard from radiologists and other practitioners in multiple places saying some of what dr. Hong alluded to which is that they're struggling with some a CR tie rods features because they're reporting them and looking at them for the first time so my first question is which ones are the most challenging for stenographers and physicians deal with and as i alluded to during the last webinar there are a number of features that seem to give people more problems i specifically talked about mixed solid and cystic nodules being problematic and i talked about comparing a mixed nodule that has a roughly even distribution like the one on the left versus one with eccentric solid material like the one on the right and this is based on my anecdotal experience but i'd also like to talk about this in the context of variability that dr. hong also presented a little bit earlier and that is the importance of interobserver variability stated another way ultrasound features aren't helpful if observers don't agree on whether they're there or not for example if one observers says that a nodule is very hypoechoic and another says that a nodule is hyper echoic they'll receive different scores and be subjected to different management recommendations in terms of variability was assessed in this paper that dr. hung was lead author on and this is the paper that she showed with eight board-certified radiologists looking at 100 proven nodules assessing their agreement on feature assignment and biopsy recommendations based on their practice and the a section from the same table that dr. Huang just showed and highlighting that some features notably margin and punk aid echogenic foes I had low Kappa values indicating look high in terms of variability not surprisingly macro calcification which is usually pretty easy to determine that it is present or not had a pretty good cap of value and the others were in between so what do you do with this we look at this table and say there are some features that people can't agree on where do we go there are basically two routes that we can take the first would be to eliminate or change the definition of some features unfortunately these features have all been shown to have utility in distinguishing benign from malignant nodules therefore doing this would not be productive the alternative and this was the conclusion in this paper is to apply more education and this is something that we're doing by means of these webinars by means of publications and it's something I'm going to touch on in the end when I talk about the future of a CRT rods but the key point is that decreasing in terms of our variability requires people to look at nodules look at features and one thing that and and try to increase their their confidence that a feature is there or not and one thing that really helps in practices is have several observers review cases and look at nodules and see if they agree on whether features are present or absence to try to improve there in terms of agreement the next question which I'll spend a little more time on is what are the major differences between a CR tie rods compared to the American Thyroid Association and the SRU recommendations and I chose those two because the ATA recommendations are frequently requested predict by our endocrinologist refers and the SRU recommendations I chose because many practices are still using them particularly in radiology practices at the highest level really there are a few differences all the risk stratification systems start with imaging then progress the characterization of nodules and finally come up with management decisions so this is at a very high level so-called fifty thousand foot view but there really are differences when you zoom in let's start off with imaging well you'd think that imaging is imaging right but the truth is that there are some distinctions between the systems that are highlighted in this table starting off with the number of nodules to scan a report ACR Tyrod specifies up to four and by this I mean that this is the maximum number of nodules that are formally reported and for which all the features are enumerated and they are classified it doesn't mean that you don't see the other nodules it just means you pick up two for to formally document in your report neither the SRU nor the ATA guidelines state a limit and we chose this because we thought it was something that could be applied practically in a typical ultrasound practice there are slight differences in measurements but the ACR Tyrod system and the ATA specified that nodule should be measured in three dimensions interestingly the SRU doesn't really talk about three dimensional measurements although it refers to the maximum two measurements and that indicates the maximum measurement of a particular nodule there are more differences in the characterization step ACR tie rods began with a feature lexicon that was published in 2015 Oh show more about that in a moment and we did that because we wanted to have a basis on which ACR tie rods classification would be founded now they're the SRU lines nor theater guidelines specifically referred to a lexicon of ultrasound features in fact in both of them it was sort of assumed that a particular feature that was listed as a criterion you would know when it was present or not and we felt that having a lexicon with examples was an important foundation for what we were doing of course there are differences in how risk assignments occur in the three systems ACR tie rods points assigned in five feature categories compared to a tabular method for the SRU based on composition in size and by composition I referred to all the internal characteristics of a nodule and for the ATA patterns of appearances and composition and size and finally lymph nodes which are not formally considered in a CRT rods and we're considered by the SRU and ata and this is the front page from the lexicon paper from 2015 and here's the table of the five feature categories that you see here on the left and I've outlined