- Why Research
- Our Impact
- Get Involved
- About BCRF
- Research is the reason
- Contact Us
You are here
David Agus, MD
Professor of Medicine and Engineering
Keck School of Medicine and Viterbi School of Engineering
Director, USC Center for Applied Molecular Medicine
Director, USC Westside Cancer Center
University of Southern California, Norris Comprehensive Cancer Center
Los Angeles, California
Creating novel tools for more precise cancer diagnoses
Machine learning techniques are utilized to define cellular composition of tumors from biopsy images.
These approaches are advancing precision medicine and the ability to deliver the right drug for each patient.
Traditional pathology relies on human interpretation of tissue patterns observed under a microscope. These interpretations form the basis of a cancer diagnosis and treatment decisions. Dr. Agus is directing a team of diverse scientists to apply machine learning to create a programming language to enhance the information gleaned from a digital imagine and refine both diagnosis and prognosis predictions.
Full Research Summary
Cancer tissue contains information that is difficult to discern by eye, but that can be assessed with deep neural networks (the technology used to train self-driving cars). Dr. Agus and his team have used these networks to predict breast cancer biomarker status through biopsy images, laying the foundation to directly predict patient response to treatment.
With support from BCRF, Dr. Agus is leading an international team of data scientists, mathematicians, biologists, and clinicians to develop a "data language" for cancer cells and tissues. Using machine learning techniques similar to those used to detect faces in digital photographs, they developed a novel digital tissue fingerprint that can distinguish cancer tissues from normal tissues, as well as the major types of breast cancer.
This year, they are using these methods to learn the tumor patterns that best predict survival or recurrence during therapy. First, they will predict clinical response biomarkers, such as estrogen receptor and HER2 tumor status, from tumor patterns. Second, they will detect immune cells in a tumor and relate their patterns to patient survival. They have accumulated many tumor microscopy images to validate the results of combining these efforts into a unified machine learning approach and hope to learn fundamental insights into the biology of microscopic tumor patterns that predict patient outcome.
In addition to the value in predicting response to therapy, the team believes that the additional biological insights will provide researchers new directions for targeting breast cancer.
In the future, this automated platform could be readily expanded to offer unbiased, data-driven recommendations toward the care of all cancer patients, even in regions without access to dedicated pathology staff
Dr. David B. Agus is a professor of medicine and engineering at the University of Southern California Keck School of Medicine and Viterbi School of Engineering and heads the USC Westside Cancer Center and the USC Center for Applied Molecular Medicine. Dr. Agus leads a multidisciplinary team of researchers dedicated to the development and use of technologies to guide doctors in making health-care decisions tailored to individual needs, and directs a National Cancer Institute Physical Sciences in Oncology Center at USC. He is a medial oncologist, and the co-founder of two personalized medicine companies, Navigenics and Applied Proteomics. Dr. Agus is an international leader in new technologies and approaches for personalized healthcare. Dr. Agus’ first book called "The End of Illness" was published in 2012 and is a New York Times #1 and international best seller and was the subject of a PBS series, and his most recent book "A Short Guide to a Long Life" is also a New York Times and international bestseller.
BCRF Investigator Since
The Blizzard Entertainment Award