Predicting progression events in multiple myeloma from routine blood work

Area of Science:

• Computational biology
• Medical informatics
• Oncology

Background:

• Multiple myeloma is a cancer of plasma cells, requiring continuous monitoring for disease progression.
• Accurate prediction of disease progression is crucial for timely intervention and personalized treatment strategies.

Purpose of the Study:

• To develop and validate a hybrid neural network system for predicting disease progression events in multiple myeloma.
• To forecast future laboratory results and identify disease progression markers from historical data.

Main Methods:

• Utilized a hybrid neural network architecture for predictive modeling.
• Trained and validated the model on datasets from the CoMMpass (N=1186) and GMMG-MM5 (N=504) studies.
• Incorporated routine blood work measurements for accessibility and interpretability.

Main Results:

• The model accurately predicted future blood work parameters, outperforming baseline estimators.
• Disease progression events were reliably predicted from forecasted data.
• External validation confirmed the model's reproducibility and significant predictive performance.

Conclusions:

• The developed system enables early detection and personalized monitoring of multiple myeloma patients at risk of progression.
• The modular design enhances interpretability and facilitates integration, contributing to a virtual human twin system.
• This approach supports optimized healthcare resource utilization and improved patient outcomes in multiple myeloma care.