Value of a BRAFV600E and lymphocyte subset-based nomogram for discriminating benign lesions from papillary thyroid carcinoma in C-TIRADS 3 and higher nodules.

BACKGROUND

The BRAFV600E mutation and lymphocyte subsets may be associated with papillary thyroid carcinoma (PTC). This study established and validated a nomogram model to quantitatively predict the malignant risk of papillary thyroid carcinoma in thyroid nodules classified as C-TIRADS category 3 or higher, providing a reference for precise diagnosis and treatment of these moderately or highly suspicious nodules.

METHODS

This retrospective study analyzed 210 patients with thyroid nodules (C-TIRADS ≥3), stratified by fine-needle aspiration biopsy (FNAB) results into benign and PTC groups. Clinical and laboratory parameters were systematically collected for all patients. Variable selection was performed using least absolute shrinkage and selection operator (LASSO) regression, with multicollinearity assessed using variance inflation factors (VIF < 5). Subsequently, significant predictors were incorporated into a multivariate nomogram. Binary logistic regression analysis was employed to identify independent risk factors for PTC following adjustment for potential confounding variables. Internal validation was performed using bootstrap resampling (1,000 iterations) to assess the model's predictive accuracy, clinical utility, and discriminative ability. Comparative analysis was conducted against the conventional C-TIRADS classification system to evaluate relative performance.

RESULTS

Significant differences were observed between benign thyroid nodules and PTC in age, BRAFV600E genotype, natural killer (NK) cell counts, NK cell percentages, CD4+ T cell percentages, and ultrasound characteristics including size, echogenicity, composition, boundary, and morphology (P < 0.05). Five variables, including age, BRAFV600E genotype, NK cell counts, NK cell%, and CD4+ T cell%, were selected through LASSO regression with collinearity diagnostics for nomogram construction. The model demonstrated excellent discrimination (AUC=0.861, C-index=0.861), good calibration (Hosmer-Lemeshow χ²=6.72, P=0.57), and superior accuracy compared to random prediction (Brier score=0.1061, P<0.05). Decision curve analysis confirmed its clinical utility across relevant probability thresholds. Finally, the comparative analysis demonstrated superior diagnostic performance of the novel nomogram relative to the C-TIRADS system (AUC: 0.862 vs. 0.752; P<0.01).

CONCLUSION

The BRAFV600E-lymphocyte subset nomogram demonstrates robust clinical utility for discriminating benign lesions from PTC in C-TIRADS 3+ thyroid nodules, offering superior diagnostic performance to conventional risk stratification systems.