Leveraging Explainable AI for Early Risk Prediction and Type Classification for Leukemia: Insights using Clinical Data from Pakistan

Leukemia, a prevalent childhood cancer affecting the blood and bone marrow, necessitates a proactive approach to risk mitigation through lifestyle adaptations. Despite previous investigations into similar aspects across different cancer types, a comprehensive inquiry focusing on all leukemia subtypes within Pakistan remains a significant research gap. Acknowledging the influence of regional variations on individuals' lifestyles, this study aims to identify lifestyle and demographic factors associated with leukemia development and predict specific leukemia subtypes using clinical data. Our data collection included 364 leukemia cases and 896 control subjects, gathered from different cancer or tertiary care hospitals in Islamabad and Peshawar, Pakistan. The data was meticulously categorized into laboratory results, demographic characteristics, and lifestyle parameters. For demographic and lifestyle analysis, we employed advanced techniques of Machine Learning, and statistical and graph-based methodologies to assess leukemia development risk. This study highlights factors associated with a higher risk of leukemia, including passive smoking, rural residence, and poor nutrition. These insights emphasize the promotion of healthier lifestyle choices to potentially reduce leukemia incidences. Additionally, we transformed the clinical dataset into graph data, which was utilized for leukemia classification and subtype prediction. We conducted classification on both graph and structured (tabular) data, with the structured data, achieving a 96% accuracy rate, notably on oversampled data. To enhance the interoperability of our classification outcomes, we employed the SHapley Additive exPlanations (SHAP) algorithm to explain the classification, offering comprehensive insights into the rationale behind leukemia classification.