Construction and validation of novel nomograms based on the log odds of positive lymph nodes to predict the prognosis of papillary thyroid cancer: a retrospective cohort study.

Abstract

Objective: This study aims to assess the long-term prognostic significance of the log odds of positive lymph nodes (LODDS) in patients diagnosed with papillary thyroid cancer (PTC) and to develop a novel nomogram for predicting long-term overall survival (OS).

Methods: The cohort was randomly divided at a ratio of 7:3 from the Surveillance, Epidemiology, and End Results (SEER) database. Additionally, patient data from a medical center in China served as an external validation cohort. Nomograms were constructed using data from the training cohort and subsequently validated using both internal and external validation cohorts to predict 120- and 180-month OS in PTC patients. The predictive performance and clinical utility of the nomogram were assessed using various metrics, including the concordance index (C-index), time-dependent receiver operating characteristic (ROC) curves, calibration curves, decision curve analysis (DCA), Integrated Discriminant Improvement Index (IDI), and Net Reclassification Improvement Index (NRI).

Results: LODDS is an independent prognostic factor for PTC, a nomogram demonstrating high accuracy in predicting long-term OS. The C-index values, and time-dependent area under the curve (AUC) indicated well discriminatory ability of the nomogram. Calibration plots exhibited high concordance, while DCA, NRI, and IDI analyses revealed superior performance of the nomogram compared to AJCC staging system.

Conclusion: The clinical prediction model incorporating LODDS exhibits robust predictive performance, aiding in the assessment of long-term prognosis post-surgery in PTC patients. It serves as a valuable adjunct to the AJCC system, offering a scientific basis for guiding interventions and rehabilitation strategies for PTC patients following surgery.