Real-world outcomes from 3D-PREDICT glioma chemosensitivity testing of primary CNS tumors.

Abstract

2060 Background: 3D-Predict Glioma is an ex vivo chemosensitivity assay that predicts response to 12 systemic agents using patient-derived 3D cell cultures. Prospective studies showed predictive accuracy, but real-world data linking assay-guided therapy selection to clinical outcomes remain limited. We analyzed outcomes at a tertiary cancer center. Methods: This retrospective–prospective chart review included patients with primary CNS tumors who underwent chemosensitivity testing at an urban cancer center since 3/2023. Demographics, tumor characteristics, assay results for the 12-drug panel, treatment selection, and clinical outcomes were abstracted. Drug response was categorized as response, moderate response, or no response. Survival analyses were performed using the Kaplan–Meier method. Results: Among the 96 patients, ECOG was 0 in 21 (22.1%), 1 in 37 (38.9%), 2 in 26 (27.4%), and 3 in 11 (11.6%). Insurance coverage included private (69, 71.9%), Medicaid/Medicare (14, 14.5%), Private plus Medicaid/Medicare (12, 12.5%), or self-pay (1, 1.0%). At the time of analysis, 85 patients (88.5%) were alive. Median age was 64 years (range 25–86; IQR 56–71), and 53 (55.2%) were female. Most were diagnosed with adult-type diffuse glioma (92, 95.8%), including 84 (87.5%) glioblastoma. Chemosensitivity testing failed in 10 (10.4%) cases due to insufficient tissue quantity or quality control. All remaining cases yielded 1 or more drugs with assay-predicted responses. Response rates for each drug are detailed in the Table. Mean time to results was 8.8 days (range 5–15). Assay-guided therapy was selected in 59.5%, predominantly temozolomide. The remaining patients received other agents on trial, standard-of-care physician's choice, no further therapy, or transferred care. Key endpoints analyzed included progression-free survival (PFS), PFS at 6 months, overall survival (OS), OS at 12 months, and time-to-progression (TTP) ratio. Conclusions: In a real-world CNS tumor cohort, 3D-Predict Glioma identified potentially active agents and influenced treatment selection in more than half of patients. Our series suggests that chemosensitivity-guided therapy is feasible in routine neuro-oncology practice. Response rates by drug. Drug Newly Diagnosed Progression Pseudoprogression All Abemaciclib 17 (37.8) 10 (47.6) 2 (100.0) 29 (42.6) Carboplatin 30 (60.0) 11 (50.0) 2 (100.0) 43 (58.1) Dabrafenib 4 (8.0) 4 (17.4) 0 (0.0) 8 (10.7) Etoposide 13 (25.5) 5 (21.7) 0 (0.0) 18 (23.7) Everolimus 44 (100.0) 22 (100.0) 2 (100.0) 68 (100.0) Irinotecan 23 (45.1) 9 (34.6) 0 (0.0) 32 (40.0) Lomustine 2 (3.9) 1 (4.3) 0 (0.0) 3 (3.9) Osimertinib 8 (17.4) 5 (23.8) 0 (0.0) 13 (18.8) Procarbazine 50 (100.0) 23 (95.8) 2 (100.0) 75 (98.7) Rucaparib 43 (86.0) 17 (73.9) 2 (100.0) 62 (82.7) Temozolomide 20 (39.2) 10 (40.0) 2 (66.7) 32 (40.5) Trametinib 20 (44.4) 6 (28.6) 0 (0.0) 26 (38.2) Denominators defined by available results per drug; N (%).