Survival Analysis of Multiple Myeloma Cancer (MMC) Using the Cox-Proportional Hazard Model

Though multiple myeloma cancer (MMC) remains incurable, research into improving the therapeutic strategy has increaseddramatically in recent years. But it is unclear if sustained improvements have been achieved. We studied the survival times of48 patients diagnosed and treated with alkylating agents. The semi-parametric Cox proportional hazard model was employedto examine the survival probability taking into account the sixteen risk factors presumed to be contributing to the survivaltimes. A careful and rigorous assessment of the risk factors based on the AIC of the stepwise selection technique revealed sevenrisk factors, and one interaction term are statistically significantly contributing to the survival times. They are blood ureanitrogen (BUN)/serum creatinine, white blood cells (WBC), Bence Jone protein in the urine (BJPU), fractures, proteinuria,gender, platelets, and the interaction of infections and serum calcium. The final Cox-PH model was well-validated and satisfiedthe key assumptions. The identified risk factors are rank according to the prognostic effect on the survival time based on thehazard ratio. Blood urea nitrogen (BUN)/serum creatinine was the greatest prognostic factor (most contributing factor, andhighly negatively related to the MMC deaths or survival times), followed by white blood cells (WBC), and normal plateletwas found to be the minimum prognostic factor (least contributing factor to MMC death or survival times). This study offersprognostic and therapeutic significance for further enhancement in the treatment strategy of the multiple myeloma cancerdisease.