More precision for personalized medicine
Mount Sinai center aims to use computer vision technology for diagnosis and treatment
NEW YORK—The Department of Pathology at the Icahn School of Medicine at Mount Sinai has established the Center for Computational and Systems Pathology to explore efforts to more accurately classify diseases and guide treatment using computer vision and machine learning techniques. Using advanced computer science and mathematical techniques with cutting-edge microscope technology and artificial intelligence, the center hopes to “revolutionize pathology practice,” according to a press release.
The Center for Computational and Systems Pathology will be a hub for the development of new diagnostic, predictive and prognostic tests, and will partner with Mount Sinai-based Precise Medical Diagnostics (Precise MD), which has been under development for more than three years by a team of physicians, scientists, mathematicians, engineers and programmers.
Dr. Carlos Cordon-Cardo, who will oversee the new center located at Mount Sinai St. Luke’s, will continue to be chair of the department of pathology at the Mount Sinai Health System and a professor of pathology, genetics and genomic sciences and oncological sciences at the Icahn School of Medicine. Dr. Gerardo Fernandez, an associate professor of pathology and genetics and genomic sciences at the Icahn School of Medicine at Mount Sinai, will serve as the center’s medical director. He will work closely with Dr. Michael Donovan, a research professor of pathology at the Icahn School of Medicine, and Dr. Jack Zeineh, director of technology for Precise MD.
“Our goal is to provide a precise mathematical approach to classifying and treating disease, which will assist our clinicians with information for effective patient care and health management,” said Cordon-Cardo. “By refining diagnoses, we can save patients from unnecessary treatments.”
Mount Sinai’s Department of Pathology processes more than 80 million tests a year, reportedly making it the largest department of its kind in the country. “Computer vision and machine learning techniques are used by our group to remove subjectivity out of the process of microscopically characterizing disease,” Fernandez explained. “Image analysis and computational approaches open up the possibility of developing features that were previously too subtle or too complex to be robust and reproducible in an analog environment.”
Precise MD is developing new approaches to characterizing an individual’s cancer by combining multiple data sources and analyzing them with mathematical algorithms, offering a more sophisticated alternative to standard approaches. One such example is Precise MD’s approach to improve upon the Gleason score, a grading system that has been used since the 1960s to establish the prognosis for a prostate cancer and guide the patient’s treatment options.
“We’re characterizing tumors based on the combination of their architectural patterns and biomarkers,” said Fernandez. “Computer vision analysis, leveraging multispectral fluorescence microscopic imaging, enables us to see what the human eye cannot. Characterizing complex biomarker and morphological relationships in patient cohorts with known outcome allows us to identify patterns that correlate with behavior.”
He added, “Intramural collaborations are encouraged, making the precise platform available to researchers interested in a computational pathology and systems pathology approach to help answer question in their respective areas of research. The goal of each collaboration is dependent on the question being addressed by each researcher. Goals may include better prognostic stratification of patients or more accurate therapeutic guidelines among many others.”
In its initial phase, Precise MD will complete a test used for patients who have had prostatectomies at Mount Sinai Health System, to help determine which of them are more likely to have a recurrence of cancer and may need additional therapy such as chemotherapy. A second, higher-impact test, which will be used to characterize prostate cancer in newly diagnosed patients, will follow in 2017. At that time, Cordon-Cardo says all prostate cancer patients at Mount Sinai will have the option to receive this test.
It is anticipated that in 2017 other current efforts will yield additional novel computer vision and machine learning tools to better characterize breast cancer. The Center for Computational and Systems Pathology and the Precise MD platform could eventually be used to characterize any number of disease states, including but not limited to melanoma, lung and colon cancers, as well as chronic inflammatory conditions such as inflammatory bowel disease.
This works hopes to enable more precise cancer treatment “By using patient-specific phenotypic characteristics at the critical decision point to establish the likelihood that a patient’s cancer will be appropriately treated,” Fernandez summarized.