Interventional Department, The Second Hospital of Shandong University, Jinan, 250033, China.
Department of Hepatobiliary Surgery, Shandong Provincial Hospital, Shandong First Medical University, Jinan, 250021, China.
J Cancer Res Clin Oncol. 2023 Sep;149(12):9577-9586. doi: 10.1007/s00432-023-04858-0. Epub 2023 May 24.
To establish and validate a prognostic nomogram of cholangiocarcinoma (CCA) using independent clinicopathological and genetic mutation factors.
213 patients with CCA (training cohort n = 151, validation cohort n = 62) diagnosed from 2012 to 2018 were included from multi-centers. Deep sequencing targeting 450 cancer genes was performed. Independent prognostic factors were selected by univariate and multivariate Cox analyses. The clinicopathological factors combined with (A)/without (B) the gene risk were used to establish nomograms for predicting overall survival (OS). The discriminative ability and calibration of the nomograms were assessed using C-index values, integrated discrimination improvement (IDI), decision curve analysis (DCA), and calibration plots.
The clinical baseline information and gene mutations in the training and validation cohorts were similar. SMAD4, BRCA2, KRAS, NF1, and TERT were found to be related with CCA prognosis. Patients were divided into low-, median-, and high-risk groups according to the gene mutation, the OS of which was 42.7 ± 2.7 ms (95% CI 37.5-48.0), 27.5 ± 2.1 ms (95% CI 23.3-31.7), and 19.8 ± 4.0 ms (95% CI 11.8-27.8) (p < 0.001), respectively. The systemic chemotherapy improved the OS in high and median risk groups, but not in the low-risk group. The C-indexes of the nomogram A and B were 0.779 (95% CI 0.693-0.865) and 0.725 (95% CI 0.619-0.831), p < 0.01, respectively. The IDI was 0.079. The DCA showed a good performance and the prognostic accuracy was validated in the external cohort.
Gene risk has the potential to guide treatment decision for patients at different risks. The nomogram combined with gene risk showed a better accuracy in predicting OS of CCA than not.
利用独立的临床病理和基因突变因素，建立并验证胆管癌（CCA）的预后列线图。
本研究纳入了 2012 年至 2018 年间多中心诊断的 213 名 CCA 患者（训练队列 n=151，验证队列 n=62）。对 450 个癌症基因进行深度测序。通过单因素和多因素 Cox 分析选择独立的预后因素。使用列线图预测总生存期（OS），分别结合（A）和不结合（B）基因风险的临床病理因素。采用 C 指数、综合判别改善（IDI）、决策曲线分析（DCA）和校准图评估列线图的判别能力和校准度。
训练和验证队列的临床基线信息和基因突变相似。SMAD4、BRCA2、KRAS、NF1 和 TERT 与 CCA 预后相关。根据基因突变将患者分为低危、中危和高危组，OS 分别为 42.7±2.7ms（95%CI 37.5-48.0）、27.5±2.1ms（95%CI 23.3-31.7）和 19.8±4.0ms（95%CI 11.8-27.8）（p<0.001）。系统化疗改善了高危和中危组的 OS，但对低危组没有改善。列线图 A 和 B 的 C 指数分别为 0.779（95%CI 0.693-0.865）和 0.725（95%CI 0.619-0.831），p<0.01。IDI 为 0.079。DCA 显示出良好的性能，并在外部队列中验证了预测准确性。
基因风险有可能指导不同风险患者的治疗决策。与不结合基因风险的列线图相比，结合基因风险的列线图在预测 CCA 的 OS 方面具有更高的准确性。
