PAPILLON COURSE on THYROID ULTRASOUNDSection 2 The nodular goiterPart 9 TIRADS
|
|
|||||||||||||||||||||||||||||||||
INTRODUCTION |
|||||||||||||||||||||||||||||||||
In this chapter we give an overview of the evolution of Thyroid Imaging Reporting and Data Systems (TIRADS). Then, we analyze the performance of TIRADS is the three most frequent types of thyroid cancer. Thereafter, we briefly discuss the concerns about TIRADS. |
Finally, we present the ultrasound (US) characteristics which are involved in various TIRADS. Essentially, this Supplement of the manual is the collection of written material presented in other parts of Section Nodular goiter. |
||||||||||||||||||||||||||||||||
EVOLUTION OF TIRADS |
|||||||||||||||||||||||||||||||||
There has been a continuous development over the past 40 years in the selection of thyroid nodules that qualify for fine-needle aspiration cytology (FNA). The main initial driver used to be recognizing all malignancies; the principle of reaching the highest sensitivity was supported by the introduction of thyroid ultrasound (US) in everyday practice. From the pre-US era, when FNA was indicated in palpable 'cold' thyroid nodules (1) the number of biopsies gradually rose and peaked around the publication of the AACE/ACE guideline in 2006 (AACE-2006). This guideline suggested FNA of all nodules larger than 10 mm, and even of smaller ones presenting clinical or US suspicion of cancer (2). In the past 10 years focus has changed and is mainly driven by the principle of balancing diagnostic sensitivity and specificity and reducing the number of FNAs. The main tool for this change has been the consideration of suspicious ultrasound characteristics for the decision of performing FNA (3-11). Practically all guidelines published since 2006 link indications for cytology to suspicious US signs. There are some differences between various TIRADS (5-9) in what is considered suspicious, and in the smallest thyroid nodule size worth considering for cytology. Using the same approach, the various TIRADS have clarified what ultrasound patterns merit cytology; further, sampling is not recommended for isoechoic nodules between 1 and 2 cm in diameter, and for those between 1 and 1.5 cm that do not show suspicious signs, as stated by the AACE-TIRADS, ACR-TIRADS, EU-TIRADS (5, 6, 8), and ATA-TIRADS and KSTR-TIRADS (7, 9), respectively. |
The basis of all TIRADS systems are those US characteristics which have diagnostic importance in the recognition of papillary thyroid carcinomas (PTC) (12). The US characteristics of FTC differ profoundly from PTC characteristics (13-15). Only a limited number of studies deal with the importance of TIRADS in medullary thyroid cancer (MTC) (16-20), and follicular thyroid cancer (FTC) (21-23). To the best of our knowledge, there has been only one paper which focuses on the diagnostic performance of TIRADS in patient selection for FNA in FTC (23). The fact that TIRADS' performance has been tested almost exclusively in PTC patients, supports the view that the clinical validity of TIRADS systems cannot be unconditionally extended to other forms of thyroid malignancy (24). Therefore, while the US-centered diagnostic work-up can effectively identify PTCs, it is not known if TIRADS systems are similarly successful in finding FTCs and MTCs burdened by substantially higher risk of mortality (25-27). |
||||||||||||||||||||||||||||||||
THE DIAGNOSTIC PERFORMANCE OF TIRADS IN THE THREE MOST FREQUENT THYROID CANCERS |
|||||||||||||||||||||||||||||||||
As has been mentioned multiple times earlier, the suspicious characteristics means that a feature is suspicious of conventional variant of papillary cancer (CV-PTC). There are numerous studies which have confirmed that decision on FNA based on TIRADS can reveal most CV-PTC. Most PTCs where FNA is not indicated belong to the follicular variant of papillary cancer (FV-PTC). TIRADS fails in around 15% of PTC, these are mostly FV-PTC. Much less is known about the diagnostic performance of TIRADS in medullary cancer (MTC). Studies focusing on this topic have shown that TIRADS has a similarly good sensitivity in MTC than in PTC. FNA is indicated on TIRADS in around 85-90% of MTCs. This is not surprising because MTC and PTC behaves similarly regarding the occurrence of suspicious US characteristics. |
Compared with the former two subtypes, much less is known regarding the TIRADS in follicular thyroid cancer (FTC). The few papers have demonstrated a significantly lower TIRADS score in FTC compared with PTC. Given the marked difference in the frequency of suspicious tickets, this is again not surprising. FTC and PTC showed just the opposite FTC and PTC behave just the opposite with respect of all (!) suspicious signs, and this difference is not simply statistical, but significant in practice. As a consequence, FNA is indicated in only 55% of FTCs in the range where the TIRADS score influence the decision on FNA. |
