THURSDAY, Nov. 19, 2020 — An algorithm that combines deep learning on histology slides and patient clinical data may predict immune checkpoint inhibitor (ICI) response in patients with advanced melanoma, according to a study published online Nov. 19 in Clinical Cancer Research.
Paul Johannet, M.D., from the NYU Grossman School of Medicine in New York City, and colleagues used a training cohort from New York University and a validation cohort from Vanderbilt University to develop a multivariable classifier that integrates deep learning on histology specimens with clinical data to predict ICI response in advanced melanoma. A receiver operating characteristic curve was generated, and patients were stratified as high versus low risk for progression using an optimal threshold. Progression-free survival was compared between the groups. Validation of the classifier was performed on two slide scanners (Aperio AT2 and Leica SCN400).
The researchers found that the multivariable classifier predicted response with areas under the curve of 0.800 and 0.805 on images from the Aperio AT2 and Leica SCN400, respectively. Patients were accurately stratified into high versus low risk for disease progression with the classifier. For Vanderbilt patients, those classified as high versus low risk had significantly worse progression-free survival.
“We believe this computational approach has the potential for integration into clinical practice,” the authors write. “This could help oncologists identify patients who are at high versus low risk for progression through immunotherapy.”
Several authors disclosed financial ties to the biopharmaceutical industry; several authors disclosed holding patents related to this work.
Abstract/Full Text (subscription or payment may be required)
Source: Read Full Article