Serum Heavy-Light Chains (HLC) and Free Light Chains (FLC) As Predictors For Early CR In Newly Diagnosed Myeloma Patients Treated With Carfilzomib, Lenalidomide, and Dexamethasone (CRd)

Effective anti-myeloma combination therapies including carfilzomib and lenaldiomide (Ln) induce deep clinical responses and durable remissions. Attainment of stringent (s)- complete response (CR), as well as negative minimal residual disease (MRD) by sensitive molecular techniques and multiparameter flowcytometry (MFC) in bone marrow is predictive of improved outcome. Here we report the association of HLC, FLC, and other prognostic factors to predict early CR and MRD in our ongoing prospective clinical CRd trials for multiple myeloma (MM) and high risk smoldering MM (SMM) patients.

Treatment consisted of 8 cycles of CRd combination, followed by 24 cycles of Ln maintenance; the treatment schedule for SMM and MM was the same. Serum FLC, and immunoglobulin HLC pairs were determined using the Freelite and Hevylite assay (Binding Site, Birmingham, UK). The clonal FLC and HLC burden was measured as the difference between involved and uninvolved FLC (FLC-diff) and involved and uninvolved HLC pair (HLC-diff), as well as FLC κ to λ ratio (FLC ratio; normal range 0.26 to 1.65), and HLC ratios IgGκ/IgGλ and IgAκ/IgAλ; normal ranges, 1.12 to 3.21 and 0.78 to 1.94, respectively. Patients were evaluated for clinical responses by standard criteria and for MRD by sensitive MFC. Patients with any type of CR (CR, near(n)CR or sCR) were divided into groups according to whether their best response occurred in cycles 1-4, 5-8 or 9+, or 1-4 vs. 5+. Dichotomous parameters were compared using Fisher’s exact test. The associations between cycle categories and dichotomous parameters were determined using a Cochran-Armitage test. Continuous parameters were compared between two groups using an exact Wilcoxon rank sum test, while continuous categories were compared across cycle categories using a Jonckheere-Terpstra test for trend. Logistic regression analysis was used to predict if any type of CR would occur within 1-4 cycles. All p-values are two-tailed and reported without formal adjustment for multiple comparisons.

We included 38 MM and 12 SMM patients (27 male, 23 female; median age 59.5; range 40–88). Among these patients, median M-spike was 2.5g/dL, median FLC-diff was 28.94mg/dL (range -0.06–1619.87); isotypes included 37–IgG, 7–IgA, 5–κ, and 1–λ; immunoparesis was present in 46 (92%); and abnormal FLC ratio in 47 (94%). Six patients with light chain only disease were excluded from HLC analyses; among 43 analyzed, all had abnormal HLC ratio, and median HLC-diff was 22.14 g/L (range 1.85–109.13). Of 8 patients, in whom M spike could not be precisely evaluated due to migration in beta region, accurate HLC pair analysis was possible in 7. A total of 33 patients achieved CR: 22–sCR, 3–CR, and 8–nCR (66%). The number of CRd cycles administered to first CR of any type was 1-4 in 15 patients, 5-8 in 12, and 9-17 in 6; median number of cycles was 6. Among patients who achieved CR, 23 CR/sCR + 4 nCR patients underwent MRD assessment, and all were negative. There was a strong trend of an association between a normalized FLC ratio after 1 cycle of CRd and all types of CR (p= 0.003) obtained in first 4 cycles. Similarly normalization of the HLC ratio after 2 cycles of CRd was significantly associated with obtaining a nCR/CR/sCR (p=0.0098) or sCR (p=0.001) in 1-4 cycles; CR was reached in 1-4 cycles in 7/9 (78%) patients who normalized their HLC ratios after 2 cycles, as compared with 4/15 (27%) among those who did not. In a univariate analysis, FLC-diff after 1 cycle of CRd (p=0.0035), and HLC-diff after 2 cycles of CRd (p=0.0092) was associated with a significant trend to achieving nCR/CR/sCR in 1-4 vs.5-8 vs.9+ cycles. In a multivariate model that initially included M spike, FLC-diff, normalized FLC ratio, HLC-diff, as well as normalized HLC ratio, only HLC-diff after 2 cycles of CRd (<2.6 g/L vs. ≥ 2.6 g/L) remained as an independent factor predictive of response. In this model, 90.9% (10/11) of those whose best response occurred in cycles 1-4, as well as 76.9% (10/13) of those whose best response occurred in cycles 5+ could be predicted by HLC-diff after 2 cycles of CRd.

Early monitoring (after 2 cycles) of HLC patterns is an independent predictor of rapid response among high risk SMM and newly diagnosed MM patients treated with CRd therapy. Biomarkers predictive of rapid responses to therapy have a major clinical potential as they will allow better tailoring of treatment with the goal to maximize survival and minimize toxicities.

Off Label Use: Treatment of smoldering multiple myeloma- carfilzomib and lenalidomide.