We have developed, validated, and open-sourced algorithms to perform accurate segmentation of the articular cartilage and meniscus from knee MRI scans. Such segmentations are used to quantify morphological biomarkers as well as to establish quantitative MRI values from MRI sequences that we have previously developed. To mitigate the paucity of labeled training datasets for image segmentation, we are developing data-efficient techniques using principles of self-supervision. In conjunction with the datasets that we have shared with the research community that enable accelerated MRI, we are developing end-to-end techniques for facilitating rapid MRI acquisition and automated analysis.

Select Publications:

• Desai A, Caliva F, Iriondo C, Khosravan N, Mortazi A, Jambawalikar S, Torigian D, Ellerman J, Akcakaya M, Bagci U, Tibrewala R, Flament I, O’Brian M, Majumdar S, Perslev M, Pai A, Igel C, Dam E, Gaj S, Yang M, Nakamura K, Li X, Deniz C, Juras V, Regatte, Gold G, Hargreaves B, Pedoia V, and Chaudhari A. The International Workshop on Osteoarthritis Imaging Knee MRI Segmentation Challenge: A Multi-Institute Evaluation and Analysis Framework on a Standardized Dataset. Radiology: Artificial Intelligence (2021) 3:3. doi: 10.1148/ryai.2021200078
• Wirth W, Eckstein F, Kemnitz J, Baumgartner C, Konukoglu E, Furst D, and Chaudhari A. Accuracy and Longitudinal Reproducibility of Quantitative Femorotibial Cartilage Measures Derived from Automated U-Net-based Segmentation of Two Different MRI Contrasts – Data from the Osteoarthritis Initiative Healthy Reference Cohort. Magnetic Resonance Materials in Physics, Biology and Medicine (2021). 34(3):337-354. doi: 10.1007/s10334-020-00889-7
• Eckstein F, Chaudhari A, Fuerst D, Gaisberger M, Kemnitz J, Baumgartner C, Konukoglu E, Hunter D, Wirth W. A Deep Learning Automated Segmentation Algorithm Accurately Detects Differences in Longitudinal Cartilage Thickness Loss – Data from the FNIH Biomarkers Study of the Osteoarthritis Initiative. Arthritis Care & Research (2020). doi: 10.1002/acr.24539
• Schmidt A, Desai A, Watkins L, Crowder H, Mazzoli V, Rubin E, Lu Q, Black M, Kogan F, Gold G, Hargreaves B, and Chaudhari A. Generalizability of Deep-Learning Segmentation Algorithms for Measuring Cartilage and Meniscus Morphology and T2 Relaxation Times. Intl Soc Magn Reson Med, (virtual), 2021
• Dominic F, Desai A, Schmidt A, Rubin A, Gold G, Hargreaves B, and Chaudhari A. Self-Supervised Deep Learning for Knee MRI Segmentation using Limited Labeled Training Datasets. Intl Soc Magn Reson Med, (virtual), 2021.
• Desai A, Barbieri M, Mazzoli V, Rubin E, Black M, Watkins E, Gold G, Hargreaves B, and Chaudhari A. DOSMA: A Deep learning, Open-Source Framework for Musculoskeletal MRI Analysis. Intl Soc Magn Reson Med, Montreal, 2019.

Using routine abdominal computed tomography (CT) scans, we extract quantitative biomarkers that depict the status of muscle and adipose tissue. Such biomarkers have been linked to future disease onset, post-surgical outcomes, and all-cause mortality.

Select Publications:

• Desai A, Boutin R, Tan J, Lenchik L, and Chaudhari A. An Evaluation of Automated Body Composition Analysis from Abdominal Computed Tomography Scans using Deep Learning. Society of Advanced Body Imaging Annual Meeting, (virtual), 2020.
• Boutin RD, Barnard RT, Kim J, Tan JC, Chaudhari A, Lenchik L. Opportunistic CT Assessment of Biological Aging: Comparing 2D vs. 3D Metrics for Muscle and Adipose Tissue. Radiological Society of North America, Chicago, 2021
• Zambrano JM, Chaudhari A, Wentland A, Jeffrey B, Rubin D, and Patel B. Opportunistic Screening for Ischemic Heart Disease Risk Using Abdominopelvic Computed Tomography and Medical Record Data: A Multimodal Explainable Artificial Intelligence Approach. Society of Abdominal Radiology 2021 (virtual).