Time：2021/12/22 10:30-11:30

Form：Tencent Meeting（ID：266 499 959）

Topic：Recent developments of accuracy metrics for biomarker evaluation with applications to ovarian cancer

Abstract:

The development of biomarkers into diagnostic and prognostic tests can be categorized into three broad phases: discovery, performance evaluation, and impact determination when added to existing clinical measures. This talk covers some key concepts including classification types and classification metrics in performance evaluation, from a statistician’s perspective. The limitations of existing classification metrics and the importance of using appropriate classification metrics are highlighted.  Specifically, this talk addresses one of the common pitfalls in biomarker evaluation, namely “naïve pooling”. Although biomarkers are evaluated using pooled data as a standard practice and such results are reported routinely in scientific journals, the consequences of “naïve pooling” in biomarker evaluation have not been investigated thoroughly. This talk points out severe hidden problems induced by “naïve pooling” and presents new appropriate performance measures.  Furthermore, the inefficiency and lack of flexibility in existing performance metrics used in multiple ordered classification are addressed and new flexible accuracy measures are presented.  An ovarian cancer data set from the PLCO cancer study is analyzed.

Resume:

Lili Tian, Ph.D., is a full professor, associate chair, and director of graduate studies at the Department of Biostatistics, University at Buffalo, State University of New York.   Dr. Tian is also an adjunct full Professor of Oncology at Roswell Park Comprehensive Cancer Institute.  Dr. Tian currently serves as an Editor-in-Chief for the world-renowned statistical/biostatistical journal “Statistical Methods in Medical Research”. She has published more than one hundred   SCI papers in the fields of medical diagnosis, biomarker evaluation,  statistical inference, cancer research, epidemiology, and nursing research.  She has been in charge of experimental design and statistical analysis for many large-scale studies funded by the US federal government including clinical trials and epidemiology studies. She is the president of the Buffalo/Niagara Chapter of the American Statistical Association (ASA) and serves on   Committee for Applied Statisticians for ASA. She also serves on the Scientific Advisory Board, Canget BioTekpharma, LLC, USA.