Machine Learning Model Predicts Monoclonal Gammopathy Using Routine Laboratory Values
A machine learning model flags patients likely to develop precancerous blood protein changes years early, using only standard blood tests already done during routine checkups.
An ML model trained on routine CBC and metabolic panel data from 232,813 patients achieved AUC 0.84 for predicting monoclonal gammopathy (MGUS/plasma cell myeloma precursor) within 5 years. This represents the first broadly validated ML model for M-protein detection from standard laboratory data, offering a scalable, low-cost pathway to identify MGUS patients before myeloma onset.
What the study was
- Study design
- Retrospective cohort, ML validation
- Population
- Outpatient adults aged 50-85 years, large US network
- Sample size
- 232813
- Category
- Early Detection
- Maturity
- Validated
- Journal
- JCO Clin Cancer Inform
Why it surfaced
First ML model predicting monoclonal gammopathy from routine CBC data; large real-world cohort (232K); MGUS early detection could prevent progression to plasma cell myeloma with high unmet need. CBC/ML topic breakout — first HIGH result after 8 consecutive zero-HIGH runs.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.