AUTHORS
John M. Burke, MD, Moya Young2, Marcelo Bellesso, MD3, Mark Yan, PhD4, Matthew Sugidono, PharmD5, Connie Batlevi6, Adriano Sa, MD, MSc7, Lukas Enderlein8
AFFILIATIONS
1 Sarah Cannon Research Institute at Rocky Mountain Cancer Centers, Aurora, CO, United States
2 Kent and Canterbury Hospital, Kent, United Kingdom
3 Hospital Santa Paula - Rede Americas, Sao Paulo, Brazil
4 Hoffmann-La Roche, Mississauga, Canada
5 Genentech, Inc, South San Francisco, United States
6 Genentech, Inc, South San Francisco, United States
7 F. Hoffmann-La Roche, Basel, Switzerland
8 F. Hoffmann-La Roche, Basel, Switzerland
ABSTRACT
Background
In the absence of head-to-head clinical trials directly comparing polatuzumab vedotin plus R-CHP (Pola-R-CHP; POLARIX, NCT03274492) against tafasitamab and lenalidomide plus R-CHOP (Tafa-Len-R-CHOP; frontMIND, NCT04824092) for previously untreated diffuse large B-cell lymphoma (DLBCL), an anchored matching-adjusted indirect comparison (MAIC) was conducted. Because both regimens were evaluated against a shared R-CHOP control arm, this method enables an estimation of relative efficacy and safety across the two separate trials.
Methods
Individual patient data (IPD) from POLARIX were matched to the available aggregate baseline characteristics of the frontMIND cohort. The use of a shared R-CHOP control arm facilitated an anchored MAIC, which restricts adjustments to effect-modifying covariates. The base-case analysis incorporated mutually reported effect modifiers. Patients with an IPI score of 2 from POLARIX were excluded to align with frontMIND eligibility criteria. Reconstructed pseudo-IPD derived from published frontMIND Kaplan-Meier curves were utilized for time-to-event endpoints. Additionally, two sensitivity analyses were executed: the first omitted covariates with a high proportion of missing values, while the second treated "missing cell of origin (COO)" status as a matching factor. Relative comparisons between Pola-R-CHP and Tafa-Len-R-CHOP are expressed as hazard ratios (HRs) for time-to-event outcomes and odds ratios (ORs) for binary metrics.
Results
The analysis achieved an effective sample size (ESS) of ~423 against the reference frontMIND population of 899, with balance achieved across all included covariates. A favorable outcome for Pola-R-CHP is indicated by an HR < 1.0 for time-to-event endpoints, an OR < 1.0 for safety outcomes, and an OR > 1.0 for efficacy response rates (CRR and ORR). No statistically significant differences in efficacy endpoints were observed between Pola-R-CHP and Tafa-Len-R-CHOP. The adjusted HRs were 0.977 (95% CI: 0.647, 1.476; p=0.913) for investigator-assessed (INV) progression-free survival (PFS), 0.939 (95% CI: 0.563, 1.565; p=0.808) for overall survival (OS), and 0.883 (95% CI: 0.614, 1.27; p=0.504) for INV event-free survival (EFS). Additional efficacy endpoints included adjusted ORs for INV overall response rate (ORR) of 1.195 (95% CI: 0.656, 2.177; p=0.560) and an INV complete response rate (CRR) of 1.339 (0.803, 2.233; p=0.263).
Regarding tolerability, the safety profile significantly favored Pola-R-CHP. Patients receiving Pola-R-CHP experienced significantly fewer serious adverse events (SAEs) (OR 0.57, 95% CI: 0.348, 0.932, p=0.025), Grade 3-5 adverse events (AEs) (OR 0.538, 95% CI: 0.319, 0.908, p=0.020), and toxicities leading to treatment discontinuation (OR 0.178, 95% CI: 0.051, 0.63, p=0.007). Other safety findings trended similarly, such as neutropenia (OR 0.611, 95% CI: 0.373, 1.003, p=0.051), Grade 5 AEs (OR 0.525, 95% CI: 0.152, 1.82, p=0.310), and discontinuations of all therapies due to AEs (OR 0.508, 95% CI: 0.157, 1.64, p=0.257).
The results of both sensitivity analyses, one adding COO missingness as a matching variable and another removing all variables with high missingness (COO, double hit/triple hit status, and extranodal disease) from matching, remained consistent with the primary findings across all evaluated safety and efficacy parameters.
Conclusions
This exploratory anchored MAIC suggests that Pola-R-CHP, when used as initial therapy for patients with aggressive B-cell lymphomas with IPI 3-5, offers a favorable safety profile with no statistically significant differences in efficacy compared to Tafa-Len-R-CHOP. This safety advantage is particularly notable regarding treatment discontinuation due to toxicity, Grade ≥ 3 AEs, and SAEs.These observations held true across sensitivity testing. Nevertheless, as with any anchored indirect treatment comparison, these conclusions require cautious interpretation due to potential residual bias from unmeasured or imbalanced effect modifiers, different follow-up times, and the reliance on reconstructed pseudo-IPD for the frontMIND survival curves.
DISCLOSURES
John M. Burke, MD: Genentech/Roche - Other (including travel support); Genmab - Consultancy (Includes expert testimony); Genentech/Roche - Consultancy (Includes expert testimony); Natera - Consultancy (Includes expert testimony); Seagen - Consultancy (Includes expert testimony); Bristol Myers Squibb - Consultancy (Includes expert testimony); Immunome - Consultancy (Includes expert testimony); GSK - Consultancy (Includes expert testimony); CRISPR Therapeutics - Consultancy (Includes expert testimony); AstraZeneca - Consultancy (Includes expert testimony); BeOne - Consultancy (Includes expert testimony); Novartis - Consultancy (Includes expert testimony); AbbVie - Consultancy (Includes expert testimony); Regeneron - Consultancy (Includes expert testimony); Eli Lilly - Consultancy (Includes expert testimony)