September 2019 HSC Section 1 Congenital and Pediatric Problems

Canfarotta et al

Cytology FNAB has been shown to be a sensitive and highly specific tool in the clinical management of pediatric thyroid nodules. 33 Indeterminate FNAB results, characterized as AUS (Bethesda III) or SFN (Bethesda IV), continue to pose the greatest challenge in deciding when to intervene surgi- cally. The malignancy rate for adults with Bethesda III and IV nodules ranges from 5%-15% to 15%-30%, respec- tively. 17 The risk of indeterminate nodules has been shown to be even higher for children, with malignancy rates for Bethesda III and IV nodules ranging from 18%-28% to 50%-100%, respectively. 14,19,34-36 Interestingly, our sample population showed a much lower malignancy rate than that seen for adults, with 5.6% of AUS and 25% of SFN nodules diagnosed as malignant on final pathology. A more recent study also showed lower malignancy rates of AUS and SFN nodules in a pediatric population—8.3% and 10%, respec- tively. 37 Multigene expression classifiers have been devel- oped to identify benign disease in indeterminate nodules in adults 38 ; however, there is insufficient evidence to recom- mend the use of such an instrument in the pediatric popula- tion at this time. 14 Given the lack of supportive tools to better risk stratify the indeterminate nodules in pediatric populations, the cur- rent ATA guidelines recommend surgical removal of nodules with indeterminate cytology. 14 Tools such as the MTNS and molecular testing may complement the results of FNAB and be useful in the future for determining extent of surgery, treatment, and other evaluations. Furthermore, a clinical tool could provide a more cost-effective model in the management of these nodules. Recently, Varshney et al showed that the MTNS can be useful in the preoperative risk stratification of Bethesda III and IV nodules in adults. 39 Our study suggests a similar utility for children, as indeter- minate nodules with malignant histology tend to have a higher MTNS on average than their benign counterparts. The major limiting factor to this study and others that attempt to estimate the risk for malignancy in indeterminate nodules is the very small sample sizes, which are difficult to avoid given the rarity of the disease. Regardless, these findings highlight the uncertainty in the management of these FNAB categories and emphasize the need to elucidate the role of adjuvant tools, such as the MTNS and molecular testing. Future research should focus on validation of this scoring system to provide a comprehensive clinical suspi- cion index for multidisciplinary teams to use in the discus- sion of patient management. We do not intend to propose a cutoff value to be used in isolation in the preoperative decision-making process but, rather, an accurate assessment tool that facilitates communication between the clinician and the family. Cumulative MTNS Our results show very high specificity and positive predictive value, as all nodules with cumulative MTNS . 11 were malig- nant on final pathology. Furthermore, the test is extremely

Figure 2. Distribution of cumulative McGill Thyroid Nodule Score (MTNS) score and final pathology for Bethesda III and IV nodules.

Table 4. Sensitivity, Specificity, and PPV of the Pediatric MTNS.

Score

Sensitivity, %

Specificity, %

PPV, %

10 11

100

94.4

83.3

80

100

100

Abbreviations: MTNS, McGill Thyroid Nodule Score; PPV, positive predic- tive value.

assumed to increase risk—hence, their inclusion in our modi- fied MTNS. However, future studies with larger cohorts are needed to elucidate the degree of risk. Ultrasonography Ultrasound has also proven to be a useful tool in the preopera- tive discrimination of benign from malignant nodular disease in children, and characteristics predictive of malignancy include microcalcifications, intranodular vascularization, hypoechoic pat- tern, nodule growth, abnormal lymph nodes, and dimensions taller than wide. 22,28 In our sample population, hypoechoic pat- tern, microcalcifications, and abnormal lymph nodes were asso- ciated with malignant disease. Shape was more likely to be taller than wide in malignant nodules but failed to reach a level of statistical significance. Our results further confirm many of the concerning ultrasound features associated with malignancy. There was no relationship with nodule size and cancer risk within our sample. The pediatric ATA guidelines acknowledge that size criterion is problematic, as thyroid volume changes with age and multiple studies have not identified an association with malignant histology. 14,29-32 However, more recent studies have found a correlation. 9,19 For this reason, we have kept size as a risk factor in our modified scoring system. In the future, as studies continue to define the degree to which there is an associ- ation of nodule size with malignancy, it is possible that further modifications in the pediatric MTNS will be necessary.

159