基于中国患者遗传背景的MDS预后评估：IPSS-M与IPSS/IPSS-R体系的比较

Prognostic Assessment of MDS Based on the Genetic Background of Chinese Patients: A Comparison of the IPSS-M versus IPSS and IPSS-R Systems

摘要: 目的：比较分子国际预后评分系统(IPSS-M)、修订版国际预后评分系统(IPSS-R)及国际预后评分系统(IPSS)在中国骨髓增生异常综合征(MDS)患者中的预后评估能力，并探讨其在中国患者遗传背景下的适用性。方法：回顾性分析2018年1月至2024年12月在安徽医科大学第一附属医院确诊的82例MDS患者临床资料及随访数据，所有患者均接受染色体核型分析及二代测序检测。依据IPSS、IPSS-R及IPSS-M进行风险分层，比较不同评分系统下的风险分布及生存预测能力。采用Kaplan-Meier法绘制生存曲线并行log-rank检验；使用加权及非加权中位生存时间残差(weighted median survival residual)比较模型预测误差；基于time-dependent ROC分析评估不同时间节点的预测效能。结果：与IPSS-R相比，IPSS-M对42例患者(52.4%)进行了重新分层。Kaplan-Meier分析显示三种评分系统分层的总生存期差异均有统计学意义(P < 0.001)。加权中位生存时间残差分析显示，IPSS-M的加权中位绝对残差为4.9个月(95%CI：3.2~6.8)，显著低于IPSS-R的16.1个月(95%CI：12.3~20.7；Wilcoxon V = 1540.5，P < 0.001，效应量r = 0.74)。时间依赖性ROC曲线显示，IPSS-M在40个月时AUC最高(0.954)，优于IPSS-R (0.782)及IPSS (0.729)。结论：本研究结果支持IPSS-M在中国MDS患者中较传统系统具有更高的预后预测准确性。考虑到不同人群的遗传差异，未来仍需大样本、多中心、前瞻性研究进一步验证。

Abstract: Objective: To compare the prognostic performance of the Molecular International Prognostic Scoring System (IPSS-M), the Revised International Prognostic Scoring System (IPSS-R), and the original International Prognostic Scoring System (IPSS) in Chinese patients with myelodysplastic syndromes (MDS), and to assess their applicability within the context of the Chinese genetic background. Methods: We retrospectively analyzed clinical and follow-up data of 82 patients with newly diagnosed MDS at the First Affiliated Hospital of Anhui Medical University between January 2018 and December 2024. All patients underwent conventional cytogenetic karyotyping and next-generation sequencing. Risk stratification was performed according to IPSS, IPSS-R, and IPSS-M. The distribution of risk categories and survival prediction performance across scoring systems were compared. Kaplan-Meier survival curves were generated and compared using the log-rank test. Weighted and unweighted median survival residuals were calculated to assess model prediction error. Time-dependent receiver operating characteristic (ROC) analyses were conducted to evaluate predictive accuracy at different time points. Results: Compared with IPSS-R, IPSS-M reclassified 42 patients (52.4%). Kaplan-Meier analysis demonstrated statistically significant differences in overall survival across all three systems (P < 0.001). Weighted median survival residual analysis showed that IPSS-M had a weighted median absolute residual of 4.9 months (95% CI: 3.2~6.8), significantly lower than that of IPSS-R (16.1 months; 95% CI: 12.3~20.7; Wilcoxon V = 1540.5, P < 0.001, effect size r = 0.74). Time-dependent ROC analysis revealed that IPSS-M achieved the highest AUC at 40 months (0.954), outperforming IPSS-R (0.782) and IPSS (0.729). Conclusion: In Chinese MDS patients, IPSS-M demonstrates superior prognostic accuracy compared with IPSS and IPSS-R. Given potential genetic differences among populations, large-scale, multicenter, prospective studies are warranted to further validate these findings.