Methodological Considerations in Comparing Ultrasound and SPECT/CT for Lymph Node Metastasis in Papillary Thyroid Carcinoma

To the Editors:

I read with great interest the article by Upadhyaya et al ¹ recently published in the Journal of Ultrasound Medicine. The authors provided a meticulously designed retrospective cohort study comparing the detection rates of ultrasound (US) and post‐therapeutic I‐131 Single‐Photon Emission Computed Tomography/Computed Tomography (SPECT/CT) for cervical lymph node metastasis in 172 post‐thyroidectomy patients with papillary thyroid carcinoma (PTC). The study's detailed mapping of imaging profiles, particularly the identification of a substantial subset (32.7%) of radioactive iodine (RAI)‐refractory disease that is US‐positive but SPECT/CT‐negative, offers highly valuable clinical insights for optimizing postoperative surveillance strategies. The authors also elegantly mapped the distribution of these nodes, highlighting Level VI as the most frequently involved nodal station. However, from a methodological and statistical perspective, I would like to respectfully raise 2 primary considerations regarding potential confounding factors that might inherently skew the comparative diagnostic outcomes.

First, I am concerned about a potential selection bias introduced by the reference standard pathway, which may disproportionately affect the calculated detection rates. The authors explicitly stated that their design eliminates verification bias by focusing exclusively on pathologically proven metastatic nodes. However, according to the methodology, 72.2% (192 of 266) of the lymph nodes in the cohort were confirmed via US‐guided fine‐needle aspiration cytology (US‐FNAC) or thyroglobulin washout. In routine clinical practice, the decision to perform an invasive biopsy is predominantly triggered by suspicious sonographic morphological findings. ² , ³ Consequently, metastatic nodes that are US‐negative but functionally SPECT/CT‐positive possess an inherently lower probability of being targeted for biopsy, thereby losing the opportunity to be included in this pathologically confirmed cohort. This inherent diagnostic workflow naturally creates a survival bias, enriching the final cohort with US‐visible lesions. As a result, the reported 75.0% detection rate for US may be inadvertently overestimated, while the 55.8% detection rate for SPECT/CT might be artificially penalized. The subgroup of US‐negative/SPECT/CT‐positive metastases (13.5%, 36 nodes) might be vastly underrepresented in the broader clinical population.

Second, the direct head‐to‐head statistical comparison of overall detection rates may be fundamentally confounded by the physical resolution disparity between the 2 modalities, particularly when applied to a cohort containing micro‐metastases. The authors reported that 38 lymph nodes (14.3%) measured less than 5 mm in short‐axis diameter, and 55.3% of these sub‐centimeter nodes were correctly classified by US, primarily due to the detection of microcalcifications. High‐frequency US possesses exceptional spatial resolution, capable of identifying punctate hyperechoic foci as small as 1 mm. In stark contrast, while I‐131 SPECT/CT is the gold standard for metabolic activity, its detection capability for sub‐centimeter lesions is physiologically restricted by inherent spatial resolution limits and partial‐volume effects, a limitation the authors also acknowledged regarding the low‐dose CT component. Furthermore, the functional avidity is dependent on the therapeutic dose administered (100–200 mCi) and tissue differentiation. Evaluating and directly comparing overall detection rates in a cohort containing a significant proportion of micro‐lesions inherently disadvantages SPECT/CT based on immutable physical properties rather than comparative clinical utility.

Despite these methodological nuances, I highly commend Upadhyaya et al for their rigorous multidisciplinary integration and robust demonstration of US's crucial role in detecting anatomically altered, RAI‐refractory recurrent disease. Their work undoubtedly advances our understanding of multimodal imaging dynamics. To further refine these comparative paradigms and mitigate selection bias in the future, prospective studies could employ a composite reference standard that integrates standardized, independent triggers for biopsy (eg, performing CT‐guided or subsequent targeted US evaluations specifically for SPECT/CT‐positive/US‐negative hotspots) or utilizing long‐term biochemical and radiological follow‐up for lesions unamenable to immediate biopsy. Such designs would build beautifully upon the foundational evidence established by this study.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.