The Impact of Autologous Stem Cell Transplantation on the Genetics of High-Risk Relapsed Multiple Myeloma

Introduction: Autologous stem cell transplant (ASCT) followed by maintenance lenalidomide remains standard of care for newly diagnosed eligible myeloma patients. An understanding of how these treatments impact the genetic profile of the myeloma clone at relapse are lacking. We have previously shown that depth of response impacts clonal evolution at relapse. Here we explore the impact of melphalan conditioned ASCT as an additional factor in inducing genetic change. We used whole exome sequencing (WES) data from paired samples for a series of 56 newly diagnosed patients treated in the NCRI Myeloma XI Trial.

Methods: The Myeloma XI trial recruited patients to both ASCT and non-ASCT pathways. Within both pathways patients were randomised to an induction regime of either thalidomide or lenalidomide with cyclophosphamide and dexamethasone. Following induction, and ASCT if eligible, patients were further randomised to observation or lenalidomide maintenance. WES was undertaken on presentation and relapse malignant plasma cells (122x, n=22 ASCT, 34 non-ASCT). The mutational load and profile, structural aberrations, and evolutionary mechanism leading to relapse was determined. The series of patients were all phenotypically high risk, defined as relapse within 30 months of treatment initiation. The median PFS from maintenance randomisation was 19 months. Best response prior to relapse was determined for all; CR/nCR = 50% ASCT, 39% non-ASCT, VGPR/PR = 50% ASCT, 61% non-ASCT.

Results: In the patients who had undergone ASCT there was a significant increase in non-synonymous (NS) mutations between presentation and relapse (32 vs 49, p=0.005). This was not seen in the non-ASCT patients (43 vs 44, p=0.53). When breaking down the mutational load according to response, only ASCT patients achieving a CR/nCR had a significant increase in the NS mutational load from 28 to 58 (p=0.003).

We profiled 24 mutations known to be recurrent in myeloma. To infer possible clonal evolution, we determined whether there was a change in the profile of the mutations at relapse i.e. gain of new mutations at relapse or loss of mutations noted at presentation. In the ASCT patients 55% (12/22) had a change in the profile of these mutations, compared to only 29% (10/34) non-ASCT patients (p=0.09). Depth of response did not impact on this in the ASCT patients. In the non-ASCT patients, there was a difference between response groups, with 62% (8/13) of patients in the CR/nCR having a change in the profile, compared to 10% (2/21) of the VGPR/PR patients (p=0.002).

We assessed high-risk structural lesions. Acquisition of +1q during the disease course was noted as a new event in 18% of ASCT and 9% of non-ASCT patients (p=0.41). There was no impact on +1q change according to depth of response. Acquired tMYC was seen as a new event at relapse only in non-ASCT patients (n=3) with no apparent difference due to response. The tumour suppressor gene (TSG) regions of copy number loss -1p, -12, -14 and -17p were reviewed. ASCT patients were more likely to have a change in the profile of TSG deletions at relapse, with 41% of patients either gaining or losing evidence of a lesion, compared to 9% of the non-ASCT patients (p=0.007). Depth of response did not impact on the profile of TSG deletions.

Branching evolution was the dominant pathway leading to relapse, noted in 77% (17/22) of ASCT patients and 58% (20/34) non-ASCT patients (p=0.25). There was no impact on depth of response and evolutionary pathway according to response, although as per our previous findings, stable evolution was confined only to patients achieving a VGPR/PR.

Patients in the trial are randomised to lenalidomide maintenance or observation. There was no impact of maintenance strategy on the mutational profiles at relapse, in both the ASCT and non-ASCT patient groups.

Conclusion: We show that newly diagnosed, phenotypically high-risk patients who have undergone ASCT have a greater mutational load and significant change in the profile of TSG deletions at relapse, suggestive of clonal evolution. A deep response (CR/nCR) also led to a significant increase in the mutation load in ASCT patients and a change in the profile of recurrent mutations in non-ASCT patients, consistent with our previous work. We conclude that although depth of response is a key determinant of genetic evolution, exposure to the high dose melphalan may also play a role, consistent with the presence of mutational signatures described previously.