the more suspicious categories in red that I'm highlighting here and these were translated into the well known by now a CR tie rods table which are charts which I won't go through in detail the five categories in which you choose one in each of the first four and then all that apply and the last one add up the points two and combine that with the size to come up with a TR level ranging from one to five and decide based on that whether to recommend an F an A or follow-up compare that to the SRU table on the left and the ATA on the right on the left we see that this SRU table showed these different appearances and each of those was associated with the management recommendation in essence these are really patterns the ATA is definitely pattern based you basically take the No jool that you're looking at and try to match it as closely as you can with one of these patterns in the chart and as you go up in the chart from bottom to top the suspicion level and the recommendations for biopsy change the problem with patterns that we've highlighted in previous webinars is it not every nodule fits a pattern in fact in one recent study of 2900 or 5 percent were not classifiable by the ATA guidelines and those nodules had a 19% malignancy rate there are even greater differences in management between the systems the size thresholds particularly ACR tie-red specifies thresholds of two point five one point five and one centimeter at TR three four and five the SRU talks about either strongly considering or just considering biopsies depending on features in size and the ATA says fna or consider F&A depending on the pattern and size follow up in a CR tie rise as i mentioned earlier is specified as each TR level so if you have a nodule for which fna isn't needed you know whether to recommend follow-up or in some cases no follow-up and the SRU had no specific recommendations for follow-up intervals and the ATS recommendations are limited as I mentioned earlier lymph nodes are not formally considered in a CR ty rads although we said that lymph node should be biopsied of abnormal the SRU said EPSA lateral nodule should be biopsied if there's an abnormal lymph node on that side and the ATA said to biopsy and irregular nodule if there's a suspicious note the differences highlighted in this chart from a recent paper in radiology that compared these seven systems and I've highlighted the sensitivity specificity and accuracy for a CRI rads and you could see how that compares to the others we knew in designing a CR tie rods that our sensitivity would be lower than other systems but you can see that our specificity and our accuracy were considerably higher and Bill will talk about this shortly why does a CR tie rod specify larger size thresholds for f na of some nodules going back to the pattern system in the ATA guidelines you can see all the patterns and the different size thresholds as you go up in suspicion if you overlay on that the point levels you can see that for example at the high suspicion levels most of the nodules would be classified as a CR tie rods 5 and would be but I've seen an ACR tie rods at a threshold of one centimeter concordant with ata this pattern hypo will cope with irregular margins would be classified as TR 4 and biopsied at a 1 point 5 centimeter threshold similarly for intermediate suspicion nodules which would both the class about is tr for the ACR guidelines would biopsy at one point five centimeters versus one centimeter for a TA at low suspicion TR 3 these 2 TR 3 nodules would be biopsied at threshold of 2.5 centimeters in the ACR tie rods compared to one and a half centimeters for the ATA what underlies our decision to use larger size thresholds and the questions we looked at it what's sizes do nodules become problematic and our ultrasound measurements in quotes accurate because accurate is a relative term many of the other guidelines including the ATA based their recommendations for size on this study from 2005 which looked at gross tumor size versus risk of dis metastasis and they said in this paper that there was a an abrupt takeoff at two centimeters however they were using gross tumor size and I'll talk more about that in a moment since then dr. new yen and several of us published a paper in thyroid earlier this year and we found no threshold effect on risk of distant Mets or nodal Mets or local invasion for tumors less than four centimeters and moreover didn't find a great change in mortality until the size was greater than two and a half centimeters as well previous research showed that ultrasound over measure measures nodules by an average of about five millimeters and we took all these things into consideration in our larger size thresholds as well the fact that we supply follow-up recommendations mitigates the likelihood of us missing a cancer over the long term finally can use a CR tie rods in the ATA guideline simultaneously well yes you can but you need to be prepared to deal with discrepancies and those are highlighted again in the ATH art showing the nodules for which there would be significant discrepancies between the ATA and the ACR tie rods guidelines if you report both you have to be prepared to deal with referring physicians who were going to notice the discrepancies and since patients get to see reports more often than they used to in fact it's becoming the norm they would often have questions so with that I'll turn the control over to dr. Middleton for his presentation thank you okay this is dr. Middleton speaking from Mission Control so my assignment was to compare the results of the ACR Tyrande to other existing stratification systems and I really want to focus on three specific studies a little bit on a fourth if we have time this is the one that I will start with and it comes from the Society of radiologists and ultrasound who developed a group of investigators that looked at modules between 2006 and 2010 it was a group of six different academic medical centers and I'm focusing on this article first because