LBA52 - A randomized phase Ib/II study of the selective small molecule AXL inhibitor bemcentinib in combination with either dabrafenib/trametinib or pembrolizumab in patients with metastatic melanoma

Background

AXL has been linked with both a reduced response to PD-1 blockade and resistance to BRAF directed therapies in melanoma. Bemcentinib (Bem) is a first-in-class, orally bioavailable, highly selective AXL inhibitor. BGBIL006 (NCT02872259) is an open label Phase Ib/II clinical study designed to explore whether combinations with Bem are safe and improve overall response rate (ORR) relative to standard of care therapies in metastatic melanoma.

Methods

Six patients were enrolled in the dabrafenib/trametinib (D/T) + Bem dose finding part 1. In part 2, 68 patients received D/T or pembrolizumab (Pem) based on BRAF mutation status and tumour load and were randomized 2:1 to receive D/T or Pem with or without Bem. Upon progression, 17 BRAF+ patients switched from D/T to Pem or vice versa in part 3. Safety was assessed by NCI CTCAE v 4.03 and tumour responses were assessed using RECIST v1.1.

Results

Patients were enrolled between 2017 and 2022. At the interim database lock on August 21 2023, median follow time was 53 months (10-78). A daily dose of 200 mg Bem was well tolerated in combinations with Pem or D/T; 24% (18/73) of Grade ≥ 3 AEs were regarded as related to Bem. Most frequent Bem-related AEs were rash, diarrhoea, fatigue and increased transaminases. In the efficacy population (n=70), we did not observe significant differences in ORR, PFS or OS between standard treatments and the combinations with Bem. In pre-planned analyses of baseline biomarkers, we did not identify subgroups of patients with increased response to combinations with Bem compared to standard treatment. The following biomarkers predicted improved ORR to Pem; high CD8+ count (p<0.01), high FOXP3 count (p<0.01), high PD-L1 in TIL (p=0.02), high AXL expression in inflammatory cells (p=0.04). High FOXP3 count (above median) in tumours, compared with low, predicted increased PFS in Pem treated patients (HR 0.2, p<0.01).

Conclusions

Bem was well tolerated in combinations with Pem or D/T in melanoma, but did not increase responses or survival in the efficacy population nor in biomarker defined subgroups. High FOXP3 count in tumour infiltrating inflammatory cells was a strong predictive marker for response to Pem.

Clinical trial identification

NCT02872259.

Editorial acknowledgement

Legal entity responsible for the study

Department of Oncology, Haukeland University Hospital, Bergen, Norway.

Funding

KlinBeForsk, Program for treatment research Norway. The Norwegian Cancer Society. The Norwegian Research Council. BerGenBio ASA provided the study drug bemcentinib free of charge and provided pharmacovigilance.

Disclosure

C. Schuster: Financial Interests, Personal, Other, Travel Grant: BerGenBio; Financial Interests, Personal, Invited Speaker: BMS, Pierre Fabre. K. Davidsen: Financial Interests, Personal, Invited Speaker: BMS. J. Karlsen: Financial Interests, Personal, Invited Speaker: AstraZeneca, Pfizer, Novartis, Pierre Fabre. A.J. Rayford: Financial Interests, Personal, Full or part-time Employment, Part time: BergenBio ASA. H. Løvendahl Svendsen: Financial Interests, Personal, Stocks/Shares: BerGenBio ASA. B.T. Gjertsen: Financial Interests, Personal, Leadership Role: Alden Cancer Therapy II, Kinn Therapeutics; Financial Interests, Personal, Advisory Role: BerGenBio, Novartis, AbbVie, Pfizer, Jazz Pharma; Financial Interests, Personal, Funding: Mendus; Financial Interests, Personal, Licencing Fees or royalty for IP: Alden Cancer Therapy II. L. Akslen: Financial Interests, Personal, Stocks/Shares: BerGenBio. J. Lorens: Financial Interests, Personal, Stocks/Shares: BerGenBio; Financial Interests, Personal, Other, Previously employed by BerGenBio: BerGenBio. All other authors have declared no conflicts of interest.