CHARM can also distinguish isocitrate dehydrogenase-mutant tumors from wild type, and classify three types of molecularly defined gliomas
By Elana Gotkine HealthDay Reporter
THURSDAY, July 13, 2023 (HealthDay News) — An artificial intelligence system can identify cancer cells and predict genetic mutation status of gliomas using intraoperative cryosection samples, according to a study published online July 7 in Med.
Noting that intraoperative cryosection evaluation remains the gold standard for guideline surgical treatments for gliomas, but that the tissue-freezing process often generates artifacts complicating histologic interpretation, MacLean P. Nasrallah, M.D., Ph.D., from the University of Pennsylvania in Philadelphia, and colleagues used samples from 1,524 glioma patients from three different patient populations to analyze cryosection slides and develop the context-aware Cryosection Histopathology Assessment and Review Machine (CHARM).
The researchers found that the CHARM models were able to identify malignant cells (area under the receiver operating characteristic curve [AUROC], 0.98 in the independent validation cohort), differentiate between isocitrate dehydrogenase (IDH)-mutant and wild type tumors (AUROC, 0.79 to 0.82), classify three main types of molecularly defined gliomas (AUROC, 0.88 to 0.93), and identify the most prevalent subtypes of IDH-mutant tumors (AUROC, 0.89 to 0.97). In addition, clinically important genetic alterations in low-grade glioma were predicted by CHARM, including ATRX, TP53, and CIC mutations, CHKN2A/B homozygous deletion, and 1p/19q codeletion via cryosection images.
“Right now, even state-of-the-art clinical practice cannot profile tumors molecularly during surgery,” a coauthor said in a statement. “Our tool overcomes this challenge by extracting thus-far untapped biomedical signals from frozen pathology slides.”
One author is an inventor on a patent, and a consultant for Curatio.DL.
Copyright © 2023 HealthDay. All rights reserved.
