A next generation probabilistic approach to analyze cancer patients data with inference and applications

Abstract

This study addresses the critical challenges faced in cancer care, particularly in predicting survival times for patients with lung cancer and acute myeloid leukemia. Despite recent advancements in medical science, existing models often fall short in accurately capturing disease progression, leading to less effective clinical decision-making, and compromised patient outcomes. The need for advanced predictive models is urgent to improve survival time forecasts and enhance treatment strategies. In response to this, we introduce a novel probabilistic approach, the New Weibull (NEWE) model, which is part of a newly generated class designed to model cancer patient data more effectively. Our methodology includes using seven well-known estimation methods, each rigorously evaluated for consistency through Monte Carlo simulation studies focused on key metrics such as absolute bias, mean square error, and mean relative error. The datasets analyzed include survival times for twenty acute myeloid leukemia patients, 121 breast cancer patients from 1929 to 1938, 33 patients with acute myelogenous leukemia, data from eighteen individuals who died from causes unrelated to cancer, and survival times of advanced lung cancer patients undergoing standard chemotherapy. The NEWE model outperformed competing models, particularly in Anderson-Darling, Cramer-von Mises, and Kolmogorov-Smirnov tests, with significantly higher p-values. These findings highlight the NEWE model’s potential to transform predictive oncology by offering more precise survival time predictions, improving the quality of care and decision-making in cancer treatment.