A match-adjusted indirect treatment comparison of lenalidomide plus rituximab versus rituximab plus chemotherapy for relapsed and/or refractory follicular lymphoma.

Purpose of the Study: The RELEVANCE trial found similar efficacy and a different safety profile for lenalidomide plus rituximab (R2) com- pared with rituximab plus chemotherapy (R-chemo) in patients with advanced untreated follicular lymphoma (FL) [1]. In the absence of head-to-head trials in the relapsed/refractory (R/R) setting, we com- pared the efficacy and safety of R2 with bendamustine and rituximab (BR) and rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) for the treatment of R/R FL using a match- adjusted indirect treatment comparison (MAIC).

Summarized Description of the Project: A systematic literature review of clinical trials and observational studies reporting survival and adverse events (AEs) of systemic treatments for R/R FL up to September 1, 2017 was conducted using MEDLINE, EMBASE, CEN- TRAL, and ClinicalTrials.gov databases, and EU and US conference proceedings. Primary outcomes of interest were progression-free sur- vival (PFS) and overall survival (OS). The following AEs were analyzed: all-grade nausea/vomiting, rash, and pneumonia; grade 34 neutrope- nia, febrile neutropenia, and diarrhea/colitis; and serious diarrhea/ colitis.

Pseudo-individual patient data (IPD) were generated from digitized KaplanMeier (KM) curves, if available, and compared with IPD from the AUGMENT and MAGNIFY R2 trials. An unanchored MAIC was conducted to adjust for potential effect modifiers and prognostic vari- ables (EM/PV) between trials. EM/PV were adjusted for based on clin- ical input and data availability. Cox proportional hazards and parametric survival models were fitted to match-adjusted IPD and pseudo-IPD. Match-adjusted AE data were compared using odds ratios (ORs).

Results: Aggregate data from 3 BR studies [24] and 1 R-CHOP study [5] were compared with the AUGMENT [6] and MAGNIFY [7] data. When the outcomes of interest were considered, PFS did not significantly differ between R2 and R-CHOP (hazard ratio [HR] 0.94, 95% confidence interval [CI] 0.521.70; P = 0.837) or R2 and BR (HR 1.26, 95% CI 0.702.27; P = 0.441) (Table). OS did not significantly differ between R2 and R-CHOP (HR 0.77, 95% CI 0.252.43; P = 0.660); no OS data were available for BR. No significant differences in AEs were observed between R2 and BR except for lower rates of nausea/ vomiting for R2 (17% vs 50%; OR 0.21, 95% CI 0.090.47; P = 0.0001).

Grade 3 neutropenia, the only AE of interest with data reported for R- CHOP, was 48% for R2 versus 55% for R-CHOP (OR 0.77, 95% CI 0.421.40; P = 0.386).

Conclusion: The ITC showed comparable outcomes between R2 and R-chemo; however, due to important limitations, results must be interpreted with caution. Point estimates varied depending on the ability to match for variables, including prior rituximab exposure, which was limited in the BR study with available PFS data and not possible for the R-CHOP study as all patients were rituximab-naive. Limitations include the small patient numbers, the inability to simultaneously adjust for all potential EM/PV, and the heterogeneity associated with prior rituximab exposure. Future studies should use real-world data to further explore the differences between R2 and R-chemo.

References

1. Morschhauser F, et al. N Engl J Med. 2018;379:934-947.
2. Rummel M, et al. Lancet Oncol. 2016;17:57-66.
3. Matsumoto K, et al. Int J Hematol. 2015;101:554-562.
4. Soler Campos JA, et al. Haematologica. 2017;102 Suppl 2:PB1887.
5. Van Oers MH, et al. Blood. 2006;108:3295-3301.

6. Leonard JP, et al. J Clin Oncol. 2019;37:1188-1199.
7. Andorsky DJ, et al. J Clin Oncol. 2019;37 Suppl 15:7513.