Pembrolizumab (Pembro) plus R-CHOP is efficacious as first-line therapy for primary mediastinal B-cell lymphoma (PMBL) with high rates of ctdna negativity – interim efficacy analysis of the ALLG-Pacific (NHL35) trial

Abstract Introduction: PMBL comprises 10% of diffuse large B-cell lymphoma, primarily affecting young adults. Chemoimmunotherapy is effective, but the optimal regimen is unclear. Pembro is active and tolerable in relapsed/refractory PMBL. There is rationale to combine R-CHOP and pembro as first-line therapy. Methods: ALLG-PACIFIC is a phase II, single arm, open-label study of R-CHOP with pembro for patients(pts) with newly diagnosed PMBL. Pts receive 2 cycles rituximab(R) (375mg/m2) + pembro (200mg) every(Q) 21 days(d) ('window') followed by 6 cycles of R-CHOP + pembro (200mg) Q21d ('induction') followed by 5 cycles of pembro (400mg) Q42d ('consolidation'). Adults ≥18 years with treatment naïve PMBL and adequate organ function, suitable for R-CHOP and no active autoimmune disease are eligible. PET scans are performed at baseline, end of window (EOW), end of induction (EOI), and 18 months from enrolment. For circulating tumor (ct)DNA analysis, libraries were prepared using KAPA HyperCap DS NHL panel (Roche Diagnostics) and sequenced on NovaSeq X Plus (Illumina). Reporter variants(RV) were identified from baseline ctDNA and genomic data was analysed using modified workflow from ctDNAtools (Alkodsi et al., BioRxiv 2020). Logistic regression analysis was performed to evaluate the relationship of baseline total metabolic tumor volume (TMTV) and EOW reduction in SUVmax (ΔSUV) with EOW ctDNA negativity. The primary endpoint is 18-month event free survival, with secondary endpoints ORR, CR, OS, immune-related toxicity (irAE), rates of discontinuation and need for radiotherapy; considered an event. Here, we present results from a planned interim efficacy analysis. Results: At 15th May 2025, 35 pts initiated treatment and are efficacy and safety evaluable. 30 (86%) completed window therapy, with 3 proceeding to induction after cycle 1 due to persistent symptomatic disease. 26 completed induction, and 10 consolidation. Median age was 34 years (range 19-71); 63% were female. 4/35 have withdrawn consent; 2 during window (rash, alopecia/nausea), 1 declining chemotherapy after first induction cycle, and 1 following discontinuation of window due to grade(G)4 pneumonitis. No pts required early transition to induction due to disease progression (PD). 29 pts were evaluable for EOI response, having either completed EOI or reached this timepoint but discontinued therapy prior. EOW ORR was 26/29 (90%), CR 1/29 (3%) and EOI ORR was 25/29 (86%), CR 18/29 (62%), with a single pt having PD at EOI. Excluding the early withdrawals (n=3, above), the EOW ORR was 26/26 (100%), and EOI ORR was 25/26 (96%), CR 18/26 (69%). None of the 7 pts with EOI PR, all Deauville(D) 4, have experienced PD or required radiotherapy. 25/28 pts with baseline ctDNA analysis had suitable RV for monitoring. 11/25 (44%) were EOW ctDNA-. 19 pts had complete imaging for central review (baseline, EOW, EOI) and EOW ctDNA samples. At baseline, median TMTV was 447.7ml (range 68.0-1366.9), and median SUVmax was 22.7 (range 16.0-32.9). The median ΔSUV was 14.9 (range -2.4-28.8). Neither baseline TMTV (p=0.236) or ΔSUV (p=0.259) were associated with ctDNA negativity. The pt with PD had EOW PR (D5) and EOW ctDNA+ (VAF 20.2%) with subsequent EOI PD. They received bridging radiotherapy followed by CAR-T cell therapy, with ongoing MRD- CR at day 90. All other pts had EOW ctDNA VAF 10% (range 0-8.96%). There were no fatal AEs. There were 21 potential irAE; 5 leading to pembro discontinuation-1 during window (n=1 G4 pneumonitis) and 4 during consolidation (n=2 G3 asymptomatic lipase elevation, n=1 G3 cystitis, n=1 G2 esophagitis), with all events resolved or resolving. Other G3 irAE not leading to pembro discontinuation were ALT increase (n=5) and rash (n=4); all during window with potential causative concomitant medications. Other irAE were low grade; rash (G2 n=6), hyperthyroidism (G2 n=1), hypothyroidism (G2 n=1), ALT increase (n=2). Conclusion: This interim analysis of ALLG-PACIFIC suggests that pembro with R-CHOP is effective first-line therapy for PMBL. Almost half of pts achieved ctDNA negativity after 2 cycles of R-pembro prior to initiation of chemotherapy. Outcomes appear comparable to intensive chemotherapy, with low rates of PD at completion of induction therapy. Pembro toxicity may be mitigated in future with treatment de-escalation given the high early ctDNA- rates observed. Enrolment is completed and efficacy, safety and correlative analyses are ongoing.