Long-Term Remission (LTR) Assumptions in Health Technology Assessments (HTAs) of Bispecific Treatments for Relapsed or Refractory Diffuse Large B-Cell Lymphoma (R/R DLBCL) Following Two or More Lines of Systemic Therapy (3L+)

OBJECTIVES: Bispecific antibodies (BsAbs) represent promising treatment options for patients with 3L+ R/R DLBCL. Economic models submitted to HTA agencies included LTR assumptions which substantially impacted survival extrapolations and cost-effectiveness results. This study investigated LTR assumptions in HTA assessments of BsAbs for R/R DLBCL.

METHODS: National Institute for Health and Care Excellence (NICE) technology appraisals (TAs) and submissions to the Canadian Agency for Drugs and Technologies in Health (CADTH) for BsAbs in 3L+ R/R DLBCL were reviewed to understand the methodology of LTR assumptions, and the critiques made by HTA reviewers.

RESULTS: Glofitamab and epcoritamab NICE (TA927, TA954) and CADTH submissions were identified and reviewed. While NICE evaluated 3L+ R/R DLBCL, CADTH further considered only Chimeric Antigen Receptor T-cell-ineligible or -experienced patients. LTR assumptions were unaffected by this difference. In both NICE submissions, LTR was defined as no disease progression 2 years after treatment initiation, and decreased mortality after 3.5 (glofitamab) or 2 years (epcoritamab). Assumptions were justified based on previous non-BsAb models, validated by clinicians. Both submissions were critiqued, noting uncertainty around definition of cure and time-point for LTR compared to limited trial follow-up. LTR assumptions were accepted by NICE committees, albeit at 3 years for both BsAbs. In CADTH submissions, LTR was assumed at 2 (glofitamab) and 3 years (epcoritamab) after treatment initiation. CADTH removed all LTR assumptions and set post-progression outcomes equal across treatments due to glofitamab and epcoritamab being associated with longer post-progression survival than comparators without a sufficient causal explanation.

CONCLUSIONS: Heterogeneity exists in the applications and justifications of LTR assumptions, posing risks of inconsistent decision-making, with NICE and CADTH having different views of the modelling choices. Approaches to modelling LTR should be evidence-driven, clinically plausible, and require validation versus mature follow-up data, especially when associated with clinical benefits post-progression or after treatment stopping.