列线图术前预测早期黏膜下右半结肠癌淋巴结转移:双中心回顾性队列研究
本研究纳入628例T1右半结肠癌患者,通过logistic回归识别淋巴结转移的独立术前预测因素并构建列线图。模型训练集AUC为0.84、验证集为0.83,校准曲线与决策曲线分析显示良好区分度与临床净获益。该模型可术前量化淋巴结转移风险,辅助个体化治疗决策。
要点
- 年龄、肿瘤位置、分化程度、CT淋巴结肿大、CEA及CA19-9为独立预测因素。
- 列线图训练集AUC 0.84、验证集AUC 0.83,校准良好。
- 模型可辅助筛选低危患者避免不必要的结肠切除。
核心结论
- 基于常规术前参数构建的列线图可准确量化T1右半结肠癌淋巴结转移风险,指导个体化治疗并避免低危患者不必要的结肠切除。
研究设计与样本
- 双中心回顾性队列研究,纳入628例T1右半结肠癌患者,采用单因素和多因素logistic回归识别独立术前预测因素并构建列线图。
主要发现
- 年龄(OR=0.92,95%CI:0.88-0.97)、肿瘤位于肝曲(OR=3.50,95%CI:1.68-7.31)、低分化肿瘤(OR=4.21,95%CI:2.23-7.94)、CT检出淋巴结肿大(OR=5.14,95%CI:2.79-9.46)、CEA升高(OR=3.26,95%CI:1.77-6.01)及CA19-9升高(OR=1.98,95%CI:1.04-3.77)为独立预测因素。
- 模型训练集AUC为0.84,验证集AUC为0.83。
- 校准曲线显示预测与观察结果高度一致,Hosmer-Lemeshow检验无统计学意义。
- 决策曲线分析显示在多个阈值概率下具有稳健的临床净获益。
临床怎么用
- 术前可利用年龄、肿瘤位置、分化程度、CT淋巴结状态、CEA及CA19-9评估T1右半结肠癌淋巴结转移风险。
- 列线图可用于辅助选择内镜切除或结肠切除,低风险患者或可避免不必要的结肠切除。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
BACKGROUND: Preoperative lymph node metastasis (LNM) risk stratification is crucial for selecting endoscopic resection or colectomy in T1 right-sided colon cancer (T1 RCC), yet validated tools integrating preoperative predictive indicators remain lacking. METHODS: This retrospective study included 628 patients with T1 RCC. Single-predictor and multi-predictor logistic regression models identified independent preoperative predictors of LNM to construct a nomogram. Model performance was assessed using area under the curve (AUC) and calibration metrics. RESULTS: Six preoperative factors constituted independent predictors of LNM: age (OR = 0.92, 95% CI: 0.88-0.97), tumor location (hepatic flexure: OR = 3.50, 95% CI: 1.68-7.31), poorly differentiated tumor (OR = 4.21, 95% CI: 2.23-7.94), CT-detected lymphadenopathy (OR = 5.14, 95% CI: 2.79-9.46), elevated CEA (OR = 3.26, 95% CI: 1.77-6.01), and elevated CA 19-9 (OR = 1.98, 95% CI: 1.04-3.77). This predictive model demonstrated high discriminatory ability: 0.84 AUC in training preserved at 0.83 upon validation. Calibration curves demonstrated optimal agreement with observed outcomes, supported by a non-significant Hosmer-Lemeshow test. Decision curve analysis further established robust clinical net benefit across threshold probabilities. CONCLUSION: This preoperative predictive model accurately quantifies the risk of LNM in T1 RCC using routine parameters, guiding individualized treatment decisions and preventing unnecessary colectomy in low-risk patients.
引用
- Han F, Liu Y, Wang Y, et al. Nomogram for preoperative prediction of lymph node metastasis in patients with early submucosal right-sided colon cancer: a dual-center retrospective cohort study. Surg Endosc. 2026. PMID: 42151464. doi:10.1007/s00464-026-12889-0
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。