Molecular Profiling of Relapsed and Refractory Myeloma Identifies Frequent Bi-Allelic Driver Events — Interim Results of the 'Myeloma UK Seven' Biomarker Trial

Introduction

Pomalidomide and low-dose dexamethasone (PomDex) are approved for relapsed and refractory myeloma (RRMM) with Overall Response Rates of about 30%, including durable remissions. Combining PomDex with biologic agents can improve response, but there are currently no markers to predict which patients benefit from the addition of drugs to PomDex alone. Myeloma UK (MUK) seven is a multi-center UK phase II trial that is designed to generate evidence for improved outcomes and treatment stratification in RRMM by: 1) building a bio-repository for molecular characterization of RRMM and discovery of predictive markers for response 2) patient randomisation to Cyclophosphamide-PomDex (CyPomDex) or PomDex to investigate if the cost-effective CyPomDex combination increases progression free survival compared to PomDex 3) a flexible design that allows addition of trial arms with other promising PomDex combinations. We report high patient and investigator bio-sampling adherence and first comprehensive genetic profiling results in RRMM using a novel molecular assay platform.

Methods

MUKseven (NCT02406222) will in total recruit 250 RRMM patients who are currently randomised 1:1 to CyPomDex or PomDex therapy. Registered patients undergo bone marrow sampling at screening, Cycle 1 Day 14, Cycle 4 Day 14 and at disease progression, and peripheral blood sampling on Cycle 1 Day 1. Patient consent includes mandatory bio-sampling at screening. Material is shipped to central laboratories for processing and storage of CD138-purified (>95%) MM tumor and non-tumor bone marrow microenvironment cells. Participants are currently recruited from over 30 NHS Hospitals in the MUK Early Phase Clinical Network.

Digital multiplex ligation-dependent probe amplification (dMLPA®; MRC Holland) is a novel assay for high-throughput genome-wide copy number aberration (CNA) profiling. The MM specific D006 assay contains 160 probes that interrogate regions frequently affected by CNA in MM, including hotspots CDKN2C, CKS1B, MYC and regions encoding for therapeutic targets such as MCL1 or SLAMF7. In addition, 219 probes are distributed over all chromosomal arms and inform a virtual karyogram. The targeted design allows for analysis of over 80 patient samples in a single MiSeq (Illumina) run, using 10-20 ng DNA input material.

Results

Between March 2016 and July 2017 a total of 83 patients were randomised to MUKseven with a median follow up of 6 cycles (1-13). Bone marrow material was received by central laboratories for all patients (100%) at screening, 69/79 (87.3%) at Cycle 1 day 14, 47/54 (87.0%) at Cycle 4 day 14 and 13/29 (44.8%) at disease progression. Peripheral blood was received for 73/79 (92.4%) on Cycle 1 Day 1. These return rates are higher than those reported for recent commercial trials. Of importance, for 52% of patients, matched bone marrow samples from earlier disease stages collected as part of UK academic research studies were available.

We profiled the first 81 RRMM samples from trial entry with dMLPA and compared CNA frequency to a recently published newly diagnosed MM (NDMM) cohort which was assessed with MLPA (Shah V, Leukemia 2017). We found significantly more focal bi-allelic CNA of driver lesions in RRMM than in NDMM (P = 0.002). These included homozygous deletions of CDKN2C (6.2 vs 1.9%) and RB1 (3.7 vs 0.8%), amplifications of MYC (8.6 vs 2.3%), and heterozygous losses of TP53 (16.0 vs 9.6%). There were significantly more RRMM patients with one or more of the following bi-allelic changes: CDKN2C, RB1 or CYLD homozygous loss, or 1q amplification (18.5 vs 8.2%; P = 0.007), and 3.7% of patients in the RRMM cohort showed more than one bi-allelic change. We identified frequent homozygous deletions in NFκB pathway genes such as CYLD (4.9%), and recurrent amplifications of therapeutic targets like MCL1 and SLAMF7 on chromosome 1 (8.6% amplifications). The trial is recruiting and profiling of additional cases as well as longitudinal profiling of cases with material from earlier disease stages is ongoing.

Conclusion

We demonstrate highly successful conduct of a multi-center biomarker discovery trial by the Myeloma UK Clinical Trial Network and the UK Myeloma Research Alliance. Our results demonstrate that bi-allelic CNA increase significantly in RRMM and some of these may help guide therapy. Recruitment and analyses of additional cases enrolled into MUKseven are ongoing and updated data will be presented at the meeting.