Neurocognitive function (NCF) outcomes in patients with glioblastoma (GBM) enrolled in RTOG 0825.

Abstract

2004 Background: RTOG 0825 evaluated overall survival (OS) and progression-free survival (PFS) differences in patients with newly diagnosed GBM treated with standard chemoradiation, maintenance temozolomide and placebo (Arm 1) or bevacizumab (Arm 2). While OS was equivalent, PFS was longer in Arm 2. Longitudinal NCF testing was performed to evaluate clinical benefit. Methods: NCF was evaluatedat baseline and while on study and progression free with the Hopkins Verbal Learning Test-Revised (HVLT-R), Trail Making Test (TMT) and Controlled Oral Word Association (COWA). Change in NCF from baseline was categorized using the reliable change index. Differences between treatment arms were compared at follow-up evaluations. Additionally, baseline (B) and early changes (EC) (B to week 10) in NCF were examined as prognostic factors. Results: 542 patients consented and 507 randomized patients participated, with test completion rates at weeks 0 (B), 10, 22, and 34 of 94-97, 69-73, 59-64, and 53-58%, respectively. Mean test performance at B was equivalent between arms and ranged from -0.8 to -4.8 SDs below healthy population norms with global NCF on a composite variable at -2.0 SDs. There were no statistically significant between arm differences in frequency of improvement through week 34. Decline on COWA (verbal measure of executive function) at week 34 relative to baseline was more common (16.1 vs 5.7%) in patients in Arm 1 (p<0.05); whereas, there were trends for more decline in Arm 2 on a visuomotor measure of executive function (TMT B, p< 0.06; 22.2 vs 35.4%). B performance and EC in global NCF, memory, executive function and processing speed were prognostic for OS. At B, global NCF was prognostic for PFS. EC in global NCF, memory and executive function were prognostic for PFS. Conclusions: There was a statistically significant difference in the frequency of decline on a verbal test of executive function at week 34 favoring Arm 2. However, this was not found at earlier time points and was not found on a visuomotor test of executive function. B and EC in NCF were prognostic for OS and PFS. Longitudinal modeling is ongoing to further evaluate the impact of treatment on patients’ NCF. Support: U10 CA21661, U10 CA37422 and Genentech, Inc. Clinical trial information: NCI-2009-01670.