||||||||||||||||||||||||||||||||
CONCERNS REGARDING INDICATION OF FNA ON SUSPICIOUS SIGNS |
|||||||||||||||||||||||||||||||||
First, we remained the reader, that the potential weakness of this decision is related not to the recently developed TIRADSs but is related to the approach that has appeared in all guidelines since 2006, which links the performance of cytology to the presence of suspicious signs instead of the previous 1 cm size limit. |
In this respect, TIRADS is not a novelty in principle, but a systematization of previous proposals. |
||||||||||||||||||||||||||||||||
Concern regarding missing a significant proportion of follicular thyroid cancers |
|||||||||||||||||||||||||||||||||
There are two fundamentally different approaches on the same observations. The guideline of the American Thyroid Association considers it acceptable not to recognize FTCs smaller than 20 mm in their largest diameter, because distant metastases occur rarely in such lesions (7). The other view states that one has to exert every effort to recognize FTCs well before they reach 20 mm in diameter, because once the nodule becomes larger, distant metastases may evolve (31). Opinions may, of course, differ. At the same time, it is worth noting that the conservative approach common in medicine, according to |
which we have changed paradigm only on the basis of careful, repeatedly confirmed studies, is missing from the ATA-type opinion. We have no data on the real risk of not recognizing half of follicular cancers smaller than 2 cm. You can override the cautious approach if both thyroid professionals and patient organizations accept that the cost we have spent so far on recognizing each of the follicular cancers over time should be utilized differently. There is no such consensus. We ourselves fully agree with Trimboli et al. who stated that TIRADS should be regarded as accurate tools only to diagnose PTC (11). |
||||||||||||||||||||||||||||||||
A selection bias which hinders the recognition of cancers missed by the TIRADS |
|||||||||||||||||||||||||||||||||
The true sensitivity of TIRADS, can be tested only in such a non-restricted cohort of patients. Most, if not all, recently published studies have estimated the diagnostic performance of TIRADS in cohorts selected for FNA and surgery based on the same principles on which the tested TIRADS method, is based (3, 10, 11, 14, 33, 34). It means that those groups of investigators who tested TIRADS on a surgical cohort of patients in which FNA was indicated on suspicious characteristics, could not reveal those cancers (mainly FTCs) which did not show suspicious signs. This implies that nodules which failed to show the required features of high suspicion escaped |
FNA and surgery, and were therefore not diagnosed as malignant. This selection bias has prevented recognition of the low sensitivity of EU-TIRADS in detecting FTCs in the 10 -20 mm range. This self-fulfilling prophecy (TIRADS is excellent because reveals cancers harboring suspicious characteristics) can lead to even more liberal approach in FNA indication. |
||||||||||||||||||||||||||||||||
Issue of subcentimeter nodules |
|||||||||||||||||||||||||||||||||
While very lifelike and practical, it makes it difficult to apply the suggestions of TIRADS allowing contradictory actions. |
The main tables of all TIRADS stands for not to perform FNA in subcentimeter nodules showing suspicious features while in the text of all TIRADS they permit FNA in such lesions. |
||||||||||||||||||||||||||||||||
No room for uncertain answers |
|||||||||||||||||||||||||||||||||
TIRADS do not provide an opportunity for a doubtful answer. On the one hand, this is unrealistic, and on the other hand, an endocrinologist who does not |
perform ultrasound, or is a less experienced ultrasonographer colleague, may have the false impression that we can give answers of mathematical level. |
||||||||||||||||||||||||||||||||
The stigmatizing effect of a formalized classification system |
|||||||||||||||||||||||||||||||||
Mentioning this problem may seem pedantry. However, for many patients, this is a real problem having a 'high suspicion' nodule. |
Especially since the associated malignancy risk is perceived by many to be threatening even after a negative cytological finding. |
||||||||||||||||||||||||||||||||
SUMMARY |
|||||||||||||||||||||||||||||||||
TIRADS was originally developed to make it easier to compare the appearance of nodules in different patient groups. For this purpose, the TIRADS performs excellently. Moreover, the huge effort of the various authors revealed basic problems in thyroid US. Such as, the lack of a common interpretation, and the often confusingly large differences in interpretation. Over time, TIRADS has played a much more important role in deciding |
