基于代谢综合征的列线图用于预测局部晚期鼻咽癌放疗后1年转移的开发和验证
Development and validation of a metabolic syndrome-based nomogram to predict 1-year metastasis after radiotherapy in locally advanced nasopharyngeal carcinoma.
文献信息
| PMID | 42860552 |
|---|---|
| 原文 | 在 PubMed 查看原文 ↗ |
| 发表日期 | 2026 |
| 作者 | Yanyu Chen |
| 作者单位 | Department of Cancer Center, The Second Affiliated Hospital, Hengyang Medical School, University of South China, Hengyang, Hunan, China. |
| 期刊 | Annals of medicine |
| SCI 分区 | Q1 |
| IF | 5.2 |
| 研究类型 | 临床研究 · 临床 |
| 所属专科 | 鼻咽癌 |
中文摘要
背景： 代谢综合征（MetS）可预测多种恶性肿瘤的结局。本研究开发了一种整合MetS的列线图，用于估计局部晚期鼻咽癌（LA-NPC）放疗后一年内的转移，扩展了现有模型并促进个体化管理。
方法： 在这项回顾性单中心研究中，1333例接受基于放疗的多模式治疗的LA-NPC患者被随机分为训练队列（n = 888）和验证队列（n = 445）。采用单因素和多因素logistic回归确定列线图的独立转移危险因素。评估包括C指数、校准图、概率密度函数（PDFs）和临床效用曲线（CUCs）。使用DeLong检验对受试者工作特征（ROC）分析进行区分度评估。
结果： 训练队列中有56例（6.3%）、验证队列中有28例（6.3%）患者在一年内发生转移。治疗前T分期、N分期、MetS、血浆EBV DNA、乳酸脱氢酶（LDH）、C反应蛋白（CRP）和吸烟与转移相关；多因素分析证实MetS和EBV DNA为独立预测因素。纳入T分期、N分期、MetS和EBV DNA的列线图在训练队列和验证队列中的C指数分别为0.763（95% CI：0.694-0.832）和0.715（95% CI：0.629-0.801），优于单一预测因素（均P < 0.05）。PDF和CUC分析提示8.7%的阈值可优化真阳性和真阴性的识别，正确分类82.0%的非转移患者和39.3%的转移患者。
结论： 这种基于MetS的列线图能够对LA-NPC放疗后一年内转移进行个体化风险分层。8.7%的阈值具有探索性，但在外部验证后有望用于识别高危患者以接受进一步治疗或监测。
英文摘要
BACKGROUND: Metabolic syndrome (MetS) predicts outcomes in multiple malignancies. This study developed a MetS-integrated nomogram to estimate metastasis within one year after radiotherapy in locally advanced nasopharyngeal carcinoma (LA-NPC), extending existing models and facilitating personalized management.
METHODS: In this retrospective single-center study, 1333 LA-NPC patients undergoing radiotherapy-based multimodal therapy were randomly divided into training (n = 888) and validation (n = 445) cohorts. Univariate and multivariate logistic regression identified independent metastasis risk factors for the nomogram. Evaluation included C-index, calibration plots, probability density functions (PDFs), and clinical utility curves (CUCs). Discrimination was assessed using DeLong test for receiver operating characteristic (ROC) analysis.
RESULTS: Metastasis within one year occurred in 56 (6.3%) training and 28 (6.3%) validation patients. Pretreatment T stage, N stage, MetS, plasma EBV DNA, lactate dehydrogenase (LDH), C-reactive protein (CRP), and smoking were associated with metastasis; multivariate analysis confirmed MetS and EBV DNA as independent predictors. The nomogram incorporating T stage, N stage, MetS, and EBV DNA achieved C-indexes of 0.763 (95% CI: 0.694-0.832, training) and 0.715 (95% CI: 0.629-0.801, validation), surpassing single predictors (all P < 0.05). PDF and CUC analyses suggested an 8.7% threshold to optimize the identification of true positives and negatives, classifying 82.0% non-metastatic and 39.3% metastatic patients correctly.
CONCLUSION: This MetS-based nomogram enables individualized risk stratification for metastasis within one year after radiotherapy in LA-NPC. The 8.7% threshold is exploratory but promising for identifying high-risk patients for further therapy or surveillance after external validation.