Predictors of Malignancy and False‐Negative Risk in Large (≥ 3 cm) Thyroid Nodules: Insights From a Single‐Centre Thyroidectomy Cohort

BACKGROUND: Whether nodule size compromises cytologic reliability remains controversial in managing large thyroid nodules. This study evaluated malignancy prevalence, predictors and false-negative (FN) cytology in a surgical cohort enriched for nodules ≥ 3 cm. METHODS: We retrospectively analyzed 285 thyroidectomy patients (2018-2024) with a ≥ 3 cm dominant nodule. Each patient represented one analytic unit. The study assessed malignancy rates, clinical and imaging predictors, and FN risk among cytologically benign (Bethesda II) nodules. Demographic, biochemical and sonographic data-including nodule size, composition, EU-TIRADS category and suspicious lymphadenopathy (LAP)-were retrieved from records. Multivariable logistic regression with bootstrap validation identified independent predictors of malignancy and factors linked to false negativity. RESULTS: Observed malignancy proportion was 13.3% in this thyroidectomy cohort (38/285), predominantly papillary carcinoma. Independent predictors: EU-TIRADS TR4-5 (OR 7.6) and suspicious LAP (OR 6.0); size ≥ 4 cm (OR 2.6) and solid composition (OR 3.3) exerted smaller effects. Among 111 Bethesda II cases, postoperative malignancy occurred in five (4.5%), all in ≥ 4 cm nodules (7.5% vs. 0%; p = 0.023). Post hoc review of FN cases suggested cystic/infiltrative features potentially contributing to sampling limitations. Nodule size had a moderate association with FN probability (AUC 0.75). CONCLUSIONS: In this surgically selected cohort of large thyroid nodules, sonographic risk, suspicious LAP and clinical context outperformed size alone in predicting malignancy. Even with benign cytology, nodules ≥ 4 cm with higher-risk imaging or structural features may warrant repeat or targeted FNAB, core biopsy, or surgery, whereas lower-risk phenotypes may still be considered for surveillance. These findings support a risk-stratified, node-informed approach beyond purely size-driven surgery in surgically referred large nodules, although extrapolation to all large nodules in the general population should be made cautiously.