which nodule should be examined cytologically. We are forced to mention that TIRADS could only be suitable for this role if there were no follicular cancers or if we accepted that nearly half of follicular cancers smaller than 2 cm were not recognized. Until one of these two is met, TIRADS can only be used responsibly as a decision support system. TIRADS may be worth using as one of the most important, but by no means exclusively, determinants in the indication of cytology. |
||||||||||||||||||||||||||||||||
| References
|
|||||||||||||||||||||||||||||||||
Supplement: THE ROLE OF ULTRASOUND CHARACTERISTICS IN TIRADS |
|||||||||||||||||||||||||||||||||
| The supplement contains those parts of the manuals presented in the Nodular goiter section which present the role of US characteristics in the various TIRADS. The references rely to the Manual in the relevant section. | |||||||||||||||||||||||||||||||||
Composition and TIRADS |
|||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||
Spongiform nodules are grouped among the possible lowest category in all (12,14,15,17) but the ATA TIRADS which puts this nodule into category 2 (16). On the other hand, ETA restricts the term spongiform for nodules which are composed entirely of spongiform areas, while the other 4 TIRADS define a nodule as spongiform if the microcystic areas > 50% (12). Three TIRADS group only completely cystic nodules into the lowest category while AACE does so if fluid component >80% (14). This difference has practical relevance in such lesions presenting suspicious signs; these lesions are grouped among the most benign category or into the most suspicious category, AACE or ATA, ETA, Korean TIRADS, respectively (12,14,16,17). The presence of comet tail artifact has no relevance in the ATA and ETA TIRADS (12,16). In contrast to them, all lesions presenting comet tail artifacts |
are grouped among the most benign category in the event of Korean TIRADS (17). It emphasizes the correct interpretation of intranodular echogenic figures, which is not always possible. E.g. the ACR TIRADS enlist small comet-tail artifacts found within a solid portion into the same category as microcalcifications because these are difficult to differentiate (15). The ACR system differs from the four other TIRADS (15). In complete or almost complete cysts the system does not deal with the solid part. The issue is that this guideline did not define 'almost' complete cysts. The other important and unique consideration in this system is the distinction between small (< 1mm) and large comet-tail artifacts. The latter is handled as a benign feature while the former, if found in solid part, is classified together with microcalcification as 'punctate echogenic foci' and is worth the highest possible points in this categorization. Considering the possibility of colloid crystals indeed or only apparently in solid part, this might lead to overestimation of colloid crystals. |
||||||||||||||||||||||||||||||||
Echogenicity and TIRADS ROLE OF ECHOGENICITY IN VARIOUS TIRADSThe echogenicity is the most important sign in characterizing of thyroid nodules. |
|||||||||||||||||||||||||||||||||
The AACE/ACE-AME, the ATA, the European and the Korean TIRADS |
|||||||||||||||||||||||||||||||||
At first, we give an overview of those four TIRADS which have a similar structure of decision. A hypoechoic nodule should be grouped among the most or among the most but one suspicious subgroup in all four systems. Essentially there are two types of systems, those which make distinction between deeply/markedly and minimally/moderately hypoechoic nodules (AACE/ACE-AME and ETA TIRADS) and those which handle all hypoechoic nodules uniformly (ATA and Korean TIRADS). In the first two systems, a deep hypoechogenicity itself is enough to categorize a lesion among the most suspicious subgroup, |
while minimally/moderately hypoechoic nodules are grouped among the most or most but one category depending on the presence of other suspicious features. The remaining two TIRADS do not differentiate deeply and minimally/moderately hypoechoic lesions. The categorization of a hypoechoic lesion as intermediate or high suspicion lesion depends on the lack or presence of other suspicious findings. However, in these two TIRADS the prerequisite is nodule's hypoechogenicity to classify a nodule as most suspicious. |
||||||||||||||||||||||||||||||||
ACR TIRADS |
|||||||||||||||||||||||||||||||||
This system makes a distinction between minimally/moderately and very hypoechoic nodules, the former is worth 2 points while the latter is worth 3 points according to the algorithm. The only hypoechoic nodule which does not belong to any of the two most |
suspicious subgroups is that minimally/moderately hypoechoic lesion which is dominantly cystic and lacks any other suspicious findings. Any other hypoechoic lesion is grouped among the two most suspicious subgroups depending on other characteristics.
