Analysis of chemo-predictive assay for targeting cancer stem cells in 41 glioblastoma patients.

• DOI: 10.1200/jco.2017.35.15_suppl.e13544
• Publication Date: 2018-09-06T15:41:11Z
• Authors: Candace M Howard; Anthony Alberico; Jagan Valluri; Michael Griswold; Pier Paolo Claudio
• Keywords: Temozolomide

e13544 Background: We have developed a drug response assay (ChemoID) that identifies effective chemotherapy against cancer stem cells (CSCs) vs. bulk of tumor from panel potential treatments. A prospective study was conducted evaluating the use CSC in glioblastoma patients treated with standard-of-care. Methods: 41 were enrolled an IRB-approved protocol. Patients all standard-of-care TMZ plus radiation or without maximal surgery. prospectively monitored for response, time to recurrence, progression-free survival (PFS), and overall (OS). Results: For every 5% increase in-vitro cell kill by TMZ, 12-month patient (non-recurrence cancer) increased 2-fold, OR=2.2 (p=0.016). Similar associations test seen, OR=2.75 (p=0.07) each TMZ. Combining results single model yielded statistically supported association, OR=2.36 (p=0.036) but much attenuated remaining OR=1.46 (p=0.472). Areas under curve (AUC) [sensitivity/specificity] at optimal outpoints (>40% >55% kill) AUC=0.989 [Sens=100/Spec=97], 0.972 [100/89] 0.989 [100/97] only, combined models, respectively. Risk categorization improved 11% when using conjunction (risk reclassification non-event NRI = 0.111, p=0.030). Median recurrence 20 months positive versus only 3 negative test, whereas median 13 4 (>55% negative. favorable observed PFS OS. Results across 14 other treatments indicated 34/41 (83%) 27/41 (66%) potentially more alternative therapies may been chosen Bulk results, Conclusions: The ChemoID has accuracy assays.