AI-Driven Benefit-Risk Analysis for Atopic Dermatitis

The class-wide boxed warning for JAK inhibitors (MACE, VTE, malignancy, serious infections, mortality) was primarily driven by the ORAL Surveillance study of tofacitinib in RA patients aged 50+ with at least one cardiovascular risk factor. Our platform models how these risks translate to the AD population, which is typically younger with lower baseline cardiovascular risk. We perform cross-indication safety extrapolation using age-adjusted, risk factor-adjusted analyses that quantify the absolute risk magnitude in AD-specific populations. This enables sponsors to present regulators and payers with contextualized risk data showing that the AD benefit-risk profile is materially different from the RA-derived warning, while still acknowledging the class-wide signal.

Yes. Our platform provides BRA models stratified across four age groups: infants and toddlers (6 months to 5 years, dupilumab-eligible), school-age children (6–11 years), adolescents (12–17 years, JAK inhibitor-eligible), and adults (18+). Each age stratum has distinct safety considerations including growth velocity monitoring, vaccination timing during immunosuppression, the theoretical long-term malignancy risk of JAK inhibition initiated in adolescence, and developmental pharmacokinetic differences. The platform models lifetime cumulative exposure projections for children initiating biologic or JAK therapy in early childhood, supporting treatment duration and switching strategy decisions.

ArcaScience models the full topical-to-systemic treatment continuum, including topical corticosteroids (with HPA axis suppression and skin atrophy risk), topical calcineurin inhibitors (with the longstanding lymphoma boxed warning), topical JAK inhibitors (ruxolitinib cream), topical PDE4 inhibitors (crisaborole), and the escalation to systemic therapy with dupilumab, tralokinumab, JAK inhibitors, or traditional immunosuppressants (cyclosporine, methotrexate). The platform models the benefit-risk threshold for systemic therapy initiation, accounting for topical therapy failure patterns, body surface area involvement, and quality-of-life impact metrics including DLQI, POEM, and SCORAD scores.

Yes. Our platform provides comparative BRA across dupilumab (anti-IL-4Rα, blocking both IL-4 and IL-13), tralokinumab (anti-IL-13), and lebrikizumab (anti-IL-13) using indirect treatment comparison and network meta-analysis methodologies. Key comparative safety endpoints include conjunctivitis rates (typically higher with dupilumab than IL-13-selective agents), injection site reactions, eosinophilia patterns, and infection rates. The platform also compares efficacy endpoints (EASI-75, IGA 0/1, NRS itch) to provide comprehensive benefit-risk positioning, supporting formulary committee presentations and treatment guideline development.

Our platform integrates data from AD registries including TREATgermany, A-STAR, TREAT NL, and the RAPID registry alongside long-term extension studies for dupilumab (up to 5 years), baricitinib, upadacitinib, and abrocitinib. This real-world evidence captures treatment persistence, switching patterns, long-term safety signals including the critical JAK inhibitor MACE and malignancy outcomes beyond 2-year clinical trial exposure, and effectiveness in patients with comorbid asthma, rhinosinusitis, and food allergy commonly seen in the atopic march. The platform reconciles clinical trial and registry data using Bayesian meta-analytic frameworks calibrated for AD-specific confounders and the typically higher disease severity seen in registry populations.