A Multi-center Study on Intraoperative Glioma Grading via Deep Learning on Cryosection Pathology.

Abstract

Intraoperative glioma grading remains a significant challenge, primarily due to the diminished diagnostic attributable to the suboptimal quality of cryosectioned slides. Precise intraoperative diagnosis is instrumental in guiding surgical strategy to balance the resection extent and the neurological function preservation, thereby optimizing patient prognoses. This study developed a model for intraoperative glioma grading via deep learning on cryosectioned images, termed IGGC. The model was trained and validated on The Cancer Genome Atlas (TCGA) datasets and one cohort (n_train = 1603, n_validate = 628), and tested on five cohorts (n_test = 213). The IGGC model achieved an AUC value of 0.99 in differentiating between high grade glioma (HGG) and low grade glioma (LGG), and an AUC value of 0.96 in identifying grade 4 glioma. Integrated into the clinical workflow, the IGGC model assisted pathologists of varying experience levels in reducing inter-observer variability and enhancing diagnostic consistency. This integrated diagnostic model possesses the potential for clinical implementation, offering a time-efficient and highly accurate method for the three-grade classification of adult-type diffuse gliomas based on intraoperative cryosectioned slides.