Epid-13. Predictors Of Histopathologic Results Of Brain Biopsy In Patients Suspected Of Secondary Cns Lymphoma

A Basit Khan,Dima Suki,Eric Goethe,Luis Fernandez,Rajan Patel,Jason Westin,T Linda Chi,Jeffrey Weinberg

DOI: https://doi.org/10.1093/neuonc/noae165.0549

2024-11-29

Neuro-Oncology

Abstract:Patients suspected of secondary CNS lymphoma (SCNSL) often undergo brain biopsies. Recent studies doubt biopsy utility due to high SCNSL diagnosis rates. We sought to enumerate histopathologic biopsy results in suspected SCNSL patients to evaluate utility of biopsy. We furthermore assess clinical and radiographic factors based on ability to distinguishing SCNSL from non-SCNSL. 109 patients with non-CNS lymphoma who subsequently presented with brain lesions that underwent biopsy at MD Anderson Cancer Center were included. Histopathologic diagnoses were obtained. Pre-operative imaging was assessed for contemporary radiologic diagnoses by obtaining the leading differential reported at the time of the study. Subsequently, a staff neuroradiologist, blinded to biopsy outcomes and clinical considerations, provided a retrospective diagnosis. 61 patients had SCNSL whereas 48 did not. Non-SCNSL biopsy results included infection (n=16, 33%), glioma (n=12, 25%), nonspecific (n=12, 25%), and other metastases (n=3, 6%). Patients with non-indolent lymphoma (DLBCL, mantle cell, Burkitt's, and lymphoblastic lymphoma) were more likely to have SCNSL on biopsy (p<0.001, OR 12.55). Patients with SCNSL had shorter time interval since initial diagnosis (median 18.1mo vs. 63.2mo, p=0.008). Systemic progression of disease at biopsy was not associated with SCNSL (p=0.35). There was no difference in age, stage, LDH, Albumin, ECOG, IPI, anatomic location, or B-symptoms at times of diagnosis or neurosurgery. No patient had positive CSF cytology prior to biopsy. Neither retrospective nor contemporary radiographic diagnoses had sufficient accuracy to replace biopsy. A positive retrospective diagnosis trended towards association with SCNSL pathology (sensitivity=0.94, specificity=0.23, PPV=0.65, NPV=0.70, p=0.08). A positive contemporary diagnosis was associated with SCNSL pathology (sensitivity=0.67, specificity=0.73, PPV=0.76, NPV=0.64, p<0.001). Similar biopsy proportions yielded SCNSL and non-SCNSL. Although some factors are associated with SCNSL, none can reliably distinguish patients. Given the divergent treatments for the reported pathologies, our study reiterates the necessity of biopsy in suspected SCNSL.

oncology,clinical neurology