AstraZeneca has entered a multi-year partnership with BostonGene, leveraging the latter’s AI foundation model for tumor and immune biology to drive early clinical trial efficiency and de-risk its oncology pipeline. Under the agreement, AstraZeneca will apply BostonGene’s multimodal analytics—integrating cell-free RNA sequencing, tumor microenvironment profiling and pre-trained foundation models—to generate patient-level predictions of safety and efficacy outcomes. The collaboration aims to shorten development timelines by up to 20%, reduce cohort attrition in Phase 1/2 studies and inform biomarker-guided trial designs across multiple solid-tumor indications. AstraZeneca’s Chief of AI for Science Innovation, Dr. Jorge Reis-Filho, highlighted that predictive insights from BostonGene’s platform could enable selection of high-responding patient subgroups and anticipate tolerability issues before First-in-Human dosing, potentially saving hundreds of millions in development spend per program.

AstraZeneca reported topline data from the TULIP-SC Phase 3 study of Saphnelo (anifrolumab) as a self-injectable subcutaneous therapy for moderate to severe systemic lupus erythematosus. The trial demonstrated a 56.2% reduction in disease activity at Week 52 using the BICLA endpoint versus 37.1% for placebo (p<0.001), consistent with prior intravenous studies. The safety profile showed comparable rates of serious adverse events (12.5% vs 11.8% for placebo) and low injection-site reaction incidence (5.4%). Already approved in over 70 countries and prescribed to more than 40,000 patients globally, Saphnelo’s new formulation is projected to enhance patient adherence and expand market reach in the self-care segment. AstraZeneca estimates that subcutaneous delivery could increase annual Saphnelo prescriptions by 25%, contributing significantly to its specialty immunology revenue stream.