Gender, Age, and Ethnic Differences in Ultrasound Predictors of Malignancy Risk in Thyroid Nodules with Indeterminate Low-Risk Cytology (TIR3A): A Systematic Literature Review

Background: Thyroid nodules classified as TIR3A (Italian Reporting System) or Bethesda III (atypia of undetermined significance/follicular lesion of undetermined significance) represent a diagnostic challenge with malignancy rates of 10–30%. While ultrasound features guide risk stratification, the influence of patient demographics—particularly gender, age, and ethnicity—on ultrasound predictors of malignant transformation remains incompletely characterized. Methods: We conducted a systematic literature review of studies evaluating ultrasound features in TIR3A/Bethesda III thyroid nodules, with specific focus on gender differences (male vs. female), age-related patterns, and ethnic variations (Caucasian vs. non-Caucasian). We extracted data on malignancy rates, odds ratios, relative risks, and diagnostic performance metrics (sensitivity, specificity, positive predictive value, negative predictive value, area under the curve) for individual ultrasound alarm signs stratified by demographic variables. A total of 20 studies were included (17 retrospective cohort studies, 3 prospective cohort studies); findings were synthesized qualitatively. Results: Male sex was reported as a risk modifier in small cohorts (crude relative risk 2.3; independent predictor status in one multivariate analysis without a reported adjusted odds ratio), but larger series found no significant sex effect; adjusted estimates are unavailable, and the association should be considered inconsistent and of uncertain magnitude. Age below 55 years was associated with higher malignancy rates and more aggressive ultrasound features in heterogeneous single-centre observational data; however, this threshold has not been validated for ultrasound-based malignancy prediction across populations. Direct Caucasian versus non-Caucasian comparative data are absent; observed geographic variation cannot be attributed to ethnicity given multiple confounders including iodine intake and healthcare system differences. Microcalcifications (sensitivity 32.4–87%, specificity 75–99.2%) and irregular margins (sensitivity 70%, specificity 94.9–95%) remain the most robust ultrasound predictors across demographic subgroups. Conclusions: Gender and age show associations with malignancy risk in TIR3A nodules, but these associations are inconsistent across cohorts and should not be treated as validated independent predictors. Ethnic comparative data remain insufficient. Optimal risk stratification should integrate demographic variables with ultrasound features and repeat cytology to personalize management and reduce unnecessary surgery.