Impact of the ultrasonography assessment method on the malignancy risk and diagnostic performance of five risk stratification systems in thyroid nodules

Abstract

Purpose Ultrasonographic (US) assessment methods may affect the estimated malignancy risk of thyroid nodules. This study aimed to investigate the impact of retrospective and prospective US assessments on the estimated malignancy risk of US features, classified categories, and diagnostic performance of five risk stratification systems (RSSs) in thyroid nodules. Methods A total of 3685 consecutive thyroid nodules (≥ 1 cm) with final diagnoses (retrospective dataset, n = 2180; prospective dataset, n = 1505) were included in this study. We compared the estimated malignancy risk of US features, classified categories, and diagnostic performances of the five common RSSs between retrospective (static US images without cine clips) and prospective datasets of real-time US assessment. Results There was no significant difference in the prevalence and histological type of malignant tumours between the two datasets (p ≥ 0.216). The malignancy risk of solid composition and nonparallel orientation was higher and that of microcalcification was lower in the prospective dataset than in the retrospective dataset (p < 0.001, p = 0.018, p = 0.007, respectively). The retrospective US assessment overestimated the malignancy risk of intermediate-or high-risk nodules according to the RSSs. Prospective US assessment showed lower specificities and higher unnecessary biopsy rates by all RSSs compared to the retrospective US assessment (p ≤ 0.006, p ≤ 0.045, respectively). Conclusions The overestimated malignancy risk of microcalcification by retrospective US assessment mainly affected the estimated risk of classified categories by RSSs. The retrospective US assessment overestimated the specificities and underestimated the unnecessary biopsy rates by all RSSs.