Artificial intelligence–enhanced pulmonary embolism response teams: clinical and economic impact in acute care

Background Pulmonary embolism (PE) remains a major cause of morbidity and mortality, particularly in intermediate-high and high-risk patients requiring rapid escalation of therapy. Pulmonary Embolism Response Teams (PERT) improve multidisciplinary decision-making, yet delayed activation may limit benefit. Artificial intelligence (AI) tools offer opportunities to accelerate detection and workflow prioritization. The combined clinical and economic impact of AI-supported PERT pathways remains insufficiently characterized. Purpose To evaluate whether an AI-enhanced PERT workflow improves activation efficiency, diagnostic and treatment timelines, resource utilization, and probabilistic cost-effectiveness. Methods We retrospectively screened 2,500 patients with CT-confirmed PE between 08/2022 and 07/2024. Outcomes were compared one year before and after implementation of an AI-PERT model, including activation rates, interventional therapy utilization, time to diagnosis and treatment, ICU and hospital length of stay, and mortality. A decision-analytic model estimated 1-year incremental costs and deaths avoided, with uncertainty assessed using 1,000 Monte Carlo simulations. Results Among screened patients, 424 intermediate-high/high-risk cases were analyzed (central PE n=181; segmental n=243). AI-PERT increased activation rates nearly tenfold (9.2% vs 84.5%). Interventional therapy utilization expanded across systemic lysis, catheter-directed thrombolysis, and large-bore thrombectomy. Time from CT completion to report decreased from 2.1±0.6 to 1.3±0.5 hours, and time to treatment was reduced from 35.9±9.6 to 18.9±2.6 hours (p<0.05). Median ICU stay decreased from 9 to 7 days. 30-day mortality improved (25% vs 21%). Probabilistic modeling demonstrated that most simulations favored AI-PERT, clustering in the cost-saving and more effective domain, indicating a high likelihood of improved survival with reduced healthcare expenditure. Conclusions AI-supported PERT integration substantially enhances activation efficiency, accelerates treatment, and reduces resource utilization in acute PE care. Probabilistic analysis suggests these clinical gains are achieved with favorable economic impact. AI-PERT represents a scalable, value-based strategy to optimize PE management.