结直肠癌伴同时肝转移患者同期切除术后预后风险模型的创建及应用

目的:探讨结直肠癌伴同时肝转移（synchronous colorectal liver metastasis，sCRLM）患者接受原发肿瘤和肝转移灶同期切除术后的预后风险因素，并建立sCRLM术后的预后风险模型。方法:回顾性分析127例在我院接受原发肿瘤和肝转移灶同期切除的sCRLM患者的临床病理学因素，采用单因素和多因素Cox模型分析方法确定总生存期（overall survival,OS）和肝无复发生存时间（recurrence-free survival,RFS）的独立预后因素，根据Cox模型的风险比(hazard ratio,HR）对独立因素赋值，从而建立OS和肝RFS的新预后风险模型。根据新模型将患者分为低、中和高风险组，各组间生存差异采用Log-rank检验。ROC曲线下面积用于比较新风险模型和Fong临床风险评分的预后预测能力。结果:本研究中位随访期23.4个月，多因素Cox分析确定下列4个因素为OS较差的独立预后因素，并根据HR值对各危险因素赋值:原发肿瘤淋巴结转移≥4个（P=0.047,1分），CEA≥200 ng/ml（P=0.023,2分）,TBS 3~8分（P=0.048,1分）和TBS≥9分（P=0.021,2分）。根据下列标准建立OS的预后风险模型:低风险组（0~1分）、中风险组（2~3分）和高风险组（≥4分）。新风险模型OS的预后预测能力优于Fong临床风险评分（AUC:0.726 vs 0.564)。多因素Cox分析确定下列2个因素为肝RFS较差的独立预后因素，根据HR值赋值如下:肝转移灶双叶分布（P=0.020,1分）和TBS≥9分（P=0.016,2分）。结论:本研究纳入接受同期手术切除的sCRLM患者创建了预后风险模型，用于评估sCRLM患者术后的OS和肝RFS。低风险组患者术后的生存情况（OS和肝RFS）优于高风险组，其预测能力尚需前瞻性临床模型进一步验证。

Creation and application of prognostic risk model after simultaneous liver and colorectal resection for patients with synchronous colorectal liver metastasis

Objective:To identify risk factors associated with cancer-specific overall survival（OS) and liver recurrence-free survival（RFS),and develop a prognostic risk model for patients with synchronous colorectal liver metastasis(sCRLM) who underwent simultaneous curative resection.Methods:A cohort of 127 consecutive sCRLM patients who underwent simultaneous liver and colorectal resection in our hospital was retrospectively analysed to identify independent predictors of OS and liver RFS with univariable and multivariable methods.These predictors were combined into a new prognostic risk model based on the hazard ratio(HR) of the Cox model.The patients were classified into low-,medium-and high-risk groups according to the OS and liver RFS risk scores.The discriminatory power（area under the curve,AUC) of the new prognostic model of OS versus Fong score was assessed.Results:The median follow-up was 23.4 months.Four predictors of worse OS were identified on multivariable Cox regression analysis,and the scores of each predictor were calculated according to the HR value:primary tumour lymph node ≥4（P=0.047,1 point),CEA greater than 200 ng/ml（P=0.023,2 points),TBS between 3 and 8(P=0.048,1 point) or TBS ≥ 9（P=0.021,2 points).A prognostic risk model of OS was performed according to the following criteria:low-risk group（score 0~1 point),medium-risk group（score 2~3 points) and high-risk group（score ≥4 points).The discriminatory capacity of the new risk model was superior to that of the Fong score（AUC:0.726 vs 0.564).Two predictors of worse liver RFS were identified on multivariable Cox regression analysis,and the scores of each predictor were calculated according to the HR value:bilobar distribution（P=0.020,1 point) and TBS≥9 points（P=0.016,2 points).Conclusion:A new prognostic risk model was proposed to evaluate OS and liver RFS after simultaneous resection of sCRLM.Patients in the low-risk group have longer OS and liver RFS,while those in the high-risk group have shorter OS and liver RFS.The prognostic risk model may help to predict patient survival after simultaneous liver and colorectal resection,and its predictive ability needs to be further verified by prospective clinical models.