(Boston) – Chronic kidney disease (CKD) is caused by diabetes and high blood pressure. In 2017, the global prevalence of CRF was 9.1%, or approximately 700 million cases. Chronic kidney damage is assessed by noting the amount of interstitial fibrosis and tubular atrophy (IFTA) in a kidney biopsy sample. Although image scanning and morphometry techniques (measuring external shapes and dimensions) can better quantify the extent of histological damage, a more widely applicable way to stratify the severity of kidney disease is needed.
Now, researchers at Boston University School of Medicine (BUSM) have developed a new artificial intelligence (AI) tool to predict the grade of IFTA, a known structural correlate of progressive and chronic kidney disease.
âHaving a computer model that can mimic the workflow of an expert pathologist and assess the grade of disease is an exciting idea because this technology has the potential to increase the efficiency of clinical practices,â the author explained. correspondent explained correspondent author Vijaya B. Kolachalama, PhD, assistant professor of medicine at BUSM.
The typical microscopic pathologist workflow involves manual operations such as panning as well as zooming in and out of specific regions on the slide to assess various aspects of the pathology. In the “zoom out” assessment, pathologists look at the entire slide and perform an “overall” assessment of the renal nucleus. In the âzoom inâ assessment, they perform a thorough microscopic assessment of the âlocalâ pathology in the regions of interest.
An international team of five practicing nephropathologists independently determined IFTA scores on the same set of human kidney biopsies scanned using web software (PixelView, deepPath Inc.). Their mean scores were taken as a baseline estimate to build the deep learning model. To mimic the nephropathologist’s approach to classifying biopsy slides under a microscope, the researchers used AI to incorporate patterns and features of subregions (or patches) of the scanned kidney biopsy image as well as the Full (global) scanned image to quantify the extent of IFTA. Using this combination of patch-level and global-level data, a deep learning model was designed to accurately predict the IFTA score.
Once validated, Kolachalama believes AI models that can automatically assess the extent of chronic kidney damage can serve as second opinion tools in clinical practice. “Eventually, it may be possible to use this algorithm to study other organ-specific pathologies focused on assessing fibrosis. Such methods may have the potential to give more reproducible IFTA readings than readings. by nephropathologists, “he adds.
These results appear online in the American Journal of Pathology.
Funding for this study was provided in part by the Karen Toffler Charitable Trust, the American Heart Association (17SDG33670323 & 20SFRN35460031), the Hariri Institute for Computing and Computational Science & Engineering and Digital Health Initiative at Boston University, and the National Institutes of Health (R21 -CA253498) at VBK.
Warning: AAAS and EurekAlert! are not responsible for the accuracy of any press releases posted on EurekAlert! by contributing institutions or for the use of any information via the EurekAlert system.