Hello Friends and Colleagues,
This Friday, November 17th, we will have one presentation. Dr Zhen Zhang from Johns Hopkins will be giving a presentation about Machine learning, big data, and clinical applications.
Information about the presentation is listed below:
Title: Machine learning, big data, and clinical applications: the long road to the development of OVA1, the first in vitro diagnostic multivariate index assay (IVDMIA) of protein biomarkers cleared by FDA
Description:
In 2009, FDA cleared OVA1 as the first in vitro diagnostic multivariate index assay (IVDMIA) test of protein biomarkers for clinical use. In 2016, Overa, the successor of OVA1, was also cleared by FDA. The development of these IVDMIAs represented a continuous effort starting in the mid-1990s to apply statistically principled machine learning approaches to combine multiple clinical tests to improve diagnostic accuracy. In this talk, I will review lessons learned from this process and discuss some special requirements on machine learning algorithms and data utilization due to clinical constraints. Our group at Johns Hopkins University, together with Pacific Northwest National Laboratory, recently completed the proteomic profiling of 174 TCGA high-grade serous ovarian cancer. I will discuss our preliminary attempts on integrating knowledge from such large-scale omics data into the development of IVDMIA tests.
Dr. Zhang is the Associate Director of the Center for Biomarker Discovery and Translation at JHU. His research focuses on the design of large-scale discovery studies for biomarkers, and their translation into clinical applications. He received his PhD in Electrical Engineering with additional trainings in mathematics and computational sciences. He was one of several groups in the 90s pioneered the application of neural networks and other machine learning algorithms to improve in vitro diagnostic tests. He was responsible for the IVDMIA development and the design of multi-center clinical trials that led to the FDA clearance of OVA1 and Overa. Over the years, he developed a diverse portfolio of research interests and expertise ranging from learning algorithm development, omics data analysis, and wet-lab experiments and assay development for biomarker functional and clinical validation.
The videos from past presentations can be found here.
The conference call will begin at 7:30 am PST (9:30 am CST, 3:30 pm GMT) via GoToMeeting and presentations will begin at 8:00 am PDT (10:00 am CDT, 4:00 pm GMT). The meeting ID is 992-810-669. If you would like to dial-in, use the following information:
United States: +1 (646) 982-0002
United Kingdom: +44 (0) 20 3713 5028