Implication of Statistical Methods on Patient Data: An Approach for Cancer Survivability Prediction

With the advent of next-generation sequencing technologies, there is a reduction in sequencing costs, which on the other hand have increased the importance of cancer survival studies due to availability of high-throughput data and advancement in high-resolution cancer genome analysis with increased sensitivity. In this chapter, we have discussed selected statistical methods for survival analysis along with their strengths and limitations. We have considered global data corresponding to cancer statistics and patient survival rates. Further, we have emphasized on the biological and technical concepts of survival analysis. Here we have mainly focused on two popular female cancers, i.e., breast and ovarian carcinoma. We have employed Kaplan-Meier estimator and Cox proportional hazard regression to illustrate the case studies based on the featured design principles using cancer dataset. The performance of the Cox proportional hazard model surpasses that of the 2D statistical comparison model of Kaplan-Meier estimator as it can incorporate multiple layers of clinical information which enriches the prediction outcome.