Brigham Research Institute Poster Session Site logo-1
Close this search box.

Sharifa Sahai, MS




Research Fellow




Research Trainee




Sharifa Sahai*, Richard J. Chen, Jana Lipkova, Chengkuan Chen, Tiffany Y. Chen, Drew Williamson, Judy J. Wang, Daniel Shao, Astrid Weins*, and Faisal Mahmood*

Multimodal AI for Renal Allograft Biopsy Assessment

I believe it’s important to participate in this symposium in order to show the faces behind research and highlight the achievements of women in medicine and science. Often women are not the first ones who are thought of when picturing a scientist. My research interests lie in Multimodal Medical Artificial Intelligence. I would like to create a model that is able to fuse all types of health data to create one cohesive and intuitive model. The outputs of this model could be used by physicians for diagnosis and prognosis. Currently I am focusing on creating models for kidney related disease.


There are over 100,000 kidney transplants annually. The manual assessment of renal allograft biopsies suffers from high observer variability (κ=0.22). The assessment is a complicated process involving several modalities and requires the expertise of renal pathologists which is not often available in low resource settings. Misdiagnosis or delay in treatment can have consequences ranging from partial transplant rejection to death.


Thus we propose MANTA (Multimodal AI for Renal Transplant Assessment), an objective and automated method for assessment of renal allograft rejection. MANTA utilizes weakly supervised deep learning multimodal fusion using gigapixel whole slide images and patients’ diagnosis as labels. MANTA fuses morphological features from H&E, PAS, Masson Trichrome and Jones Silver stains to get holistic predictive results.


MANTA achieves an AUC of 0.95 for assessing Interstitial Fibrosis and Tubular Atrophy, 0.82 for T-Cell mediated rejection and 0.81 for Antibody-Mediated rejection. We have gathered an international cohort of renal allograft biopsies in order to assess clinical performance across patient populations, sample preparation protocols and slide scanning instrumentation.


The evaluation of this AI system paves the way for clinical trials to establish the efficacy of AI-assisted renal allograft assessment to improve kidney transplant outcomes.