Systematic differences between ultrasound and pathological evaluation of thyroid nodules: a method comparison study.

Abstract

The size of thyroid nodules as measured by ultrasound (ultrasound size, USS) is routinely used in clinical decision-making. Reports of discrepancy between USS and pathological size (PS) evaluation have not analysed their systematic differences. The objective of this study was to uncover the lack of agreement (bias) between USS and PS measurements. A retrospective study was performed on 121 patients who had a total or hemi-thyroidectomy for a solitary nodule. Ordinary least product regression was used to detect and distinguish constant and proportional bias in unidimensional size measurements between USS and PS evaluation. Three-dimensional volume measurements were compared in a subgroup of 31 patients. Pre-specified acceptable limits of interchange were defined as 20% difference. Ordinary least product regression demonstrated no constant or proportional bias between the two methods; regression equation: USS = (0.863) + (1.040) × PS. When nodules were grouped by size, discrepancies between the two methods were observed in nodules <10 mm (P = 0.004). However, potential overtreatment of patients with USS >10 mm but PS <10 mm only accounted for 4.1% of total patients. Subgroup analysis of volume measurements showed no bias between USS and PS evaluation. USS and PS measurements were interchangeable, as there was no evidence of constant or proportional bias between the two measurements. However, USS may misclassify the size for smaller nodules and potentially lead to unnecessary workup and treatment. Discrepancy in size measurements between USS and PS should be taken into account in clinical practice, particularly in smaller nodules.