Daratumumab in first-line therapy is cost-effective in transplant-eligible patients with newly diagnosed myeloma

Abstract

Triplet regimens, such as lenalidomide, bortezomib, and dexamethasone (RVd) or thalidomide, bortezomib, and dexamethasone (VTd), are standard induction therapies for transplant-eligible patients with newly diagnosed multiple myeloma (NDMM). The addition of daratumumab to RVd and VTd has been investigated in the GRIFFIN and CASSIOPEIA trials, respectively, resulting in improvement in the rate of minimal residual disease (MRD) negativity. In this study, we conducted a cost-effectiveness analysis with a 10-year time horizon to compare first-line and second-line use of daratumumab for transplant-eligible patients with NDMM. Because long-term follow-up data for these clinical trials are not yet available, we developed a Markov model that uses MRD status to predict progression-free survival. Daratumumab was used either in the first-line setting in combination with RVd or VTd or in the second-line setting with carfilzomib plus dexamethasone (Kd). Quality-adjusted life-years (QALYs) and incremental cost-effectiveness ratios were calculated from a Japanese and US payer perspective. In the Japanese analysis, D-RVd showed higher QALYs (5.43 vs 5.18) and lower costs (¥64 479,793 vs ¥71 287 569) compared with RVd, and D-VTd showed higher QALYs (5.67 vs 5.42) and lower costs (¥43 600 310 vs ¥49 471,941) compared with VTd. Similarly, the US analysis demonstrated dominance of a strategy incorporating daratumumab in first-line treatment regimens. Given that overall costs are reduced and outcomes are improved when daratumumab is used as part of a first-line regimen, the economic analysis indicates that addition of daratumumab to first-line RVd and VTd regimens is a dominant strategy compared with reserving its use for the second-line setting.

Graphical abstract

Figure 1.

Construction of the Markov model. (A) Markov diagram. (B) Regimens used in each line of treatment. Bd, bortezomib and dexamethasone; EloPd, elotuzumab, pomalidomide, and dexamethasone; KRd, carfilzomib, lenalidomide, and dexamethasone; Obs, observation; R, lenalidomide.

Figure 2.

PFS curves according to parameters used in the model. PFS of (A) MRD⁻ and MRD⁺ patients after ASCT (IFM 2009), (B) Second-line DKd, Kd (CANDOR), and KRd (ASPIRE), (C) EloPd (ELOQUENT-3), and (D) panobinostat-Bd (PANORAMA-1). Survival curves are shown from published clinical trials; curves marked “estimated” in the key were calculated using the model.

Figure 3.

PFS and OS of first- and second-line daratumumab calculated by the models. (A) PFS, (B) OS by the GRIFFIN model, and (C) PFS, and (D) OS by the CASSIOPEIA model. Survival curves are shown from clinical trials; curves marked “estimated” in the key were calculated using the model.

Figure 4.

Costs per cycle and cumulative costs for first- and second-line daratumumab. Costs per Markov cycle (3 months) and cumulative costs over 10 years for the (A-B) GRIFFIN model and (C-D) CASSIOPEIA model.

Figure 5.

Tornado diagram of one-way sensitivity analysis of the GRIFFIN model (Japanese analysis). (A) Incremental cost and (B) incremental effectiveness of D-RVd vs RVd. Red bars represent changes in values when parameters were varied from base case to higher uncertainty value. Blue bars represent changes in values when parameters were varied from base case to lower uncertainty value. Only parameters that produce >¥1 000 000 change in costs and 0.01 QALY change in effectiveness were included in the diagram. *Probability of withdrawal from second-line treatment was the only parameter that may reverse the incremental cost of D-RVd vs RVd.

Figure 6.

Sensitivity analysis. Three-way sensitivity analysis varying the cost of carfilzomib, withdrawal rate from second-line treatment, and (A) rate of MRD-negativity with D-RVd, transition probability to calculate PFS of (B) DKd and (C) Kd, and (D) cost of daratumumab. Two withdrawal rates, 0.06 (base case) and 0.25 (25% of the patients withdraw from treatment without receiving second-line), were analyzed. Blue and red areas indicate that D-RVd and RVd are cost-effective within those values, respectively. Dashed black lines indicate the base case or other clinically important values.

Figure 7.

Scatterplot of PSA for the Japanese analysis. Results for (A) GRIFFIN and (B) CASSIOPEIA model. Red and green dots represent iterations in which RVd and D-RVd were cost-effective, respectively.