Background: Patients with relapsed or refractory diffuse large B cell lymphoma (r/r DLBCL) often do poorly with conventional salvage regimens of chemotherapy. Next-generation sequencing (NGS) panels can rapidly identify recurrent molecular abnormalities, thus helping to orient patients (pts) in appropriate targeted therapies or clinical trials. We aimed to evaluate whether selecting pts through tumor genotyping is associated with a better outcome.

Methods: From 2013 to 2018, all pts with r/r DLBCL having molecular portrait of tumor before enrollment in early clinical trials (eaCTs) for r/r DLBCL were analyzed for clinical and histomolecular characteristics, tumor response, progression-free survival (PFS) duration of response and overall survival. The objective were to evaluate the feasibility and potential benefit of using tumor genotyping for guiding enrollment in eaCTs. Molecular screening methods included immunohistochemistry, PCR-based assays and next-generation sequencing. All pts gave their written consent for the study.

Results: Sixty-two pts with r/r DLBCL had tumor molecular portraits. At the time of tumor molecular portrait, the median age was 69 years (range 26-77), median of previous line of therapies was 2 (range 1-9) and 14 pts (23%) had prior received auto stem-cell transplant. Fifteen out of the 62 pts (24%) were molecularly oriented (MO) in eaCTs. Identification of potentially actionable targets was found in 30/62 (48%) of pts, of whom 15/30 (50%) received a molecularly-informed therapy. Beyond molecular portrait, fifty pts were enrolled in eaCTs (15 pts were MO and 35 were non-MO oriented) and 12 pts were not enrolled in eaCTs. The MO-oriented group included the following therapeutic targets: CD79A/B or MYD88 (n=10 pts), EZH2 or ARIDI1A (n=3 pts), MYC (n=1 pt) and BRAF (n=1 pt). The overall responses rate was 60% (6 PR and 3 CR) in MO group versus 23% (5 PR and 2 CR) in non-MO group (p= 0.01). The median of PFS in MO and non-MO groups were 2.2 months and 1.9 months, respectively (p=0.23; HR=0.69 [CI95:0.38-1.26]). The median duration of response in MO and non-MO groups were 10.9 and 9.3 months, respectively (p=0.78; HR=0.76 [CI95: 0.26-2.18]. The mean PF2/PF1 ratio in MO and non-MO groups were 2.34 [CI95: 0.27-4.41] and 1.67 [CI:0.53-2.81], respectively (p=0.093; HR=0.61 [CI95: 0.33-1.14]. The median overall survival in MO and non-MO groups were 8.9 and 7.7 months, respectively (p=0.34; HR=0.69 [CI95: 0.33-1.47]).

Conclusions: Molecularly oriented treatments of recurrent diffuse large B-cell lymphoma demonstrates higher responses rates. A subset of patients with recurrent or refractory diffuse large B-cell lymphoma may benefit from incorporation of tumor genotyping to guide their enrollment in clinical trials. Accelerating the use of prospective genomics tumor molecular portraits may increase the chances for a precision medicine for recurrent diffuse large B-cell lymphoma.


Soria:Medimmune: Employment. Ribrag:BMS: Consultancy, Honoraria, Other: travel; Infinity: Consultancy, Honoraria; MSD: Honoraria; NanoString Technologies: Consultancy, Honoraria; argenX: Research Funding; epizyme: Consultancy, Honoraria; Incyte Corporation: Consultancy; pharmamar: Other: travel; Servier: Consultancy, Honoraria; Gilead: Consultancy, Honoraria; Roche: Honoraria, Other: travel; Amgen: Research Funding.

Author notes


Asterisk with author names denotes non-ASH members.