5284 - Second-line treatment of bevacizumab plus lomustine versus bevacizumab plus irinotecan in patients with recurrent glioblastoma

Background

Evidence for second-line treatment of recurrent glioblastoma (rGBM) is scarce and no standard of care is established. A Dutch phase II trial (BELOB) reported promising results with bevacizumab plus lomustine with improved overall survival (OS) and progression free survival (PFS). Therefore, a phase III trial (BELOREC) described maintenance of a PFS improvement; however, OS was not improved. Despite the negative trials, bevacizumab combined with chemotherapy is widely used in rGBM.

Methods

Retrospective analysis of adult patients with a first recurrence of a GBM (WHO grade 4) after temozolomide chemoradiotherapy, in the period between 2007 and 2017, at our institution. We selected patients treated with bevacizumab 10 mg/kg every 2 weeks and irinotecan 125 mg/m² every 2 weeks (BVZ+IRI) or with lomustine 90 mg/m² every 6 weeks (BVZ+LOM). Survival analysis was performed using the Kaplan-Meier method and prognostic factors assessed by univariate analysis and by the Cox regression model.

Results

Of the 154 patients included, the median age was 57 years (19-77) and 64.3% (n = 99) were men. Almost 90% were ECOG-PS 0-1 and the median treatment time was 3 months (0-29). About 70% (n = 108) of the patients were treated with BVZ+IRI and almost 30% (n = 46) with BVZ+LOM. 9-months OS was 37.6% in the BVZ+IRI group, with a median of 7 months (95% CI 5.92-8.08), and 57.3% in the BVZ+LOM group, with a median of 9 months (95% CI 4.55-13.45). For 6-months PFS, it was 35.9% in the BVZ+IRI group, with a median of 4 months (95% CI 3.13-4.87), and 40.6% in the BVZ+LOM group, with a median of 5 months (95% CI 4.29-5.71). In the univariate analysis, BVZ+LOM (p = 0.04), corticotherapy doses response (p = 0.03) and proteinuria (p = 0.04) were prognostic factors for OS. BVZ+LOM (HR 0.49; IC 95% 0.25-0.94; p = 0.03) and corticotherapy doses response (HR 0.52; IC 95% 0.28-0.98; p = 0.04) had a positive impact on OS in multivariate analysis. The most frequent toxicities were trombocitopenia (34.3% in BVZ+IRI and 43.5% in BVZ+LOM; p = 0.28), neutropenia (28.7% and 21.7%; p = 0.37) and leucopenia (23.1% and 17.4%; p = 0.43).

Conclusions

According to our data, and despite the samples size difference, BVZ+LOM has a positive prognostic impact in OS comparing to BVZ+IRI.

Clinical trial identification

Legal entity responsible for the study

Joana Simões.

Funding

Has not received any funding.

Editorial Acknowledgement

Disclosure

All authors have declared no conflicts of interest.