Machine Learning Predicts Rare Blood Cancer

Study in Journal of Clinical Pathology on predicting PV using CBC parameters.

Extreme Gradient Boosting algorithm predicted PV with 94 percent accuracy.

1,484 adults analyzed; 82 had PV.

Platelet count was the most influential CBC parameter.

EPO levels differed significantly between PV and non-PV groups.

Limitations include incomplete data on confounding factors.

Further research with larger datasets is recommended.