Response to Cherella et al. re: “The Use of the Bethesda System for Reporting Thyroid Cytopathology in Pediatric Thyroid Nodules: A Meta-Analysis”

Dear Editor:

We thank Cherella et al. (1) for their interest in our recently published meta-analysis on pediatric thyroid nodules (2), where the authors critically reappraised our findings with an alternative metric, the overall risk of malignancy (oROM) (3), and also referred to their own institutional experience (4).

We agree that oROM (i.e., a ratio of malignant nodules to all aspirated nodules) could disclose certain differences not recognized by the conventional ROM (malignant nodules out of operated nodules) statistics. To address this, we provide oROM calculations in Table 1. Indeed, as it was estimated by Cherella et al., the oROM in the atypia of undetermined significance/follicular lesion of undetermined significance (AUS/FLUS) and follicular neoplasm/suspicious for follicular neoplasm (FN/SFN) of pediatric patients were significantly higher than those in adults (21.5% vs. 9.2% and 36.9% vs. 17.1%; p < 0.001). However, such statistical significance was not seen when the conventional ROM was used. While the ROM is an essential metric in any cytology reporting system, including the Bethesda System for Reporting Thyroid Cytopathology (TBSRTC) (5), the utility of oROM is not well acknowledged at this time. In fact, the vast majority of existing publications about thyroid nodules do not utilize the oROM terminology, even when raw data are available for its calculation (6).

An important factor that could affect the oROM estimation is the application of active surveillance for adults with indeterminate thyroid nodules, particularly in Asian countries, which could lower the oROM to a greater extent in the adult group. Another concern that could affect the oROM calculation is the dropout rate during follow-up, which could be higher in the adult population due to having multiple choices of available treatment facilities. Finally, the absolute difference of oROM and ROM might be remarkable (up to 11-fold; Table 1), resulting in difficulty to estimate the actual ROM, which is supposed to be “in the midrange” between the oROM and ROM (5).

In their comment, Cherella et al. specifically put a focus on indeterminate thyroid nodules (AUS/FLUS and FN/SFN) and, therefore, overlooked one of our major conclusions that cytologically benign pediatric nodules have significantly higher resection rate compared with adults despite similar ROMs (2). This finding has been now reinforced by the oROM metric—our results showed comparable oROM in the nondiagnostic (2.3% vs. 2.0%, p = 0.793) and benign categories (0.9% vs. 0.7%, p = 0.365) of pediatric versus adult population (Table 1). Interestingly, in their own series, Cherella et al. also found that pediatric thyroid nodules with benign cytology had a higher resection rate (28% vs. 14%) but with no significant difference in the ROM (3% vs. 7%; p = 0.11) and a significantly lower oROM (0.7% vs. 1%; p = 0.001) in comparison with the adult population (4). It should be noted that benign nodules account for 55–60% of thyroid nodules in both adult and pediatric populations (2) and such an increased resection rate in children should be carefully investigated.

We appreciate the comments offered by Cherella et al. and bringing attention to the largely unrecognized significance of oROM metrics. At the same time, we stand by our conclusion that nonmalignant pediatric thyroid nodules, especially those with benign cytology, might be exposed to a potential risk of overtreatment.