Abstract / Summary
Thyroid nodules (TNs) are prevalent, particularly in women and older adults, yet only a minority are malignant. Clinical decisions on biopsy and follow-up are complicated by selective use of fine-needle aspiration cytology, indeterminate results, and efforts to limit unnecessary procedures. We systematically identified and synthesized sociodemographic, laboratory, and ultrasonographic (US) factors associated with malignancy in TNs. PubMed, Scopus, Web of Science, and EMBASE were searched (PROSPERO CRD42024557085). Two reviewers independently screened studies, extracted data, and assessed risk of bias (Joanna Briggs Institute tools). Random-effects meta-analyses were performed for predictors reported in ≥5 studies, with heterogeneity quantified using I2. Pooled odds ratios (ORs) with 95% confidence intervals (CIs) were calculated. Malignancy was defined according to each study's reference standard (histology and/or cytology). Sensitivity analyses assessed robustness across diagnostic method and key study-level characteristics (e.g. geographic region and sample size). Of 35,359 records, 194 studies met the inclusion criteria. Of these, 47 with poolable crude ORs (predictors reported in ≥5 studies) contributed to meta-analysis, and the remainder to qualitative synthesis. Male sex, family history of thyroid cancer, higher thyroid-stimulating hormone levels, and several US features (including hypoechogenicity, solid composition, calcifications/microcalcifications, taller-than-wide shape, and irregular margins) were associated with higher odds of malignancy, whereas increasing age and larger nodule size showed inverse associations, though with substantial between-study heterogeneity. Malignancy in TNs reflects both patient-related and nodule-specific characteristics. Integrating these factors may support pre-biopsy risk assessment and more consistent biopsy and follow-up decisions. https://www.crd.york.ac.uk/PROSPERO/, identifier CRD42024557085.
Topics
Primary Source
Frontiers in endocrinology
Ask Prognia AI
Have questions about this meta-analysis?
Prognia AI can search this source alongside 35M+ PubMed papers and current ESC, AHA, NICE, and ADA guidelines to give you a fully cited clinical answer.