JS01.5.A PHASE 2 TRIAL OF CARBOPLATIN AND BEVACIZUMAB FOR RECURRENT ADULT EPENDYMOMA. A CERN STUDY

Abstract

Abstract BACKGROUND Most adults with ependymoma undergo tumor resection at the time of diagnosis, which may be followed by radiation. At recurrence, re-resection and/or (re)-irradiation may be given, however, there are few established chemotherapy treatments. A previous retrospective report of 8 patients treated with carboplatin and bevacizumab showed a high response rate with 6 patients demonstrating an imaging response (Green, Neurology 2009). We sought to further investigate this regimen with a prospective trial. MATERIAL AND METHODS We performed a prospective phase 2 study in the CERN Adult Clinical Trials Network. Adult patients with recurrent or progressive ependymoma were enrolled to receive carboplatin (AUC =4-5) every 4 weeks for up to 6 cycles and bevacizumab at 10mg/kg every 2 weeks for one year, with the option to continue until progression or toxicity. The primary endpoint was 12-month PFS rate and >50% defined efficacy. Serial symptom burden measurement at baseline and at the time of disease evaluation using MD Anderson Symptom Inventory-brain tumor (MDASI-BT) or MDASI-Spine patient-reported outcomes (PROs) were used to evaluate the clinical impact of PFS. RESULTS A total of 22 patients with median age of 45 years were accrued and treated; 11 were women. WHO grade was 3 in 13 patients and grade 2 in 9 patients (3 with myxopapillary ependymoma) Ten patients had only spinal cord disease, 3 had both spinal cord and brain involvement and 9 patients had brain involvement alone (6 supratentorial, 3 infratentorial). Previous treatments included radiotherapy in all 22 patients and alkylating chemotherapy in 9 patients. Treatment was well tolerated with expected myelotoxicities and hypertension. The Kaplan-Meier calculated 12-Month PFS rate was 76.4% (95%CI 52.2%, 89.4%), median PFS = 18 months (95%CI 12.2, +∞). There were 2 partial responses (9.1%). Brain tumor responders (objective response or stable disease) showed reduction while non-responders had an increase in both neurologic and cognitive symptoms but similar report of other symptoms. Spine tumor responders and non-responders both showed worsening disease-related symptoms; autonomic symptoms worsened in responders. Activity related interference worsened for all patients. CONCLUSION This treatment regimen was safe and met the primary efficacy endpoint of 12-month PFS rate. The improvement in disease-related symptoms in brain tumor patients supports that the achieved disease stability was clinically meaningful, but the increased activity-related interference suggests that treatment-associated symptoms may impact work, general activity, and walking ability during treatment. Improvements in spine tumor disease-associated symptoms were not seen. A confirmatory trial is warranted to further investigate the findings and to determine if there are differences in response amongst ependymoma subtypes and tumor location.