Utility of the Diagnosis-Specific Graded Prognostic Assessment for Prognostication in Leptomeningeal Disease.

Abstract

Leptomeningeal disease (LMD) is the spread of cancer cells to the arachnoid mater, pia mater, and cerebrospinal fluid. It occurs in 5% to 10% of solid organ cancers, with higher rates in breast, lung, and melanoma cancers. The prognosis for patients with LMD remains poor, with a median survival of 1.5 months without treatment and 2 to 3 months with treatment, despite advances in cancer treatment. This retrospective study included 64 patients with LMD with primary cancers represented in the diagnosis-specific Graded Prognostic Assessment (DS-GPA) at a single institution over 5 years. Patient characteristics, treatment, and overall survival (OS) data were collected. Statistical analyses included descriptive statistics, log-rank tests, and Cox proportional hazards regression models. The median OS for the 64 patients with LMD was 2.6 months, with no statistically significant differences among cancer types. Though not statistically significant, those with higher DS-GPA scores trended toward longer survival in breast and lung cancer cohorts. Patients with LMD on imaging confined to 1 location (cerebrum, cerebellum, spine, or cranial nerves) and receiving systemic chemotherapy alone also had longer survival. The DS-GPA tool is promising for LMD prognostication and may be strengthened by incorporating imaging and chemotherapy characteristics. Larger, multicenter studies are needed to validate its prognostic utility. Keywords: Leptomeningeal disease, diagnosis-specific graded prognostic assessment, prognosis, overall survival, breast cancer, lung cancer.