应用"数学模型"预测皮肌炎患者合并恶性肿瘤情况的研究

目的：建立1个数学模型，预测临床皮肌炎患者合并恶性肿瘤情况。方法：22例皮肌炎合并恶性肿瘤患者作为一组，43例皮肌炎不合并恶性肿瘤患者作为另一组，对两组患者的临床及实验室指标进行分析，建立判别函数。结果：典型皮疹、皮肤异色症、吞咽困难、VCA－IgA等4个旨标两组差别具有统计学意义，根据4个指标建立的函数，阳性预测值为61．97％，阴性预测值为100％，预测准确率为86．47％。结论：本判别函数数学模型简单，在理论上可较准确的判断皮肌炎合并恶性肿瘤的情况。

Predicting the status of patients with dermatomyositis complicated by malignant tumor by a mathematical model

Objective:To establish a mathematical model to predict the status of patients with dermatomyositis complicated by malignant tumor. Methods: The dada from 22 patients with dermatomyositis complicated by malignant tumor and 43 patients with dermatomyositis without malignant tumor were analyzed and used to establish a discriminant function.Results: Four factors showed statistically significant difference between these two groups: typical rash, poikioderma, dysphagia and positive VCA-IgA. Fisher's liner discriminant function was set up using these 4 factors. According to this mathematical model, the positive predictive value, negative predictive value and predictive accuracy of dermatomyositis complicated by malignant tumor were 61.97%, 100% and 86.47% , respectively. Conclusion : The mathematical model of discriminant function could predict the status of dermatomyositis complicated by malignant tumor accurately.