Ruan Guo-Tian, Song Meng-Meng, Zhang Kang-Ping, Xie Hai-Lun, Zhang Qi, Zhang Xi, Tang Meng, Zhang Xiao-Wei, Ge Yi-Zhong, Yang Ming, Zhu Li-Chen, Shi Han-Ping
Department of Gastrointestinal Surgery/Department of Clinical Nutrition, Beijing Shijitan Hospital, Capital Medical University, 10 Tie Yi Road, Beijing, 100038, China.
Key Laboratory of Cancer FSMP for State Market Regulation, Beijing, 100038, China.
Nutr Metab (Lond). 2023 Jan 4;20(1):2. doi: 10.1186/s12986-022-00719-8.
Precisely predicting the short- and long-term survival of patients with cancer is important. The tumor-node-metastasis (TNM) stage can accurately predict the long-term, but not short-term, survival of cancer. Nutritional status can affect the individual status and short-term outcomes of patients with cancer. Our hypothesis was that incorporating TNM stage and nutrition-related factors into one nomogram improves the survival prediction for patients with colorectal cancer (CRC).
This multicenter prospective primary cohort included 1373 patients with CRC, and the internal validation cohort enrolled 409 patients with CRC. Least absolute shrinkage and selection operator regression analyses were used to select prognostic indicators and develop a nomogram. The concordance (C)-index, receiver operating characteristic (ROC) curve, and decision curve analysis (DCA) were used to assess the prognostic discriminative ability of the nomogram, TNM stage, Patient-Generated Subjective Global Assessment (PGSGA), and TNM stage + PGSGA models. The overall survival (OS) curve of risk group stratification was calculated based on the nomogram risk score.
TNM stage, radical resection, reduced food intake, activities and function declined, and albumin were selected to develop the nomogram. The C-index and calibration plots of the nomogram showed good discrimination and consistency for CRC. Additionally, the ROC curves and DCA of the nomogram showed better survival prediction abilities in CRC than the other models. The stratification curves of the different risk groups of the different TNM categories were significantly different.
The novel nomogram showed good short- and long-term outcomes of OS in patients with CRC. This model provides a personalized and convenient prognostic prediction tool for clinical applications.
准确预测癌症患者的短期和长期生存情况至关重要。肿瘤-淋巴结-转移（TNM）分期能够准确预测癌症患者的长期生存，但无法预测短期生存。营养状况会影响癌症患者的个体状态和短期预后。我们的假设是，将TNM分期和营养相关因素纳入一个列线图可改善对结直肠癌（CRC）患者的生存预测。
本多中心前瞻性初级队列纳入了1373例CRC患者，内部验证队列纳入了409例CRC患者。采用最小绝对收缩和选择算子回归分析来选择预后指标并制定列线图。一致性（C）指数、受试者工作特征（ROC）曲线和决策曲线分析（DCA）用于评估列线图、TNM分期、患者主观整体评估（PGSGA）以及TNM分期+PGSGA模型的预后判别能力。根据列线图风险评分计算风险组分层的总生存（OS）曲线。
选择TNM分期、根治性切除、食物摄入量减少、活动及功能下降和白蛋白来制定列线图。列线图的C指数和校准图显示对CRC具有良好的判别能力和一致性。此外，列线图的ROC曲线和DCA显示在CRC中比其他模型具有更好的生存预测能力。不同TNM类别的不同风险组的分层曲线存在显著差异。
新型列线图在CRC患者中显示出良好的短期和长期OS结局。该模型为临床应用提供了一种个性化且便捷的预后预测工具。
