摘要
目的开发并验证一种基于新型血小板指数评分(platelet index score,PIS)的列线图模型预测肝细胞癌(hepatocellular carcinoma,HCC)病人的预后。方法回顾性分析2017年1月至2022年6月间于郑州大学人民医院肝胆胰腺外科行手术切除治疗的692例HCC病人的病历资料。收集入组病人的术前检验、临床病理特征和手术相关资料。按照标准方案于病人术后开始随访,以无复发生存时间(recurrence-free survival,RFS)作为主要研究结局,观察HCC切除术后复发情况。其中训练队列485例,验证队列207例。应用Kaplan-Meier曲线在训练队列中分析血小板参数,建立血小板指数评分。在训练队列中采用单因素和多因素Cox比例风险回归模型分析HCC病人术后复发的独立危险因素并构建列线图模型。绘制训练队列和验证队列受试者操作特征(receiver operator characteristic,ROC)曲线评估列线图模型预测效能。绘制训练队列和验证队列校准曲线评估列线图模型预测RFS与实际值的一致性。结果通过绘制Kaplan-Meier曲线发现:较低的血小板计数(≤157.5×10^(9)/L,P=0.001)、较高的平均血小板体积(≥11.35 fL,P<0.001)、较高的血小板分布宽度(≥13.85 fL,P<0.001)与较短的RFS相关。将以上3种血小板指标整合为PIS的新型评分系统中,绘制Kaplan-Meier曲线展示出其有良好的预测性能,并根据PIS得分分为高风险组和低风险组,低风险组1、2、3年的无复发生存率分别为13.32%(51/383),28.20%(108/383),38.90%(149/383),高风险组1、2、3年的无复发生存率分别为23.53%(24/102),49.02%(50/102),67.65%(69/102)。多因素Cox比例风险回归模型分析结果显示,美国癌症联合委员会(AJCC)分期(HR=2.921,95%CI:1.83~4.67,P<0.001)、微血管浸润(HR=1.906,95%CI:1.28~2.83,P=0.001)、门静脉癌栓(HR=1.408,95%CI:1.03~1.92,P=0.031)、肿瘤卫星灶(HR=1.388,95%CI:1.03~1.88,P=0.033)、Ki-67表达(HR=1.997,95%CI:1.45~2.75,P<0.001)、甲胎蛋白(HR=1.723,95%CI:1.29~2.30,P<0.001)和PIS分组(HR=1.442,95%CI:1.08~1.92,P=0.013)均是影响HCC切除术后复发的独立危险因素。根据独立危险因素绘制列线图,1、2、3年的列线图模型在训练队列中曲线下面积(area under the curve,AUC)分别为0.826(95%CI:0.773~0.876)、0.850(95%CI:0.800~0.876)、0.909(95%CI:0.882~0.934),在验证队列中AUC分别为0.826(95%CI:0.752~0.890)、0.807(95%CI:0.741~0.863)、0.804(95%CI:0.737~0.862)。列线图评分总分越高,提示模型预测病人PFS效能越好。校准曲线结果显示,列线图模型预测HCC切除术后病人的1、2、3年RFS与实际值的一致性在训练队列和验证队列中均较好。结论结合了基于多种血小板指标的PIS的列线图模型可良好预测病人术后1、2、3年的复发状态,为HCC病人术后提供更好的风险分层。
Objective This study aims to develop and validate a novel platelet index score(PIS)-based nomogram to predict the prognosis of hepatocellular carcinoma(HCC).Methods A retrospective analysis was conducted on the medical records of 692 patients with HCC who underwent surgical resection at the Department of Hepatopancreatobiliary Surgery,Zhengzhou University People’s Hospital,between January 2017 and June 2022.Preoperative laboratory testing,clinicopathological characteristics,and surgery-related data were collected.Postoperative follow-up was performed according to standard protocols,with recurrence-free survival(RFS)as the primary outcome to assess early recurrence and metastasis after HCC resection.Platelet parameters were analyzed using Kaplan-Meier curves in the training cohort to establish the platelet index score.Independent risk factors for postoperative recurrence were identified using univariate and multivariate Cox proportional hazards regression models,and a nomogram was constructed.The predictive performance of the model was evaluated using receiver operating characteristic(ROC)curves and calibration curves in both the training and validation cohorts,aiming to assess the consistency with actual RFS outcomes.Results Kaplan-Meier analysis revealed that lower platelet counts(PLT≤157.5×10^(9)/L,P=0.001),higher mean platelet volume(MPV≥11.35 fL,P<0.001),and higher platelet distribution width(PDW≥13.85 fL,P<0.001)were associated with shorter RFS in HCC patients.These three platelet indices were integrated into a novel scoring system,namely PIS,which demonstrated a good predictive performance.Based on the PIS,HCC patients were stratified into high-and low-risk groups.The 1-,2-,and 3-year RFS in the low-risk group was 13.32%(51/383),28.20%(108/383),and 38.90%(149/383),respectively,which was 23.53%(24/102),49.02%(50/102),and 67.65%(69/102)in the high-risk group,respectively.Multivariate Cox regression analysis identified the American Joint Committee on Cancer(AJCC)staging(HR=2.921,95%CI:1.83-4.67,P<0.001),microvascular invasion(HR=1.906,95%CI:1.28-2.83,P=0.001),portal vein tumor thrombus(HR=1.408,95%CI:1.03-1.92,P=0.031),tumor satellite lesions(HR=1.388,95%CI:1.03-1.88,P=0.033),Ki-67 expression(HR=1.997,95%CI:1.45-2.75,P<0.001),alpha-fetoprotein(HR=1.723,95%CI:1.29-2.30,P<0.001),and PIS(HR=1.442,95%CI:1.08-1.92,P=0.013)were independent risk factors for postoperative recurrence.The nomogram was plotted based on independent risk factors.The results showed that the AUC of the nomogram model in the training queue for 1-,2-,and 3-years was 0.826(95%CI:0.773-0.876),0.850(95%CI:0.800-0.876),and 0.909(95%CI:0.882-0.934)respectively,and the AUC in the validation queue was 0.826(95%CI:0.752-0.890),0.807(95%CI:0.741-0.863),and 0.804(95%CI:0.737-0.862)respectively.The higher the total score of the nomogram,the better the efficacy of the model in predicting the RFS of patients.Calibration curves showed a good consistency between the predicted and actual RFS in both cohorts.Conclusion The PIS-based nomogram model,incorporating multiple platelet indices,accurately predicts 1-,2-,and 3-year recurrence status after HCC resection and provides effective postoperative risk stratification for patients with HCC.
作者
朱恒立
蔡驰宇
李炳垚
唐昌乾
任泳年
李德宇
Zhu Hengli;Cai Chiyu;Li Bingyao;Tang Changqian;Ren Yongnian;Li Deyu(Department of Hepatopancreatobiliary Surgery,Zhengzhou University People's Hospital,Henan Zhengzhou 450003,China;Xinxiang Medical University,Henan Xinxiang 453003,China;Department of Hepatopancreatobiliary Surgery,Henan University People's Hospital,Henan Zhengzhou 450003,China)
出处
《腹部外科》
2025年第3期229-238,共10页
Journal of Abdominal Surgery
基金
国家自然科学基金(82103617,82470653)
河南省科技攻关项目(232301420056,232102311024)。
关键词
血小板指数评分
肝细胞癌
预测模型
术后复发
Platelet index score
Hepatocellular carcinoma
Prediction model
Postoperative recurrence