|
||||||||||||||||||||||||||||||||
Various intranodular echogenic figures and TIRADS |
|||||||||||||||||||||||||||||||||
The presence of microcalcification automatically groups the lesion among the most suspicious subgroup in the AACE and in the European TIRADS, while a lesion presenting microcalcifications has to put either to the most but one or most suspicious category in the ATA and Korean TIRADS, isoechoic or hypoechoic nodules, respectively. Regarding the ACR TIRADS, the term 'punctate echogenic foci' is used for microcalcifications and short-tail comet-tail artifacts and worth three points. Following the logic of this system, it means that almost all lesions presenting this feature should be categorized into category 5. (See Table 4.) The presence of macrocalcification has no role in three TIRADS. While it is worth one or three points in ACR TIRADS, macrocalcification or peripheral (rim) calcification, respectively. It means that ACR treats rim calcification in a similar way to widely accepted suspicious sign; an approach which is not backed by scientific data. Rim calcifications with extrusive soft tissue component is enlisted among the most suspicious characteristics in the ATA TIRADS. |
It is interesting that comet-tail artifacts are handled differently in various TIRADS. At the one end is the Korean TIRADS which groups all nodules presenting comet-tail artifact among the possible lowest category, if comet-tail artifact is found in cystic portion and the nodule lacks suspicious US characteristics. Although AACE mentions comet tail artifacts among features characteristic for benign lesions, the lack or presence of comet tail artifact does not influence the classification of nodules in AACE. In the event of ACR TIRADS large comet-tail artifacts worth 0 points, while short comet-tail artifacts in the solid part worth the maximal as possible 3 points because this is included in the punctate echogenic foci feature. Considering the difficulties of the correct judgement of comet-tail artifacts, neither the Korean nor the ACR approach seems to be moderate enough, the former might lead to underestimation of a nodule presenting comet-tail artifacts while the latter might lead to overestimation. Until we reach a significantly better agreement in the judgement of this characteristic, the more restrained handling of this sign by the ATA, ETA (they do not involve comet tail artifacts in their system) and AACE seems more well-founded". |
||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||
Nodule' borders and TIRADS |
|||||||||||||||||||||||||||||||||
No TIRADS involves blurred borders among suspicious signs, while lobulated/spiculated margins belongs to the most suspicious features in all TIRADS (19, 21, 24-26). The presence of lobulated/spiculated margins automatically groups the lesion among the most suspicious subgroup in the AACE and in the European TIRADS, while to group a lesion presenting irregular margins into the most suspicious category in the event of |
hypoechoic nodules in the ATA and Korean TIRADS.
