Is Molecular Characterization Useful for Targeted Therapy Orientation in Patients with Relapsed or Refractory DLBCL Included in Early Phase Clinical Trials?

Background: Early drug development is associated with a high rate of failure in patients with relapsed or refractory diffuse large B-cell lymphoma (R/R DLBLC), partly due to genomic heterogeneity. Precision medicine aims at identifying patients with cancer who are likely to respond to targeted therapies. Here we assessed the clinical impact of characterizing actionable and molecular targets and matching targeted therapy in pts with R/R DLBCL prior enrollment in early phase clinical trials (epCT).

Methods: Pts with R/R DLBCL enrolled in epCT, provided paired samples including germline and fresh tumor tissue sample. Paired samples were analyzed by targeted next generation sequencing (NGS) using a customized panel of 44 genes (Ion Torrent). Clinical actionability of genomic alterations and orientation to targeted therapy were prospectively assessed at the Hematological Tumor Board from our Institution according to the ESMO Scale for Clinical Actionability of molecular Targets (ESCAT) tiers (ESCAT scale; J. Mateo, Annals of Oncology, 2018[1]). Pts were classified as molecularly matched (MM) i.e. treated based on the presence of an ESCAT actionable molecular abnormality, or non-molecularly matched (no-MM). Tumor and clinical characteristics and outcome of pts were collected.

Results: From 2013 to 2020, 90 pts with R/R DLBCL screened for early clinical trial had tumor & germline samples sequenced by NGS. The mean age was 62 years old (range, 23-83) and median prior lines of systemic therapy was 2 (range 1-9). Sixteen (18%) pts were enrolled in epCT and treated with MO drugs, 43 pts (48%) were enrolled in epCT and treated with no-MM drugs, and 31 (34%) pts were not enrolled in epCT. The main clinical characteristics of the pts were similar in both MM and no-MM groups (table 1). Based on ESCAT scale, targeted therapies of the 16 pts MM were distributed as “investigational-level” tier II in 6 pts (3 pts EZH2 mutant [m] treated with EZH2 inhibitors; 3 patients with a CD79Bm treated with BTK inhibitors); “hypothetical-level” tier III in 1 pt BRAFm treated with a BRAF inhibitor; and “hypothetical-level” tier IV in 9 pts (6 pts MYD88m treated with cereblon modulator drugs; 2 pts with BCL7Am/ARID1Am treated with investigational epigenetic modifier drugs; 1 pt with MYCm treated with a PI3K inhibitor). Median progression-free survival in MM and no-MM pts, were 3.3 and 1.7 months, respectively (p=0.0092; HR=0.47 [CI95:0.27-0.83]). Median overall survival in MM and no-MM pts were 20.0 and 6.7 months, respectively (p=0.0238; HR=0.49 [CI95: 0.26-0.91]) (figure 1).

Conclusions: Molecular orientation before inclusion in early phase clinical trials was associated with better outcomes in patients with R/R DLBCL. A molecular tumor profile could be useful to manage patients with R/R DLBCL with precision medicine helping to identify patients likely to respond to targeted therapies.

Reference.

1- Mateo J, Chakravarty D, Dienstmann R et al. A framework to rank genomic alterations as targets for cancer precision medicine: the ESMO Scale for Clinical Actionability of molecular Targets (ESCAT). Annals of Oncology 2018; 29(9):1895–1902.

Table 1.Patient’s characteristics at time of R/R DLBCL relapse and tumor sample in molecularly matched patients (MM group) and non-molecularly matched patients (no-MM group). Data presented are n (%) unless otherwise stated.

MM= Molecularly matched. no-MM= Non-molecularly matched. R/R DLBCL= relapsed or refractory diffuse large B-cell lymphoma. y.o. = years old. mo.= months. ECOG PS= Eastern Cooperative Oncology Group performance status. NHL=non-Hodgkin lymphoma. ASCT= Autostem cell transplantation. IHC= Immunohistochemistry. COO= Cell of origin. GC= Germinal center. ABC= Activated B-cell. NA= Not available. LDH= Lactate deshydrogenase.

*the p-value indicates the comparison between the proportions of the groups from a contingency table, and was calculated by the exact Fisher test. All tests were non-parametric. The p values were calculated by two-tailed and a 95% confidence interval was used, with values less than 0.005 considered being significantly different.