it's the largest of the existing comparisons and also because the first author is a very close and personal friend of mine so that cohort was 3400 about 3,400 nodules that had definitive fna and/or histology results we looked at all these modules and retrospectively determined their ultrasound features features that you all are familiar with now and then we classified each module based on the ACR Tyrande from 2017 the Korean Thai rads from 2016 and the ATA guidelines from 2016 so of the 3400 plus modules a little over 3,000 were benign and 352 were malignant and you can see the breakdown of the different types of malignancies that were included majority were papillary or the follicular variant of papillary cancer and then for each one of these modules we divided them up into categories depending on what each classification scheme would place them so they would either place them in an MA recommendation a follow-up ultrasound recommendation or no further evaluation recommendation and this is a chart that shows the breakdown of where the modules were classified on the different systems so up here at the top are the relatively innocuous looking modules very low suspicion benign lesions and the TR ones and twos for the AC art iran's and then as we move down we get into the more suspicious modules and you can see that there's a large number of modules in each one of these categories but what I really want to focus on is the non classified modules so notice that in the ACR system there were no nodules that could not be classified on the other hand in the ATA and the korean classifications there were at least a substantial minority of modules that could not be classified particularly in the ATA guidelines and when we just focus on those guidelines we found as frankly mentioned earlier that there were 14% of the ATA nodules that could not be classified into any of these risk categories and of those 9% were malignant and that's a fairly significant minority of the mount classified nodules and then if you just look at all malignant nodules 13% of them were not classified based on the ATA guidelines and here's a charge that I do want to spend a little bit of time with so it looks at the ACR the ATA and the Korean guidelines and it looks at these different parameters so a CR has definitive recommendations for which nodule should be biopsied and which module should be followed and then leaving the rest of the modules as modules that don't require any further evaluation the ATA has body recommendations about which module should be followed korean guidelines don't have any recommendations about follow-up so i'm just going to focus to begin with on the biopsied categories we'll talk about followed categories later so in this first row yield of malignancy that refers to the number of biopsied modules that turn out to be malignant and you want that to be high you want to have as many nodules at your biopsy be malignant and as few nodules as possible be benign you want your yield to be high and as you can see in this study the ACR yield was higher substantially higher than either the ATA or the Korean yield and the next row is another statistic and this is the percentage of malignant modules that would be biopsied using the different systems so this is a number that you want to be high you want to biopsy as many of the malignant modules as possible in the ACR system it was 68 percent a little bit higher significantly significantly higher in the ATA in the korean systems and this is something Franklin mentioned earlier and this is something we expected all along when we set up the guidelines if you just look at the modules that are greater than a centimeter greater and greater than or equal to a centimeter then that sensitivity or that percentage of malignant modules that are biopsy increases quite a bit to 83 percent it also increases with the other systems and there remains a significant difference between the three then we move on to this next row where we focus on benign modules so the percentage of benign nodules that are biopsied in the ACR system is 47% you want this number to be as low as possible you'd like to biopsy as few benign benign nodules as you can and in our study this was significantly lower than either the ATA or the korean systems and again if you just look at nodules that are bigger than bigger than or equal to a centimeter assuming that anything less than a centimeter is not significant regardless of what it is you see these statistics that the ACR biopsies 50% of the benign nodules that are greater than or equal to a centimeter ATA and the korean systems biopsy a little more than 80 percent okay so next let's do focus on these followed nodules so since the ACR does give discrete recommendations on what should be followed I wanted to look at this as well and of the thought of the nodules that would receive this recommendation about 12% turned out to be malignant of all the malignant modules about 21% would be followed under the ACR criteria so if you look at this if you add these two numbers up that's almost 90 percent of the malignant modules would either be biopsied or followed under the ACR system which is higher than the number that would be biopsied under ata or Korean we just look at the malignant modules greater than a centimeter 11% would be followed if we look at benign nodules 18% would be followed and again if you add these two numbers up the total continues to be less than the modules that would be biopsied with the ATA system or the Korean system much less than modules that would be followed with those systems now let's do just talk a little bit about the APA follow up so APA does have some guidelines for follow up if you look at all the nodules that did not meet guidelines for biopsy and that's a small percentage in this study only 22% of nodules didn't meet guidelines for biopsy by the ATA recommendations but of these 756 55% had no recommendation either for or against follow-up 16 of those were malignant 11% had a recommendation