A Deep Learning Model for Preoperative Differentiation of Glioblastoma, Brain Metastasis, and Primary Central Nervous System Lymphoma: An External Validation Study

Advances in Primary Central Nervous System Lymphoma

PURPOSE OF REVIEW

Optimal initial management can have a significant impact in long-term outcome in primary CNS lymphoma. This article reviews recent advances and the state of the field.

RECENT FINDINGS

Genomic analysis of CSF cell-free DNA has emerged as a new diagnostic tool for PCNSL. Treatment options have likewise evolved, with mature data from first-line chemotherapy-based prospective trials disclosing excellent results in younger (< 60-65) patients, with a cure achieved in a majority. However, results in older patients remain dismal, with several new salvage options under investigation including BTK pathway-targeted therapies, and CAR-T cell treatments. Meanwhile, low-dose radiation has emerged as an additional alternative for consolidation therapy. For younger PCNSL patients, the goal of treatment is now a cure, with the next frontier being the development of therapies affording optimized neurocognitive outcome and lower toxicity. Treatment for older patients remains however an unmet need, with several promising clinical trials ongoing.

An accessible deep learning tool for voxel-wise classification of brain malignancies from perfusion MRI

Noninvasive differential diagnosis of brain tumors is currently based on the assessment of magnetic resonance imaging (MRI) coupled with dynamic susceptibility contrast (DSC). However, a definitive diagnosis often requires neurosurgical interventions that compromise patients' quality of life. We apply deep learning on DSC images from histology-confirmed patients with glioblastoma, metastasis, or lymphoma. The convolutional neural network trained on ∼50,000 voxels from 40 patients provides intratumor probability maps that yield clinical-grade diagnosis. Performance is tested in 400 additional cases and an external validation cohort of 128 patients. The tool reaches a three-way accuracy of 0.78, superior to the conventional MRI metrics cerebral blood volume (0.55) and percentage of signal recovery (0.59), showing high value as a support diagnostic tool. Our open-access software, Diagnosis In Susceptibility Contrast Enhancing Regions for Neuro-oncology (DISCERN), demonstrates its potential in aiding medical decisions for brain tumor diagnosis using standard-of-care MRI.

Featured Papers in NeuroSci

In this topical collection, Arsiwalla et al [...]