Despite the addition of rituximab to CHOP chemotherapy in the treatment of diffuse large B-cell lymphoma (DLBCL), up to 40% of human patients ultimately succumb to their disease. Additionally, those patients effectively cured have a high risk of long-term morbidities secondary to treatment. As such, novel approaches that reduce the intensity of chemotherapy while simultaneously improving outcomes through combination with small molecule inhibitors/immunotherapeutics are desirable. Unfortunately, exploration of such treatment strategies is typically relegated to patients with relapsed/refractory disease where efficacy is likely to be lower. It is now clear that DLBCL in dogs resembles the human disease with respect to clinical presentation, molecular aberrations, and genomic alterations making it a good spontaneous large animal model of disease. As such, we have been leveraging canine DLBCL to rapidly evaluate the activity of rational small molecule inhibitors targeting PI3Kδ (RV1001), NAMPT/PAK4 (KPT-9274), or SUMO-Activating Enzyme (TAK-981) combined with immunotherapy (anti-CD20, from Elanco), with the ultimate goal of identifying the most effective combination to move forward in human patients. Using an adaptive mini-pilot trial approach in the front-line setting we are assessing the activity and toxicities of anti-CD20/targeted small molecule combinations in affected dogs and correlating clinical, biological, and genomic biomarkers to develop signatures that predict response to therapy and long-term progression-free survival. Our data demonstrate that anti-CD20 effectively depletes B-cell populations in treated dogs and that treatment with RV1001 can break tolerance as evidenced by prednisone responsive transaminitis. We have identified long-term survivors in each treatment cohort, and our preliminary data suggest that a combination of low-dose doxorubicin, anti-CD20, KPT-9274, and RV1001 may be associated with the long-term survival. We have performed integrated genomic analysis of paired tumor/normal and matched constitutional samples from the five dogs in the first cohort to identify candidate molecular correlates of treatment response. Matched tumor/normal whole genome sequencing confirmed previously described (TRAF3, SETD2, TP53) as well as novel candidate (MGA) driver mutations associated with canine DLBCL. Following completion of our pilot studies and correlative analyses, a large prospective study will be undertaken to confirm findings.