Conexiant’s news site is now MDSpire News. Learn more
Advertisement
AI may predict focal GBM recurrence
Optical analysis of resection-margin tissue identified tumor infiltration as the strongest contributor to predictions of where glioblastoma later recurred.
The study evaluated whether AI-derived tumor infiltration in resection-margin samples could predict focal glioblastoma recurrence.
2
The retrospective development cohort included 273 margin samples from 60 patients, and a separate prospective internal cohort included 98 samples from 20 patients.
3
AI-derived infiltration alone had a mean area under the curve (AUC) of 0.62, while the random forest model combining infiltration with clinical, radiographic, and molecular variables had a mean AUC of
4
The random forest model achieved an AUC of 0.80 in the separate internal validation cohort, and a reduced model using infiltration, age, and resection cavity volume had a mean cross-validation AUC of
5
The validation cohort came from the development cohort’s health system, and the study’s sampling limitations included nonstandardized margin sampling and unknown performance in patients with unresecta
Since Roswell Park Comprehensive Cancer Center first began offering Gamma Knife radiosurgery 27 years ago, the cancer center has emerged as a world leader in the procedure.
