双打击多发性骨髓瘤的Nomogram预测模型的构建与验证

Construction and verification of a Nomogram prediction model for double-hit multiple myeloma

目的 评估现有分期系统对多发性骨髓瘤(MM)的应用价值,构建双打击MM的Nomogram预测模型并验证。方法 回顾性纳入MM患者446例,将其以2∶1的比例随机分为训练组(295例)和验证组(151例)。收集所有患者的基线资料及预后情况并分组进行比较。采用受试者工作特征(ROC)曲线评估MM患者总生存期(OS)和无进展生存期(PFS)的分层预测价值。采用log-rank检验进行Kaplan-Meier生存曲线差异比较。采用单因素和多因素Cox回归分析评估OS的影响因素。构建Nomogram预后模型,采用C指数评估该模型的预测能力。结果 ROC曲线分析结果显示,R-ISS和D-S对MM患者的OS和PFS分层预测效能不高(P<0.05)。Kaplan-Meier生存曲线分析结果显示两组患者中位OS和PFS比较差异均无统计学意义(P>0.05)。多因素Cox回归分析结果显示,年龄、Double-hit、LDH、ISS评分均是影响OS的独立预后因素(P<0.05)。构建包含上述4个独立预后因素的Nomogram预后模型,对其能力进行评价,结果显示在训练组与验证组中预测OS与实际OS之间均具有良好的一致性。结论 现有分期系统对MM的生存预测仍然具有一定的局限,建立以Double-hit细胞遗传学异常为基础的Nomogram预后模型具有对OS良好的预测价值,或可用于指导临床个体化治疗。

Objective To evaluate the application value of the existing staging system for multiple myeloma(MM),and establish a Nomogram prediction model for double-hit MM for verification.MethodsA total of 446 patients with MM were retrospectively included and randomly divided into training group(295 cases) and validation group(151 cases) at a ratio of 2∶1.Baseline data and prognosis of all patients were collected and compared in different groups.The stratified predictive value of overall survival(OS) and progression-free survival(PFS) in MM patients was evaluated using receiver operating characteristic(ROC) curves.Kaplan-Meier survival curves were compared by log-rank test.Univariate and multivariate Cox regression analyses were used to evaluate the influencing factors of OS.A Nomogram prognostic model was constructed and its predictive ability was evaluated by C-index.Results ROC curve analysis results showed that R-ISS and D-S had low efficacy in stratified prediction of OS and PFS in MM patients(P<0.05).Kaplan-Meier survival curve analysis showed that there was no significant difference in the median OS and PFS between the two groups(P>0.05).Multivariate Cox regression analysis showed that age,Double-hit,LDH and ISS score were independent prognostic factors affecting OS(P<0.05).A Nomogram prognostic model containing the above four independent prognostic factors was constructed and its ability was evaluated.The results showed good consistency between the predicted OS and the actual OS in both the training group and the validation group.Conclusion The existing staging system still has some limitations in predicting the survival of MM.The established Nomogram prognostic model based on the cytogenetic abnormality of Double-hit has good value in predicting the OS,and may be used to guide clinical individualized treatment.