Direct Assessment of IgA and IgG Paraprotein By Hevylite Assay and Correlation to IMWG Response and Progression-Free Survival in Newly Diagnosed, Transplant-Eligible Multiple Myeloma Patients in the Prospective Phase III GMMG-MM5 Trial

Background: Response criteria in myeloma have been developed to predict treatment outcome and to compare results from clinical trials. The current IMWG response criteria (Kumar, Lancet Oncology 2016) are based on paraprotein assessment by immunofixation and electrophoresis of serum and urine. The hevylite (HLC) assay directly quantifies the kappa and lambda fractions of IgG, IgA and IgM as well as the corresponding kappa/lambda ratios. To compare the response assessment by IMWG criteria and by HLC we applied the test to serum samples from the prospective GMMG MM5 trial in newly diagnosed myeloma patients.

Patients and Methods: 502 transplant-eligible myeloma patients were randomised to induction with bortezomib/dexamethasone combined to either adriamycin or cyclophosphamide and to receive either continuous or response-adapted lenalidomide maintenance. After 3 cycles stem cells were collected and high-dose melphalan and autologous stem cell transplantation (HDT) was performed. Patients not in nCR or CR were planned to receive a second HDT. All patients received 2 cycles of lenalidomide consolidation followed by lenalidomide maintenance according to the initial randomisation. After each treatment phase response was assessed according to IMWG. Frozen serum samples from the same time points were used to perform the HLC test. The current analysis is focused on the response after induction and after the first HDT in patients with either IgA or IgG paraprotein. To quantify the paraprotein by HLC the difference between involved and uninvolved HLC was calculated and the ratio HLC difference after treatment/HLC difference at baseline was used as response parameter. HLC responses were defined as HLC-CR (normal HLC kappa/lambda ratio,IgA: 0.78-1.94, IgG: 1.12-3.21), HLC-VGPR (≥ 90% reduction in HLC difference), HLC-PR (between 50% and 89% reduction), HLC-SD (49% reduction to 24% increase) or HLC-PD (≥ 25% increase).

Results: Of the 502 patients 105 had IgA and 298 IgG paraprotein. Sufficient serum samples to assess HLC-responses were available for 292 (IgA: 86, IgG: 212) patients.), 18% had HLC-CR, 23% HLC-VGPR and 41% HLC-PR after induction, while IMWG-based responses were 4% CR, 13% nCR, 15% VGPR and 49% PR. Overall there was a moderate concordance between both assessments: Of 205 patients achieving at least a PR, 191 (93%) also had at least HLC-PR. For 81 patients with ≥ VGPR, 69 had ≥HLC-VGPR (85%). Conversely, of 207 patients with ≥ HLC-PR 161 (78%) showed ≥ PR by IMWG and 104 patients with ≥ HLC-VGPR 69 (66%) had ≥VGPR.

After first HDT the IMWG-responses were 18% CR, 22% nCR, 19% VGPR and 35% PR, while HLC yielded 50% HLC-CR, 24% HLC-VGPR and 18% HLC-PR. Of 214 patients with ≥ PR 201 had ≥ HLC-PR (94%), while of 134 patients with ≥ VGPR, 121 had ≥HLC-VGPR (90%). Of 169 patients with ≥ HLC VGPR 121 (72%) had ≥VGPR and ≥PR was found in 158/211 patients with ≥ HLC-PR (75%). To assess the prognostic value of both response assessments Kaplan-Meier landmark analyses of progression-free survival (PFS) and overall survival (OS) at 6 months for induction and 9 months for HDT were performed. CR and nCR after induction were associated with a better PFS than the other response categories (log-rank test: p<0.001), in contrast there was no clear trend for HLC response after induction. However patients with CR or nCR by IMWG had a trend for a longer PFS if they also achieved HLC-CR (log-rank test: p=0.063, median PFS not reached vs 37.8 months). This also translated into an excellent OS for this subgroup (87.7% at 60 months). Irrespective of IMWG response, achieving HLC-CR after induction led to a significant improvement in OS with 86.5% at 60 months versus 71.4% without HLC-CR (log-rank test: p=0.039) After HDT the IMWG categories CR, nCR, VGPR or PR did not affect PFS, whereas there was a trend for longer PFS for HLC-CR versus HLC-VGPR (median 47.8 vs. 37.3 months, p=0.106). Neither IMWG-response, HLC response nor HLC ratio after HDT showed influence on OS.

Conclusions: Our results show that HLC for IgA and IgG can identify patients achieving ≥PR and ≥ VGPR according to IMWG after induction and after HDT in newly diagnosed myeloma patients. HLC response categories were not correlated with PFS, however HLC-CR after induction identified a subgroup of patients with an excellent overall survival of 88% at 5 years. Further analyses will elucidate this finding by including MRD results and genetic risk factors.