Coiffier B, Adeleine P, Gentilhomme O, Felman P, Treille-Ritouet D, Bryon P A
Hematology Service, Hôpital Edouard-Herriot, Lyon, France.
Cancer. 1987 Dec 15;60(12):3029-32. doi: 10.1002/1097-0142(19871215)60:12<3029::aid-cncr2820601229>3.0.co;2-4.
Two groups of patients with a myelodysplastic syndrome (MDS) were analyzed by univariate (log-rank test) and multivariate (logistic regression) analyses to detect the most important prognostic factors. By stepwise analysis, the variables found to have prognostic significance for death were as follows: age, percentage of marrow blasts, presence of circulating blasts, and number of platelets. The variables found significant for predicting progression to acute leukemia (AL) were as follows: hemoglobin level, percentage of marrow blasts, and presence of circulating blasts. The first group of 193 patients was used to build a prognostic index which reflected the probability of a given patient dying or progressing to AL within 6, 9, or 12 months. The application of this prognostic index to a test group of 143 patients was used to determine the expected error rate and the validity of the prediction rule.
通过单变量（对数秩检验）和多变量（逻辑回归）分析对两组骨髓增生异常综合征（MDS）患者进行分析，以检测最重要的预后因素。通过逐步分析，发现对死亡具有预后意义的变量如下：年龄、骨髓原始细胞百分比、循环原始细胞的存在情况以及血小板数量。发现对预测进展为急性白血病（AL）具有显著意义的变量如下：血红蛋白水平、骨髓原始细胞百分比以及循环原始细胞的存在情况。第一组193例患者用于构建一个预后指数，该指数反映了特定患者在6个月、9个月或12个月内死亡或进展为AL的概率。将该预后指数应用于143例患者的测试组，以确定预期错误率和预测规则的有效性。
