Multimodal ultrasound-based nomogram for predicting occult cervical lymph node metastasis in cN0 papillary thyroid carcinoma

Occult cervical lymph node metastasis (CLNM) remains a clinically relevant issue in clinically node-negative (cN0) papillary thyroid carcinoma (PTC), while its preoperative prediction remains challenging. This study aimed to develop and internally validate a multimodal ultrasound-based nomogram for individualized prediction of occult CLNM in patients with cN0 PTC.

This retrospective study included 204 patients with pathologically confirmed cN0 PTC who underwent preoperative conventional ultrasonography (US), superb microvascular imaging (SMI), shear wave elastography (SWE), and contrast-enhanced ultrasound (CEUS) between January 2020 and April 2022. Patients were randomly assigned to training (n = 143) and validation (n = 61) cohorts. Clinical characteristics and multimodal ultrasound features were analyzed. Candidate predictors were screened using univariate logistic regression and least absolute shrinkage and selection operator (LASSO) regression. Conventional ultrasound and multimodal ultrasound nomograms were constructed and evaluated using receiver operating characteristic analysis, calibration curves, decision curve analysis (DCA), interobserver agreement analysis, and bootstrap internal validation.

The multimodal ultrasound nomogram incorporated 12 predictors derived from conventional US, SWE, and CEUS features. The model achieved bootstrap-corrected AUCs of 0.849 (95% CI: 0.776-0.922) and 0.718 (95% CI: 0.559-0.876) in the training and validation cohorts, respectively, outperforming the conventional ultrasound model. DCA demonstrated greater net clinical benefit of the multimodal model across a broader range of threshold probabilities. Interobserver agreement for included imaging features showed good reproducibility.

The multimodal ultrasound nomogram, integrating peritumoral stiffness heterogeneity and CEUS perfusion parameters, showed acceptable performance for predicting occult CLNM in patients with cN0 PTC, providing incremental value compared with the conventional ultrasound model in this single-center cohort. This model may serve as a potential tool for preoperative risk stratification; however, further external validation in larger multicenter studies is required before clinical application.