for follow-up in 6 to 12 months and 41 or half of those were malignant 27 percent of these nodules that didn't meet guidelines for biopsy had a recommendation to consider follow-up in 12 to 24 months I'm not sure what referring clinicians do with a recommendation to consider follow-up but of those 26 were malignant and there were only 50 nodules out of all of those where there was a recommendation for no further follow-up now those 50 nodules account for just 1.5 percent of all of the nodules that were evaluated so only 1.5 percent had a definitive recommendation for no further evaluation ie no biopsy or no follow-up based on the APA guidelines now you compare that to the ACR guidelines 32 percent had a definitive recommendation for no further evaluation so compare those two results ACR leaves you with final recommendations of no further resources need to be wasted on 32 percent of modules and the ATA it's just 1.5 percent of the modules so the conclusion from this study was that the ACR tie-red compares favorably with the other systems the ACR system has a higher biopsy yield of malignancy primarily because of reduced number of biopsies of benign nodules it also eliminates further evaluation ie biopsy or follow of a much higher number of modules the ATA in the korean systems result in biopsy of a higher percentage of malignant modules but this difference is at least partially mitigated by the follow-up recommendations included in the ACR system so the next paper is one that Franklin has referred to a couple of times this comes from a group of Korean radiologists who looked at 2,000 nozzles and they compared the diagnostic performance of seven different society guidelines of their 2009 about 1500 were benign and the benignity was proven by surgery or histology in a small percentage by a benign biopsy with a repeat biopsy that was also benign in a smaller percentage and most of them have benign biopsies to begin with and then were followed for 12 months so very good proof that these were benign there were 450 for malignant nodules and almost all of those were ultimately proven surgically and you can see the breakdown of the types of malignancies again papillary and follicular variant of papillary cancer were the most common and these are the standard statistics for the different societies so this is the ACR the American Thyroid Association American Association of Clinical Endocrinology the National Cancer comprehensive Network French society of endocrinology Society of radiologists and ultrasound and the Korean thyroid Association and if you just look at sensitivity you want that obviously to be high and as we've said before that's one of the things that is lower with ACR Tyrande lower than most of the other systems specificity on the other hand is one of the strengths of the ACR and as you can see it had the highest specificity of all of the guidelines positive predictive value obviously want that to be high and in fact it was the highest in the ACR guidelines and negative predicted back value you want that to be high as well and that was fairly similar in all of these categories there was a little difference in the negative predictive value and overall accuracy again was highest in the ACR system compared to all the other systems so it performed very well when compared to many other established guidelines and then these authors also evaluated the systems in a manner somewhat similar to the first study that I talked about they looked at the total number of modules that would require biopsy based on the different systems and you can see the ACR recommended biopsy and the fewest number of nodules of all of these the yield of malignancy and the yield of benignity also very similar to the previous study you want the yield of malignancy to be high and it was the highest in the ACR the yield of benignity to be low and it was the lowest of all the possibilities and then they also looked at something called unnecessary fna rate and that was defined as the number of benign biopsies divided by all of the nodules in the study and again that's something that you would want to be as low as possible and in fact it was as low as lower than any of the other guidelines and then they also looked at false positive rate and I show this just because the false positives were lowest in the ACR guidelines but they don't really define how they the formula they used to calculate that so I take that with a little bit of a grain of salt and in the final big study that come out since the Thai rads was published is from a group of primarily endocrinologist from Italy and they looked at the performance of the ATA the AAA CA and the ACR systems in identifying cytologically high risk fibroid modules and they looked at a little more than a thousand nodules fine needle aspiration was used as the gold standard there was no surgical or histologic correlation in this study they have over 900 modules that were benign and they had 113 modules that were presumably malignant based on high-risk cytology so some of these categories that were in the benign category and in the high-risk category fell into psychologies that were indeterminate they divided the indeterminate cytologic results into low-risk and determinants and high-risk indeterminate so they assumed that all of these would be malignant and all of these would be benign but clearly that's that's not the truth the majority in each case would be true but not all of them so the manuscript reviewer in me would have raised a lot of red flags about this way of defining nodules nevertheless they did provide some useful data and this again sort of highlights the fact that there are non classifiable nodules in both the ATA guidelines and the AACE guidelines five percent in a TA and 2.8 percent in the AAC e and of the ones that were non classifiable in the ATA mine were cytologically high risk ACR again was able to classify all the nodules the primary reason that I wanted to review this article is they also showed receiver operating