Using Informatics Tools to Characterize Precision Medicine Treatments for Diffuse Large B-Cell Lymphoma (DLBCL)

Diffuse large B-cell lymphoma (DLBCL) is genetically and clinically heterogeneous. Despite immunohistochemical and genomic subtyping, standard firstline DLBCL treatment remains rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP). As genetic profiling becomes more accessible and pervasive, studies identifying specific drug-gene interactions in DLBCL are needed to develop future patient-specific DLBCL treatment strategies that improve the delivery of effective and cost-effective, patient-centered care. Methods: We performed a systematic literature review using PubMed Medical Subject Heading and text word terms for DLBCL/genetics and exome sequencing and recorded DLBCL mutations, associated mutation frequencies, pathways, and implications for overall survival (OS) with a focus on recent publications from Reddy et al 2017, Chapuy et al 2018, and Schmitz et al 2018 and related articles. Data for mutation frequencies and OS were extracted from Reddy et al (2017; https://dlbcl.davelab.org/). Mutations were categorized as either favorable or unfavorable. Favorable mutations were associated with 3-year OS ≥ OS for patients equivalent or better than IPI score ≤1 in this dataset (≥90.1%), whereas unfavorable mutations were associated with 3-year OS ≤ OS for patients with an IPI score ≥2 from this dataset (≤64.0%). Potential mutation-specific treatments were identified from the initial literature review, using the Drug-Gene Interaction database (DGIdb, http://www.dgidb.org/, Cotto et al, 2017) and the Open Targets Platform (OTP, https://www.targetvalidation.org/, Koscielny et al, 2017). DLBCL therapeutic groups were defined for mutations/mutation groups that affected ≥5% of DLBCL patients and were associated with targeted therapies based on DGIdb and OTP. Specific drug-gene interactions were characterized using a series of PubMed searches. Results: The analysis identified 32 articles with 55 total therapeutic clusters. 3 mutations were associated with favorable outcomes with an expected 3-year OS of 90.1% (95% CI: 81.1%-95.0%], and 7 mutations were associated with unfavorable outcomes with an expected 3-year OS of 64.0% (95% CI: 56.3%-70.7%) (Figure). The intersection of mutations associated with unfavorable outcomes and mutations associated with targeted therapies yielded 16 therapeutic clusters and revealed a predicted population of 17.6% of DLBCL with poor risk disease and potentially targetable mutations (Table). Drug-gene interaction characterizations identified 267 pertinent studies that included: in vitro studies, mouse models, or clinical trials of the interactions' roles in any cancer. Of these, 7 studies (1 clinical trial, 6 in vitro studies, and 1 mouse model) directly investigated DLBCL specific effects for fostamatinib, ibrutinib, and crizotinib. Conclusions: Linking systematic review with informatics tools identified DLBCL populations, DLBCL specific gene-drug interactions, and potential therapies for trials for patients who relapse and frontline combinations with R-CHOP that can be used in precision medicine strategies for improving outcomes in this disease.

View largeDownload slide

View largeDownload slide

Close modal