Pembrolizumab with R-CHOP in Previously Untreated Diffuse Large B-Cell Lymphoma: Long Term Follow up and Analysis of the Mechanism of Pdl-1 Tumor Expression

Background: Rituximab with cyclophosphamide, doxorubicin, vincristine, and prednisone (RCHOP) remains standard therapy for previously untreated diffuse large B-cell lymphoma (DLBCL). We previously reported that adding pembrolizumab to RCHOP (PR-CHOP), shows no significant additive toxicity and a promising 2-year progression free survival of 83% (Smith B J Haem 2019), with no progression events among 19 pts with known PDL-1+ tumors. In that analysis, we noted 83% of patients had tumor PDL-1 expression by IHC using a centralized standard assay, including 18/23 (78%) and 13/23 (56%) demonstrating positivity in ≥5% and ≥30% of lymphoma cells respectively. Herein we report long-term clinical follow up, as well as results of high-resolution molecular karyotyping using chromosome genomic array testing (CGAT), to assess whether 9p24.1 abnormalities explained the extent of PDL1 tumor expression.

Methods: Pts age 18 years or older with previously untreated DLBCL, transformed lymphoma and grade 3 B follicular lymphoma eligible for 6-cycle, curative-intent RCHOP were eligible. Steroids within 7 days of treatment, active autoimmune disease, or history of pneumonitis were excluded. This phase I study combined pembrolizumab (200 mg IV q 3 weeks x 6) to RCHOP x 6, with supportive care per standard practice. There was no maintenance or consolidation therapy. Baseline PDL1 expression was analyzed centrally (Merck/Qualtek). CGAT was performed from DNA extracted from formalin-fixed, paraffin-embedded tissue sections using the OncoScan platform (ThermoFisher). H&E slides of adjacent sections of the tissue used for CGAT were evaluated by a pathologist to ensure a minimum of 30% tumor content. Patients were followed for relapse and survival.

Results: Among 30 treated pts with a median follow up of 32 months, there have been no additional DLBCL relapses or deaths since the published report. There has been one relapse of follicular lymphoma in a patient with initial composite lymphoma , who is now in remission 1 year after autologous stem cell transplantation. 3-year estimated PFS is 83% and OS is 86%. On univariate analysis, only tumor bulk 7.5 cm or greater (p=.02) and absence tumor PDL-1 expression by IHC (p=.001) predicted inferior PFS. Tumor bulk, (p=.03), IPI 3 or higher (p=.01), and absence of tumor PDL-1 expression (p=.001) predicted worse overall survival. PFS and OS were unaffected by cell of origin by IHC, double expresser status, or diagnosis to treatment interval greater than median (29 days).

Regarding safety, there were late cases of paraneoplastic pemphigus and rheumatoid arthritis 5 and 8 months from last dose of pembrolizumab, respectively. Both of these were managed successfully with immunosuppression.

Chromosome genomic array testing CGAT) was performed to assess copy number aberrations (CNAs) and copy neutral loss of heterozygosity (cnLOH), particularly for the presence of gains or amplifications involving locus 9p24.1, containing the CD274 (PDL1) gene. Of 15 tested pretreatment baseline diagnostic tissue specimens, 14 provided informative result with 12 abnormal and 2 normal. Notably, no gain or amplification in 9p24.1 were seen; there was one case of deletion and 1 with cnLOH. The percent genome altered among abnormal cases ranged from 2% to 37% (median 12%). Complex karyotype, defined as genomic aberrations involving at least 3 chromosome arms (or % genome altered greater than 5% by CGAT), was seen in 10/14 pts (71%). This 14-sample data set included 6/14 pts with >30% PDL1 tumor expression (and 10 with 5% or greater PDL1-tumor expression).

Conclusions: No additional safety signals have been observed, and PFS/OS remain favorable for the combination of pembrolizumab + RCHOP in this single arm trial with the best results in PDL-1+ disease. Despite frequent tumor PDL1-expression at baseline, no gains or implications of 9p24.1 were observed suggesting alternative mechanisms for overexpression of PDL-1. Further study this regimen and others testing combinations of immune checkpoint inhibition with first-line DLBCL therapy are warranted, to better ascertain the potential superiority of this strategy in this setting.