AI Model FastGlioma Can Detect Glioma Infiltration During Surgery

THURSDAY, Nov. 21, 2024 (HealthDay News) — The artificial intelligence (AI) model FastGlioma can detect glioma infiltration during surgery, according to a study published online Nov. 13 in Nature.

Akhil Kondepudi, from the University of Michigan in Ann Arbor, and colleagues presented a visual foundation model for fast and accurate detection of glioma infiltration in fresh, unprocessed surgical tissue. FastGlioma was pretrained using large-scale self-supervision on rapid, label-free optical microscopy, using about 4 million images, and was then fine-tuned to output a normalized score indicating the degree of tumor infiltration. FastGlioma was then tested in a prospective, multicenter international testing cohort of 220 patients with diffuse glioma.

The researchers found that FastGlioma was able to detect and quantify the degree of tumor infiltration (average area under the curve, 92.1 ± 0.9 percent). For detecting tumor infiltration during surgery, FastGlioma outperformed image-guided and fluorescence-guided adjuncts by a wide margin in a head-to-head prospective study involving 129 patients. Across diverse patient demographics, medical centers, and diffuse glioma molecular subtypes, the performance of FastGlioma remained high. Zero-shot generalization was seen for other adult and pediatric brain tumor diagnoses with FastGlioma.

“This model is an innovative departure from existing surgical techniques by rapidly identifying tumor infiltration at microscopic resolution using AI, greatly reducing the risk of missing residual tumor in the area where a glioma is resected,” co-senior author Shawn Hervey-Jumper, M.D., from the University of California, San Francisco, said in a statement. “The development of FastGlioma can minimize the reliance on radiographic imaging, contrast enhancement, or fluorescent labels to achieve maximal tumor removal.”

Several authors are shareholders in Invenio Imaging, which was used in the study.

Abstract/Full Text