curves for the three different systems and for those of you that are familiar with this I won't spend too much time with it but this is a plot of sensitivity on the vertical axis and one - specificity on the horizontal axis and for receiver operating curves you want the curve to be as far above this dotted line as possible you want it to be in this upper left-hand corner as much as possible and you can quantify that by calculating the area underneath the curve and as you can see here they've done that the little triangular line or the line with the triangles is a CR system the line with the black boxes is the AACE and the line with the grey circles is the ATA and you can see that the area under the curve is the best for the AC art iran's so in conclusion they wrote this that their study shows that the ultrasound classification system proposed by the ATA to AACE and the ACR differ in their ability to identify modules at high risk of malignancy in particular the ACR Tyrande classification system has the highest ROC area under the curve for the identification of high-risk nodules and is the only ultrasound scheme able to classify all thyroid nodules their research also confirm a relevant limitation of the AC or ata classification which leaves unclassified nodules at relatively high risk of malignancy so based on the current evidence there are some conclusions that you can make first of all the ACR Classifieds all nodules it's the only system in most of the existing systems that will classify all nodules the ACR tie rads reduces the total number of modules requiring fna or follow-up compared to other systems and the ACR Tyrod reduces the number of benign nodules requiring fna or follow-up a CRT rads will miss more cancers than most other systems but many of those missed cancers are less than a centimetre in size and many of those that are missed will be followed for five years based on the follow-up recommendations in the ACR ty rads so that's really all the time that I have we do have a few more points to make and I'm going to turn the controls back over to Franklin thank you Bill and in the last few minutes I want to talk about where we're going next so you've heard a fair bit this morning about a CR tie rods versus other systems but we have to remember that this is in the contest we all have the same goal which is to reduce the number of biopsies of benign nodules while detecting appropriately malignant nodules there are more similarities between all the systems out there than there are differences so to paraphrase Rodney King can we all get along I think the answer is yes and to that end I started a grassroots effort we're calling ourselves the International thyroid nodule working group our first teleconference was held one week ago and although it makes it look like we spent two days on this it was really 70 minutes the two dates are because of the different time zones involved and this involved 17 radiologists and dark Rinaldi's and surgeons from all the countries that you see here and this was an organizing call to decide where to go next but we agreed that our ultimate aim is essentially a universal tie rods we're going to do other things as well for example we're going to look at creating a internationally available Ashleigh's of ultrasound features to answer the educational question or issue that I mentioned earlier and I don't know as I speak to you this morning where this will go but or now this afternoon rather where this will go but I'm hopeful that this will be a good cooperative effort it's not going to be done quickly within a few months it's going to take time but both doctors Hung and Middleton were on the call and I think they would share my enthusiasm for what was expressed by thyroid experts from around the world who said that we should work together and try to reach consensus so with that pass control back on to dr. hyung to conclude yes and the other initiative we have is a multicenter thyroid ultrasound registry that's a grant funded from the ACR and Akron with six practices both private and academic being involved this registry eventually this research registry will eventually be a practice registry so that your sites can participate as well and get some benchmarking nationally of how you're performing with regards to biopsy or a malignancy rate and then the frequency that you're interpreting these findings so that is the other aspect of looking forward with ACR tirades gathering more evidence to compare it to other risk stratification criteria out there well we're on to our last two minutes I wanted to show you the website for ACR Torres which will have links to publications as well as links to the all of the webinars that we've had well this is a thorough adultress on registry and the participating sites and I'd like to thank them for their participation so the recording will be available all of the webinars how we do it frequently asked questions and evidence behind ACR towards will be made available I would like to take this moment to thank Lauren Hicks from the ACR for coordinating and managing these efforts again with this webinar she'll have the questions and the timestamps of when they appear on the webinar Franklin and Bill did a fantastic job in simplifying the literature and making it digestible I'm certainly grateful that we left enough time for Bill to most eloquently describe and compare the every stratification Guidelines Franklin that chart on ace Tyrod superimposed on the 88 was wonderful I think all practices should have that up in their reading room if they there refers like a TA guidelines so that is it thank you to everyone of Bill Franklin do you have any additional comments no I just would like to add my thanks to Lauren for her invaluable assistance in getting these organized and it's been it's been fun it's been interesting and educational and to help valuable to people who've watched these webinars thank you everybody oceans right thank you thanks for spending an hour with us bye-bye bye-bye hi