Dynamic Predictors for Malignancy Risk of Follicular Thyroid Neoplasms: A Real‐World, 10‐Year Cohort Study

OBJECTIVE: The malignancy risk of follicular thyroid neoplasms (FTN) is variable, yet most studies rely on static, single-time-point assessments without monitoring dynamic changes. METHODS: We included patients who received two or more preoperative ultrasound examinations and were diagnosed with follicular thyroid carcinoma (FTC) or follicular thyroid adenoma (FTA) after surgery. We first identified 11 key predictors for the malignancy risk of FTN, comprising five static predictors (age, sex, body mass index, baseline mean tumor diameter, and surveillance duration) and six dynamic predictors [average thyroid-stimulating hormone (TSH) level, TSH instability, weighted rate and instability of mean diameter, and changes in tumor margin and calcifications over time]. Then, we applied logistic regression to evaluate the association between each dynamic predictor and the malignancy risk of FTN, adjusting for the static predictors as potential confounders. RESULTS: The results showed that a higher average TSH level was associated with a decreased risk of malignancy in FTN [odds ratio (OR): 0.67; 95% confidence interval (CI): 0.49, 0.92]. However, a shift in margin from circumscribed to irregular or maintaining an irregular margin was associated with a higher risk of malignancy [OR (95% CI): 4.18 (1.49, 11.71) and 5.33 (2.51, 11.35)]. We observed no association between the rate or instability of the mean tumor diameter and the malignancy risk of FTN. CONCLUSION: Changes in tumor margins (to irregular or lobulated) and lower average TSH levels were key predictors of a higher malignancy risk, while the rate or instability of tumor growth might play a minor role.