The University of Florida’s Computational Microscopy Imaging (CMI) Lab recently made significant strides in nephrology research, showcased during the American Society of Nephrology (ASN) Kidney Week event held in Philadelphia, PA, from November 2nd to November 5th, 2023. Presenting a total of 8 abstracts, the lab’s groundbreaking work elucidated crucial aspects of kidney health and disease. ASN Kidney Week, a vital annual conference in the nephrology community, provided a platform for experts to share insights and advancements in kidney health and disease research.
The CMI Lab, known for its innovative approach to computational microscopy and imaging techniques, presented a diverse range of abstracts covering various facets of kidney health and machine learning. The research topics included disease mechanisms and advanced imaging methodologies, showcasing the lab’s multifaceted contributions to the field.
Automated Scoring of Peritubular Capillary Inflammation
Akshita Gupta, PhD student with the University of Florida’s Department of Health Outcomes and Bioinformatics, and colleagues developed an automated pipeline for quantifying peritubular capillary inflammation in kidney transplantation rejection. The innovative approach utilizing dual immunostaining for endothelium and leukocytes proved highly efficient, demonstrating a 4X increase in speed compared to manual scoring.
Machine Learning for Diabetic Nephropathy Classification
As presented by post-doctoral associate, Dr. Fatemeh Asfari, Myles Joshua T. Tan and colleagues focused on distinguishing between typical and atypical cases of diabetic nephropathy using machine learning. Their approach, based on quantitative morphological features extracted from glomeruli, showed promising results in differentiating atypical cases, potentially informing treatment decisions.
Multimodal Data Analysis with Spatial Transcriptomics
PhD candidate in biomedical engineering, Samuel Border, and colleagues delved into spatial transcriptomics methods to link histological observations with molecular mechanisms. Their visualization tool, FUSION, coupled with the SpotNet prediction model, showcased impressive accuracy in characterizing cell types and states in tissue units.
Predicting Membranous Nephropathy Progression
As presented by post-doctoral associate, Dr. Ahmed Naglah, Jamie L. Fermin, PhD student in electrical and computer engineering, and colleagues used an attention U-Net model to segment electron micrographs, extracting morphological features for predicting the progression of membranous nephropathy. While clinical and morphological features offered insights into risk levels, they proved less effective in predicting remission.
Integration of CODEX and Brightfield Histology for Cell Segmentation
PhD student in biomedical engineering, Nicholas Lucarelli, and team employed deep learning to segment and classify cell nuclei from renal tissue sections. Their work successfully automated the analysis, providing spatial context to protein expression, promising advancements in renal pathology.
Interdisciplinary Collaborations in Nephrology Research
Dr. Pinaki Sarder, the Principal Investigator of the CMI Lab, coauthored three additional posters in collaboration with colleagues from the Kidney Precision Medicine Project (KPMP) and the Human BioMolecular Atlas Program (HuBMAP). These collaborative efforts highlighted the lab’s interdisciplinary approach to addressing challenges in nephrology using computational and machine learning approaches.
ASN Kidney Week served as a significant forum for the CMI Lab, highlighting their commitment to advancing nephrology through innovative methodologies. The research presented underscores the lab’s dedication to unraveling complexities in kidney diseases, offering potential implications for improved diagnostics and treatment strategies.
