ColloquiumSpeaker:
Sunil Mathur, Ph.D.
Professor, Weill Cornell Medical College
Director, Biostatistics, Houston Methodist Neal Cancer Center
Houston, Texas
Title: Testing of Hypotheses for Achieving Higher Efficiency with Applications in Cancer Research
Abstract: We propose a new two-sample test for a two-sample location problem based on empirical distribution function which is based on rank-order to not only detect the differences between the control and treatment arm but also save time and cost in clinical trials. The test statistic is constructed as a power divergence between empirical distribution functions obtained from the two independent samples making it to more powerful than its competitors under heavy-tailed and light-tailed distributions. The permutation principle is used to implement the test. We show that our test is component-wise scale invariant. Further, the distribution of the proposed test statistic is obtained under the null hypothesis and in general. We also report the theoretical expectation and variance of the proposed test statistic when the null hypothesis is true. Using the Monte Carlo method we computed empirical power which shows that our test performs better than its competitors under heavy-tailed, light-tailed, and even elliptically asymmetric population distribution. Overall, the proposed test provides better power than its competitors considered here irrespective of the nature of the population. We apply the proposed test to Triple-negative breast cancer (TNBC) which has a higher recurrence rate of 6.7–10.5 months compared to an overall recurrence rate of 2.1–6.4 months for all breast cancer types. This type of cancer has shorter times to recurrence and a more aggressive clinical course, leading to a worse prognosis and lower survival rates than non-TNBC patients. Testing drug efficacy for treating TNBC is one of the most challenging tasks in cancer research, yet it is crucial for making informed medical decisions, supporting healthcare planning, and aiding patients in making informed healthcare choices. Our proposed test can lead to Improved drug efficacy which further can lead to enhanced quality of care, improved efficiency, increased cost-benefit, and higher satisfaction for both providers and patients.
Day & Time: Thursday, September 12, 2024 at 3:00pm
Location: TBA
Counts toward seminar attendance for MSc and PhD students in Math & Stats.