University of Boston researchers developed a deep learning algorithm that can assess kidney disease with better accuracy than trained pathologists.
Detecting kidney damage is of great importance, and unlike many other diseases, symptoms often don’t appear until the disease is very advanced. Getting this diagnosis wrong can lead to a series of life-threatening conditions.
“This rapid, scalable method can be deployable in the form of software at the point of care, and holds the potential for substantial clinical impact, including augmenting clinical decision making for nephrologists,” the team wrote in their research paper.
The team used a separate dataset, not used in training, to inference and validate the data. The convolutional neural network was able to assess and classify the images with better accuracy than the pathologist-derived fibrosis score, the researchers said.
“Although the trained eyes of expert pathologists are able to gauge the severity of disease and to detect nuances of histopathology with remarkable accuracy, such expertise is not available in all locations, especially at a global level.” The researchers wrote in their research paper. “Moreover, there is an urgent need to standardize the quantification of pathological disease severity, such that the efficacy of therapies established in clinical trials can be applied to treat patients with equally severe disease in routine practice.”
The research paper was published in the journal, Kidney International Reports.