Missing data handling technique in joint modeling context

Abstract

The maximum tolerated dose (MTD) is commonly practiced for dose selection in oncology. Higher doses work as the best effective treatment. Conventionally, the doses are selected in phase 2 from a phase. Perhaps, the quality of life gets compromised due to the high dose of chemotherapeutic regimes in the early phase of the trial. Alternative chemotherapy administration is Metronomic Chemotherapy (MC). Here very minimal doses are administered to avoid high toxicity. This work is about to handle missing data of circulating endothelial cells (CEC) and supports the optimal biological dose (OBD) for MC. It is performed with mimic data. A data simulation strategy is adopted. Simulation work is performed with R software. It helps to identify the suitable technique to handle missing data. The results conclude that the MC is efficacious by improving the Progression Free Survival (PFS) and Oveall Survival (OS) through controlled toxicity. Now the illustrated example can be extended to explore the impact of CEC for OS. This is a preliminary attempt towards MC to address some critical issues.

Keywords

• EM algorithm
• Handling Missing Data
• Imputation
• Predictive Mean Matching
• Regression method