Regarding the ACR TIRADS, the presence of irregular margins worth two points. Following the logic of this system, it means that such lesions should be categorized into category 4 or 5 depending on the presence of other suspicious signs. (See Table 4.) |
||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||
Nonparallel orientation and TIRADS |
|||||||||||||||||||||||||||||||||
Taller-than-wide sign (TWS) belongs to the most suspicious features in all TIRADS (38-42). The presence of TWS automatically groups the lesion among the most suspicious subgroup in the AACE and in the European TIRADS, while to classify a lesion showing TWS into the most suspicious category requires that the nodule should be hypoechoic nodules in the ATA and Korean TIRADS. Regarding the ACR TIRADS, the presence of TWS worth three points. Following the logic of this system, it means that such lesions should be categorized into category 4 or 5 depending on the presence of |
other suspicious signs. (See Table 2.) There is an important difference between the TIRADS. It affects the handling of nodules presenting TLS. Three TIRADS involve TLS among nonparallel orientation (40-42) while two do not (38-39). It seems to be more justified to include TLS among NPO (43). |
||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||
Extrathyroidal extension and TIRADS |
|||||||||||||||||||||||||||||||||
In the sections discussed so far, we examined the suitability of US for prediction of extrathyroidal extension (ETE). We now investigate whether the US signs of ETE can be predict thyroid cancer - regardless of the presence or absence of ETE. US signs of ETE may be useful predicting thyroid cancer because PTC is preferentially located at the edge of the thyroid, either in the upper pole (32) of the lobe or in the isthmus (33). It seems to be a reasonable assumption that the location and the potential ETE are linked. But I underline here again, that the US signs of ETE for the prediction of cancer is not influenced by the real occurrence of ETE in TIRADS scoring. There are two main issues. Studies in the literature almost exclusively focus on the prediction of ETE and not on the prediction of cancer. This is very disturbing because e.g. the three American TIRADS use US signs of ETE for prediction of cancer when they involve these signs into generating TIRADS score. To the best of my knowledges, there is only one study in the literature which focused on the diagnostic performance of US signs of ETE in predicting thyroid cancer, namely medullary cancer. This study has found that the US signs of ETE have a similar diagnostic value to other suspicious US characteristics (34). The second issue is that it is not clear and not described in the literature and in the guidelines which US signs or which combination of US signs should be taken into account that we say that the nodule shows sign of ETE. It seems to be reasonable that only abutting nodules with non-continuous nodules fulfill the criteria of suspicion. Basically, three TIRADS involve ETE in the categorization of thyroid |
nodules (see Table 5). In the TIRADS of the American Thyroid Association (18) and in the collaborative TIRADS of the American Association of Clinical Endocrinologists (AACE), American College of Endocrinology (ACE) and Associazione Medici Endocrinologi (AME) (17) the presence of ETE automatically puts the lesion into the most suspicious category. The AACE guideline has an important consideration; in contrast with other suspicious features, AACE suggests FNA even in subcentimeter nodules in the event of suspicion of ETE: 'Due to the nonnegligible prevalence of extracapsular extension in subcentimetric tumors (35), suspicious (high-risk) lesions <10 mm should receive FNA for early diagnosis and treatment when subcapsular (especially if associated with capsular abutment)' (AACE). In the system of the American Clinical Radiologist ETE is worth 3 points (19), which following the logic of this system means that the lesion belongs either to the most or the most suspicious but one subgroup. In contrast with the former systems, neither the TIRADS of the Korean Society of Radiologist (21) nor that of the European Thyroid Association (20) takes ETE into account in generating a score mainly because the US-based diagnostic criteria for ETE is somewhat subjective and has not been established yet. The handling of ETE might cause some confusions by using the KSTR TIRADS: on the one hand, the system does not involve this feature into generating the score and therefore indication of FNA based on the score and the size of the nodule. On the other hand, the Korean guideline takes ETE most seriously, because it suggests FNA even in the suspicion of ETE, moreover, in contrast with all other features, irrespectively of nodule' size (21). |
||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||
The TIRADS of the European Thyroid Association is although more consistent but seem to be less realistic: this system describes the presentation of ETE and the concerns about its usefulness but does not suggest FNA in the event of ETE (20).
|
|||||||||||||||||||||||||||||||||
The halo sign and TIRADS |
|||||||||||||||||||||||||||||||||
In the guideline of the American Thyroid Association (ATA) and the American College of Radiologist (ACR) the halo sign is not mentioned or omitted, ATA or ACR, respectively (18-19). The latter mentions that the halo has no discriminatory effect. The other three guidelines summarize the above-discussed findings in the literature (20-22). |
Neither of the TIRADS involves halo sign in the classification of thyroid nodules. The AACE is the only TIRADS which mentions the 'regular halo' among features characteristic of benign lesions (20). However, the presence of halo is not a prerequisite to categorize a lesion among the benign subgroup which means that even in this system, neither the lack or the absence of halo influences the classification of a nodule. |
||||||||||||||||||||||||||||||||
