Ultrasonography should not guide the timing of thyroidectomy in pediatric patients diagnosed with multiple endocrine neoplasia syndrome 2a through genetic screening

Background: American Thyroid Association (ATA) guidelines suggest that thyroidectomy can be delayed in some children with multiple endocrine neoplasia syndrome 2A (MEN2A) if serum calcitonin (Ct) and neck ultrasonography (US) are normal. We hypothesized that normal US would not exclude a final pathology diagnosis of medullary thyroid cancer (MTC). Methods: We retrospectively queried a MEN2A database for patients aged <18 years, diagnosed through genetic screening, who underwent preoperative US and thyroidectomy at our institution, comparing preoperative US and Ct results with pathologic findings. Results: 35 eligible patients underwent surgery at median age of 6.3 (range 3.0-13.8) years. Mean MTC size was 2.9 (range 0.5-6.0) mm. The sensitivity of a US lesion ≥5 mm in predicting MTC was 13 % [95 % confidence interval (CI) 2 %, 40 %], and the specificity was 95 % [95 % CI 75 %, 100 %]. Elevated Ct predicted MTC in 13/15 patients (sensitivity 87 % [95 % CI 60 %, 98 %], specificity 35 % [95 % CI 15 %, 59 %]). The area under the receiver operating characteristic curve (AUC) for using US lesion of any size to predict MTC was 0.50 [95 % CI 0.33, 0.66], suggesting that US size has poor ability to discriminate MTC from non-MTC cases. The AUC for Ct level at 0.65 [95 % CI 0.46, 0.85] was better than that of US but not age [AUC 0.62, 95 % CI 0.42, 0.82]. Conclusions: In asymptomatic children with MEN2A diagnosed by genetic screening, preoperative thyroid US was not sensitive in identifying MTC of any size and, when determining the age for surgery, should not be used to predict microscopic MTC.