Anant Madabhushi

Professor Madabhushi's research centers on developing computational algorithms and machine learning frameworks for quantitative analysis of biomedical data. His work aims to enhance diagnostic and prognostic capabilities through integrating multimodal medical imaging, molecular data, and pathology. The research emphasizes translational applications to improve personalized medicine and clinical decision-making by leveraging advanced computational modeling and pattern recognition techniques.

Dr Madabhushi’s pioneering work in developing novel machine vision and AI based approaches for interrogating routinely acquired digital pathology and radiology scans is helping oncologists treat patients suffering from brain, lung, prostate, colorectal, breast, and blood cancer. Most recently, to help combat COVID-19, he adapted technology from his work on lung cancer to create a new artificial intelligence tool that uses computers to digitally analyze CT scans to identify patients who require ventilators for the coronavirus5. He places special emphasis on eliminating racial disparities in medicine by identifying underlying biological variations in disease appearance between populations, which can then be specifically targeted, to enable population-based personalized risk calculators. He is also taking this logic to another minority population, Veterans, where their unique exposure to wartime environments and particular lifestyle choices, make them different from the civilian population and thus require training with Veteran data.

His work also addresses critical issues in bioethics, health disparities, and health care costs6. For example, he is part of a group of researchers at CWRU and in Uganda and Tanzania who recently received a $5M NIH global health grant to establish an HIV-associated malignancy research center focused on lung cancer in East Africa. The team will investigate novel approaches to characterize lung cancer epidemiology, somatic mutation burden, HIV and accelerated aging, and radiological features of lung cancer and the relationship to HIV-1 infection.
 

Dr Madabhushi’s work places an emphasis on eliminating racial disparities in medicine by developing AI tools that employ population-specific risk-prediction models to identify underlying biological variations in disease appearance between races—allowing disease to be medically targeted with more accuracy. He also stresses the importance of recognizing potential bias in the data and understanding the diversity of the research pool. Under a new partnership, he will be working with Hampton University, a historically Black research university in Virginia, to develop AI tools for addressing cancer disparities. The technologies will help pinpoint which Black men undergoing proton therapy for prostate cancer are likely to respond to treatment. This partnership builds on Dr Madabhushi’s work showing critical cellular variations between Black and White prostate cancer patients (discussed below).

The research thrust of Dr Madabhushi and his team is developing algorithms and similar computational approaches to visually highlight and extract complex, “sub-visual,” minute physical features from image data from MRIs, CT scans, and digital-pathology tissue images. Physicians can use the presence of these biomarkers—which they cannot typically see when reviewing the images unaided—for detecting and treating serious illnesses. For example, Dr Madabhushi has developed image-segmentation, detection, and feature-analysis algorithms to pinpoint the structure and arrangement of various cells and nuclei (i.e., biomarkers) in pathology images, physical features which are correlated with many types of cancer, rates of progression, and potential to benefit from specific treatments, e.g., immunotherapy vs. chemotherapy.
 

Dr Madabhushi’s work makes maximum use of routinely acquired medical imaging scans and tissue slide images. The digital tests he is developing are relatively low cost and can be deployed through the cloud and the internet, enabling a global footprint. His work flows from the fact that no matter how well trained and medically experienced, the limited human eye will inevitably miss crucial signs and signals in CT scans, MRIs, and digital pathology. His research is built on two further premises. First, applying computer science, machine learning, and AI to clinical problems will transform our ability to personalize treatment (precision medicine)—providing the right treatment to the right patient at the right time. Resources are also thereby conserved for both health systems, e.g., fewer trials of multiple therapies, and individuals, e.g., fewer visits and co-payments, lower out-of-pocket expenses, and a reduction in lost wages.

Second, as a scientist, Dr Madabhushi is committed to addressing the diagnostic and predictive needs of his physician colleagues. In addition to those from Case Western Reserve, he has partnered with many physician-scientists, clinicians, and researchers from Cleveland Clinic, Louis Stokes Cleveland VA Medical Center, University Hospitals, and MetroHealth, a nationally ranked public health care system in Cleveland. As noted above, international partnerships with global colleagues are vital to his modus operandi, and he has more than 40 collaborations on six continents.

Dr Madabhushi’s approach to AI in health, stresses integrating domain knowledge from the medical and biomedical spectrum into sophisticated AI, machine learning, and computational-science templates—ensuring that clinicians play a central role in the solution. Care providers are more apt to trust and use technologies that they comprehend and to whose development they have contributed.