Li Manning, Jia Sixun, Wu Han, Shou Chunyi, Song Xiaolu, Peng Ye, Wang Huafang, Wang Ying, Tong Xiangmin, Chen Yirui
Department of Hematology, Affiliated Zhongshan Hospital of Dalian University, 6 Jiefang Road, Dalian, 116001, Liaoning, China.
Cancer Center, Department of Hematology, Affiliated People's Hospital, Zhejiang Provincial People's Hospital, Hangzhou Medical College, 158 Shangtang Road, Hangzhou, 310014, Zhejiang, China.
Discov Oncol. 2025 Aug 9;16(1):1514. doi: 10.1007/s12672-025-03156-0.
Develop a diagnostic model using common hematological and immunological indicators to assist in the early screening and differential diagnosis of Multiple Myeloma (MM) in clinical settings, reducing the risk of misdiagnosis.
A retrospective analysis was conducted on 274 newly diagnosed and treated MM patients and 137 connective tissue disease patients treated at Zhejiang Provincial People's Hospital from January 2008 to August 2023. Laboratory indicators, including complete blood count, biochemistry, coagulation function, and immunoglobulin markers, were collected. The cohort was randomly divided into a 70% training set and a 30% validation set. Relevant variables were selected through univariate and multivariate analyses in the training set. A discriminative diagnostic model was developed using a multivariate logistic regression algorithm. The model's predictive accuracy and generalizability were evaluated by validating and conducting receiver operating characteristic (ROC) curves and calibration curves.
The developed differential diagnostic model in this study included the following observed indicators: IgM, glomerular filtration rate, high-density lipoprotein, red cell distribution width, and thrombin time. The model demonstrated excellent discriminatory power and good calibration, with an area under the curve (AUC) value of 0.980 (95% CI: 0.967-0.994). Additionally, the model exhibited high sensitivity (0.963), specificity (0.938), accuracy (0.955). The validation set further confirmed the generalization and accuracy of the model, with an AUC value of 0.954 (95% CI: 0.961-0.992).
The constructed differential diagnostic model in this study can accurately predict and differentiate MM patients and those with elevated Ig abnormalities, thereby enhancing the efficiency of clinical diagnostic decision-making.
利用常见血液学和免疫学指标建立诊断模型，以协助临床环境中多发性骨髓瘤（MM）的早期筛查和鉴别诊断，降低误诊风险。
对2008年1月至2023年8月在浙江省人民医院新诊断并接受治疗的274例MM患者和137例结缔组织病患者进行回顾性分析。收集全血细胞计数、生化、凝血功能和免疫球蛋白标志物等实验室指标。将队列随机分为70%的训练集和30%的验证集。通过训练集中的单因素和多因素分析选择相关变量。使用多因素逻辑回归算法建立判别诊断模型。通过验证以及绘制受试者工作特征（ROC）曲线和校准曲线来评估模型的预测准确性和泛化能力。
本研究建立的鉴别诊断模型包括以下观察指标：IgM、肾小球滤过率、高密度脂蛋白、红细胞分布宽度和凝血酶时间。该模型具有出色的判别能力和良好的校准，曲线下面积（AUC）值为0.980（95%CI：0.967 - 0.994）。此外，该模型具有高敏感性（0.963）、特异性（0.938）、准确性（0.955）。验证集进一步证实了模型的泛化能力和准确性，AUC值为0.954（95%CI：0.961 - 0.992）。
本研究构建的鉴别诊断模型能够准确预测和区分MM患者与Ig异常升高的患者，从而提高临床诊断决策的效率。
