Abstract
W
In this study, ROM was defined only among nodules with histopathological confirmation, rather than among all nodules. However, citing the BSRTC (2), the authors noted in a prior study that this method of calculating ROM raises “concern for overestimating the ROM, particularly for the ND, benign, and AUS/FLUS categories due to selection bias given the relatively small proportion of nodules that undergo resection” in adults (3). Accounting for this bias is crucial when comparing ROM between pediatric and adult nodules because lower resection rates in adults lead to greater overestimation of the ROM in adults than in children, potentially obscuring true differences in ROM. Although the authors addressed this issue in their adult meta-analysis (3), they did not do so when comparing children with adults, when this bias is even more relevant. On our simplified reanalysis of the data presented, the ROM among all atypia of undetermined significance/follicular lesion of undetermined significance (AUS/FLUS) nodules was 2.5-fold higher in children (23.2% vs. 9.2%, p < 0.00001, χ 2 analysis), and it was twofold higher among follicular neoplasm/suspicious for a follicular neoplasm (FN/SFN) nodules (35.1% vs. 17.5%, p < 0.00001). Of course, the actual ROM lies between the values calculated among all nodules and among resected nodules (2), but given the authors' emphasis on the discrepant resection rates in children and adults, failing to account for this bias is an important oversight.
As the authors note, differences between pediatric and adult nodules might also be masked by heterogeneity among the included studies. Additional heterogeneity is introduced by comparing pediatric and adult studies from different institutions, particularly given the variable interpretation of indeterminate cytology (4) and potentially limited thyroid cytopathology expertise among pediatric pathologists. The authors state that these limitations could be addressed by obtaining “data on adult and pediatric patients from the same institutions” but that “it is not practical to perform such an analysis.” In fact, such an analysis was performed and was reported in one of the largest studies included in this meta-analysis (5). In this single-center study, designed specifically to address the limitations that this meta-analysis cannot, significant differences in ROM were observed between children and adults among all AUS/FLUS and FN/SFN nodules, whereas the lack of difference in ROM among resected nodules was attributable to lower resection rates among adults.
In summary, although this meta-analysis faithfully reflects the published data on pediatric thyroid cytopathology, we believe that its comparison of pediatric and adult data ignores a critical bias and, therefore, does not support the stated conclusion that pediatric and adult ROMs are equivalent within cytological categories. Careful attention to these issues is necessary in future studies to avoid drawing potentially erroneous parallels between children and adults.
Footnotes
Author Disclosure Statement
No competing financial interests exist.
Funding Information
No funding was received for this article.