Differential Effect of Upfront Intensification Treatment in Genetically Defined Myeloma Risk Groups - a Combined Analysis of ISS, Del17p and SKY92 Scores in the EMN-02/HOVON-95 MM Trial

Background

The introduction of proteasome inhibitors (PIs) and immune modulatory drugs (IMiDs) to multiple myeloma (MM) treatment protocols has drastically improved the overall quality and duration of response. This rationalized challenging the role of upfront high dose melphalan (HDM) followed by autologous stem cell transplantation (ASCT) versus chemotherapy alone as intensification treatment in four recent phase III trials. Yet, these showed uniformly that upfront HDM is still superior to non-transplant intensification alternatives including PIs and/or IMiDs.

However, HDM's mechanism of action involves induction of interstrand crosslinks and double-strand DNA breaks that cause collateral damage in both healthy cells (toxicity) and tumor cells, which may induce strong selection pressure on the outgrowth of resistant clones at relapse. We therefore hypothesized that certain low-risk patients could save HDM as salvage treatment and that this would improve their overall survival (OS).

Methods

We chose to combine three robust classification tools in MM that historically have shown a consistent performance across diverse patient cohorts: (1) International Staging System (ISS), (2) deletion of chromosome 17p (del17p) by Fluorescent in Situ Hybridization (FISH) (cutoff ≥10%) and (3) SKY92 high-risk (HR) classifier, based on gene expression profiling (GEP) of MM cells. By combining these scores, we classified ultra-low risk (uLR), intermediate-risk (IR) and HR MM as follows:

uLR: ISS = 1, del17p = absent, SKY92 HR = absent.

IR: ISS = 2 or 3, del17p = absent, SKY92 HR = absent.

HR: ISS = 1, 2 or 3, del17p and/or SKY92 HR = present.

The effect of intensification treatment on progression-free survival (PFS) and OS in these risk groups was evaluated in the phase III EMN-02/HOVON-95 MM trial (EudraCT 2009-017903-28). We focused on chemotherapy-naive MM patients aged ≤65 years, who received bortezomib, cyclophosphamide and dexamethasone (3-4xVCD) as induction therapy, followed by randomization 1 (R1) between bortezomib, melphalan and prednisone (4xVMP) and 1xHDM+ASCT, R2 between bortezomib, lenalidomide and dexamethasone (2xVRD) versus no consolidation and lenalidomide maintenance for all. At baseline, bone marrow was collected in the MM biobank at Erasmus MC Cancer Institute, the Netherlands. On this, CD138+ enrichment was performed, followed by a flowcytometric purity check, FISH analysis for del17p and GEP using HG U133 Plus 2.0 arrays (Affymetrix). PFS was defined as the time from R1 to progression or death, whichever occurred first. P-values were not adjusted for multiple testing.

Results

In our biobank region, 435 patients were randomized between VMP and 1xHDM, of whom ISS data were available for all, del17p status for 87%, SKY92 classification for 41% and all three risk factors for 38% of patients. Of these, 33% were ISS stage I, 47% ISS stage II, 20% ISS stage III, 10% had a del17p and 20% were SKY92 HR. A total of 43 (26%), 83 (50%) and 41 (25%) patients were classified as uLR, IR and HR, respectively.

Our risk classification was both significantly associated with PFS (logrank p < 1x10^-5; 4-year PFS of 64% in uLR, 49% in IR and 24% in HR) and OS (p < 1x10^-4; 4-year OS of 85% in uLR, 77% in IR and 47% in HR).

HR patients had a superior PFS with 1xHDM over VMP (p = 0.003), confirming previous reports. OS was related to intensification arm in both uLR and HR patients, however, in opposite direction: uLR patients treated with VMP had a better OS than those treated with 1xHDM (p = 0.047, 4-year OS of 95% with VMP versus 72% with 1xHDM); whereas OS was better after 1xHDM in HR patients (p = 0.017, 4-year OS of 68% with 1xHDM versus 22% with VMP). We found a significant interaction between intensification treatment and risk group for both PFS (likelihood ratio p = 0.03) and OS (p = 0.005). This indicates that there is indeed a differential treatment effect on survival between risk groups.

Conclusions

• Although our findings are preliminary and need further validation, we are the first to identify a low-risk, transplant-eligible MM subgroup that - in contrast to high-risk MM patients - may benefit from receiving upfront VMP versus 1xHDM+ASCT.

• Moreover, this illustrates for the first time how combined molecular and clinical subclassification in MM may identify patients that benefit from alternative treatment strategies, such as prolonged bortezomib and lenalidomide. Ultimately, this may result in risk-adapted treatment approaches.