Cancer Research Center-IBMCC-USAL-CSIC, Department of Medicine and Cytometry Service-Nucleus Platform, Centro de Investigación Biomédica en Red de Cáncer (CIBERONC), University of Salamanca, Salamanca, Spain; Biomedical Research Institute of Salamanca (IBSAL), Salamanca, Spain; Spanish Network on Mastocytosis, Toledo and Salamanca, Spain.
Spanish Network on Mastocytosis, Toledo and Salamanca, Spain; Instituto de Estudios de Mastocitosis de Castilla La Mancha and CIBERONC, Virgen del Valle Hospital, Toledo, Spain.
Lancet Haematol. 2021 Mar;8(3):e194-e204. doi: 10.1016/S2352-3026(20)30400-2. Epub 2021 Jan 25.
Several risk stratification models have been proposed in recent years for systemic mastocytosis but have not been directly compared. Here we designed and validated a risk stratification model for progression-free survival (PFS) and overall survival (OS) in systemic mastocytosis on the basis of all currently available prognostic factors, and compared its predictive capacity for patient outcome with that of other risk scores.
We did a retrospective prognostic modelling study based on patients diagnosed with systemic mastocytosis between March 1, 1983, and Oct 11, 2019. In a discovery cohort of 422 patients from centres of the Spanish Network on Mastocytosis (REMA), we evaluated previously identified, independent prognostic features for prognostic effect on PFS and OS by multivariable analysis, and designed a global prognostic score for mastocytosis (GPSM) aimed at predicting PFS (GPSM-PFS) and OS (GPSM-OS) by including only those variables that showed independent prognostic value (p<0·05). The GPSM scores were validated in an independent cohort of 853 patients from centres in Europe and the USA, and compared with pre-existing risk models in the total patient series (n=1275), with use of Harrells' concordance index (C-index) as a readout of the ability of each model to risk-stratify patients according to survival outcomes.
Our GPSM-PFS and GPSM-OS models were based on unique combinations of independent prognostic factors for PFS (platelet count ≤100 × 10 cells per L, serum β2-microglobulin ≥2·5 μg/mL, and serum baseline tryptase ≥125 μg/L) and OS (haemoglobin ≤110 g/L, serum alkaline phosphatase ≥140 IU/L, and at least one mutation in SRSF2, ASXL1, RUNX1, or DNMT3A). The models showed clear discrimination between low-risk and high-risk patients in terms of worse PFS and OS prognoses in the discovery and validation cohorts, and further discrimination of intermediate-risk patients. The GPSM-PFS score was an accurate predictor of PFS in systemic mastocytosis (C-index 0·90 [95% CI 0·87-0·93], vs values ranging from 0·85 to 0·88 for pre-existing models), particularly in non-advanced systemic mastocytosis (C-index 0·85 [0·76-0·92], within the range for pre-existing models of 0·80 to 0·93). Additionally, the GPSM-OS score was able to accurately predict OS in the entire cohort (C-index 0·92 [0·89-0·94], vs 0·67 to 0·90 for pre-existing models), and showed some capacity to predict OS in advanced systemic mastocytosis (C-index 0·72 [0·66-0·78], vs 0·64 to 0·73 for pre-existing models).
All evaluated risk classifications predicted survival outcomes in systemic mastocytosis. The REMA-PFS and GPSM-PFS models for PFS, and the International Prognostic Scoring System for advanced systemic mastocytosis and GPSM-OS model for OS emerged as the most accurate models, indicating that robust prognostication might be prospectively achieved on the basis of biomarkers that are accessible in diagnostic laboratories worldwide.
Carlos III Health Institute, European Regional Development Fund, Spanish Association of Mastocytosis and Related Diseases, Rare Diseases Strategy of the Spanish National Health System, Junta of Castile and León, Charles and Ann Johnson Foundation, Stanford Cancer Institute Innovation Fund, Austrian Science Fund.
近年来，已经提出了几种用于系统性肥大细胞增多症的风险分层模型，但尚未进行直接比较。在这里，我们基于目前所有可用的预后因素，为系统性肥大细胞增多症的无进展生存期（PFS）和总生存期（OS）设计并验证了一种风险分层模型，并比较了其对患者预后的预测能力与其他风险评分。
我们进行了一项回顾性预后建模研究，纳入了 1983 年 3 月 1 日至 2019 年 10 月 11 日期间在西班牙肥大细胞网络（REMA）中心诊断为系统性肥大细胞增多症的患者。在一个由 422 名患者组成的发现队列中，我们通过多变量分析评估了先前确定的、独立的预后因素对 PFS 和 OS 的预后影响，并设计了一个全球肥大细胞预后评分（GPSM），旨在通过仅纳入具有独立预后价值的变量来预测 PFS（GPSM-PFS）和 OS（GPSM-OS）（p<0·05）。GPSM 评分在来自欧洲和美国中心的 853 名患者的独立队列中进行了验证，并在总患者系列中与现有的风险模型进行了比较（n=1275），使用 Harrells 一致性指数（C 指数）作为每个模型根据生存结果对患者进行风险分层的能力的指标。
我们的 GPSM-PFS 和 GPSM-OS 模型基于 PFS（血小板计数≤100×10 个/ L、血清β2-微球蛋白≥2·5μg/mL 和血清基础胰蛋白酶≥125μg/L）和 OS（血红蛋白≤110g/L、血清碱性磷酸酶≥140IU/L 和至少有一个 SRSF2、ASXL1、RUNX1 或 DNMT3A 突变）的独特组合的独立预后因素。在发现和验证队列中，这些模型在低风险和高风险患者之间显示出明确的区分，对 PFS 和 OS 预后较差，对中间风险患者进一步区分。GPSM-PFS 评分是系统性肥大细胞增多症 PFS 的准确预测指标（C 指数 0·90[0·87-0·93]，而现有模型的范围为 0·85 至 0·88），特别是在非进展性系统性肥大细胞增多症中（C 指数 0·85[0·76-0·92]，在现有模型的 0·80 至 0·93 范围内）。此外，GPSM-OS 评分能够准确预测整个队列的 OS（C 指数 0·92[0·89-0·94]，而现有模型的范围为 0·67 至 0·90），并在一定程度上能够预测进展性系统性肥大细胞增多症的 OS（C 指数 0·72[0·66-0·78]，而现有模型的范围为 0·64 至 0·73）。
所有评估的风险分类都预测了系统性肥大细胞增多症的生存结局。在 PFS 方面，REMA-PFS 和 GPSM-PFS 模型，以及高级系统性肥大细胞增多症的国际预后评分系统和 OS 的 GPSM-OS 模型是最准确的模型，表明基于在全球诊断实验室都可获得的生物标志物，可能能够前瞻性地进行稳健的预后预测。
西班牙卡洛斯三世健康研究所、欧洲区域发展基金、西班牙肥大细胞病和相关疾病协会、西班牙国家卫生系统罕见病战略、卡斯蒂利亚和莱昂大区、查尔斯和安约翰逊基金会、斯坦福癌症研究所创新基金、奥地利科学基金会。
