Abstract
Glioblastoma (GBM) and primary central nervous system lymphoma (PCNSL) are aggressive brain tumors that require neurosurgical treatment. For targeted treatment, biopsy and histopathological identification of the tumor entity are necessary. Here, we present a protocol for diagnosing PCNSL using a convolutional neural network (CNN)-based algorithm. We describe steps for installing the u-LINNDA (user-optimized lymphoma identification through neural network detection aid) algorithm, data preparation and preprocessing, and predicting tumor entities using u-LINNDA. We then detail procedures for predicting tumor identity and inspecting the u-LINNDA report.