鼻咽癌患者淋巴结阳性高风险预测：一种多模型方法

Identifying patients at high risk of an elevated lymph node ratio (LNR) is critical for optimizing the management of nasopharyngeal carcinoma (NPC), as LNR, defined as the ratio of metastatic to examined lymph nodes, serves as a key prognostic indicator. This retrospective observational study aimed to investigate the epidemiology and influencing factors associated with high LNR in NPC patients. Various machine learning algorithms were employed to select independent predictive variables, and both univariate and multivariate Cox regression analyses were conducted to develop predictive models. The performance of different models was evaluated using receiver operating characteristic curves, calibration plots, and decision curve analysis, and nomograms and survival curves were constructed to facilitate visualization and clinical interpretation. A total of 1563 NPC patients were included in the study. The optimal model demonstrated an area under the curve of 0.73 (95% confidence interval: 0.67-0.78) in the modeling group and 0.76 (95% confidence interval: 0.70-0.81) in the validation group. The nomogram identified N stage, M stage, type of surgery, race, and confirmation status as independent risk factors for high LNR. Survival curve analysis further indicated that patients classified as high-risk by the nomogram had significantly worse outcomes. These findings suggest that elevated LNR is strongly associated with adverse prognosis in NPC patients. The constructed nomogram serves as a practical clinical tool to stratify patients based on LNR risk, thereby enabling personalized follow-up, treatment planning, and management strategies to optimize patient outcomes.