Related Experiment Video
Updated: Aug 12, 2025
10:04
Establishment of a Human Multiple Myeloma Xenograft Model in the Chicken to Study Tumor Growth, Invasion and Angiogenesis
Published on: May 1, 2015
13.1K
Machine learning-based infection prediction model for newly diagnosed multiple myeloma patients
Ting Peng1, Leping Liu2, Feiyang Liu1
1Department of Hematology, The Third Xiangya Hospital of Central South University, Changsha, China.
Frontiers in Neuroinformatics
|January 30, 2023
Summary
This study developed a machine learning model to predict infections in newly diagnosed multiple myeloma (NDMM) patients. The model identifies key risk factors, aiming to reduce infection incidence and improve patient outcomes.
Area of Science:
- Hematology
- Oncology
- Medical Informatics
Background:
- Infections pose a significant risk to patients with newly diagnosed multiple myeloma (NDMM).
- Understanding infection characteristics and risk factors is crucial for improving patient prognosis.
Purpose of the Study:
- To develop and validate a predictive model for infection risk in NDMM patients.
- To identify key variables associated with infection development in NDMM.
Main Methods:
- A multicenter clinical dataset of 564 NDMM patients was analyzed.
- Machine learning algorithms, including XGBoost, were employed to build and evaluate a prediction model.
- Key variables were selected using the Lasso method, and model performance was assessed via AUC and accuracy.
Main Results:
- Fifteen key variables were identified, including age, ECOG, osteolytic disruption, and various laboratory indicators.
- The XGBoost model demonstrated superior predictive performance with an AUROC of 0.8664.
- The model achieved an accuracy of 68.64% on the validation dataset.
Conclusions:
- A validated, machine learning-based infection prediction model for NDMM patients was successfully established.
- This model offers a simple and convenient tool to aid in reducing infection incidence and enhancing patient prognosis.
