Abstract
PURPOSE: We aimed to establish a prediction model based on preoperative clinicopathologic features and intraoperative frozen section examination for precise prediction of metastatic involvement of lymph nodes posterior to the right recurrent laryngeal nerve (LN-prRLN) in patients with papillary thyroid carcinoma (PTC). METHODS: Clinicopathologic data pertaining to patients with PTC who underwent initial thyroid surgery between July 2015 and December 2017 were collected from electronic medical records. Multivariate logistic regression analysis was performed to identify independent predictors of LN-prRLN metastasis for incorporation in the nomogram. The performance of the model was assessed using discriminative ability, calibration, and clinical application. RESULTS: A total of 592 patients were enrolled in this study. The LN-prRLN metastatic positivity was 19% (95% confidence interval [CI], 15.61-21.89%). On multivariate analysis, ultrasonography-reported LN status, extrathyroid extension, Delphian lymph node metastasis, and number of metastatic pretracheal and paratracheal lymph nodes were independent predictors of LN-prRLN metastasis. The nomogram showed good discriminative ability (C-index: 0.87; [95% CI, 0.84-0.91]; bias-corrected C-index: 0.86 [through bootstrapping validation]) and was well calibrated. The decision curve analysis indicated potential clinical usefulness of the nomogram. CONCLUSION: This study demonstrates that the risk of LN-prRLN metastasis in individual patients can be robustly predicted by a nomogram that integrates readily available preoperative clinicopathologic features and intraoperative frozen section examination. The nomogram may facilitate intraoperative decision-making for patients with PTC.