Carlson, Katherine

Background: Pulmonary hypertension (PH) is a life-threatening disease affecting up to 1% of the global population. Diagnosis can be challenging and is often delayed due to the need for advanced imaging and invasive procedures. The use of artificial intelligence applied to ECGs (ECG-AI) has been shown to detect subtle patterns in voltage-time data and may be a valuable tool for the early detection of PH. Hypothesis and Purpose: To evaluate the performance of a previously trained, ECG-AI algorithm to detect PH (ECG-AI PH) using real-world data (RWD) collected in a multicenter, validation study. A joint primary hypothesis required sensitivity (Sn), specificity (Sp), positive predictive value (PPV), and negative predictive value (NPV) to exceed the null values of 76%, 70%, 10% and 90%. Study Design and Methods: This retrospective validation study was conducted at 5 geographically diverse U.S. health systems. Adult subjects were eligible for inclusion if they had a 12-lead ECG paired with an echocardiogram (Echo) in which tricuspid regurgitation velocity (TRV) was recorded, following presentation with dyspnea. Patients were classified according to echocardiographic criteria as either PH (PH+, TRV >3.4 m/s) or controls (PH-, TRV ≤2.8 m/s) to simulate the real-world use of ECG-AI, where a positive result could lead to a follow-up Echo. The study database was locked before processing the digital ECGs with ECG-AI PH. Performance was also estimated in a subset of subjects that later had a right heart catheterization using mPAP ≥ 20mmHg as the definition of PH. Results: A total of 14281 subjects (53% female, 63% aged 50+ years) met the inclusion criteria, including 3019 PH+ cases and 11262 PH- controls (Figure). Sn and Sp were 84.0% (95% CI: 82.6%, 85.3%) and 71.7% (95% CI: 70.9%, 72.6%), respectively. The positive and negative predictive values were 18.9% (95% CI: 18.4%, 19.4%) and 98.3% (95% CI: 98.1%, 98.4%), respectively, at 7.3% prevalence. Each endpoint met pre-defined performance criteria. In the subset of 1683 subjects with a follow up RHC, performance remained robust (Sn 85% (1155/1358); Sp 46% (151/325)). Conclusion: While ECG-AI PH was first developed as an investigational tool to detect PH, continued development as software as a medical device for clinical use demonstrated that the algorithm retained strong performance to detect PH in diverse, non-overlapping clinical settings and patient populations.

Introduction: Pulmonary hypertension (PH) is a life-threatening, progressive disease with non-specific symptoms, often leading to delayed diagnosis. Early identification of World Health Organization Group 1 (Pulmonary Arterial Hypertension, PAH) and Group 4 (Chronic Thromboembolic Pulmonary Hypertension, CTEPH) is essential, as effective therapies can improve outcomes. Hypothesis: An electrocardiogram-based AI algorithm for PH detection (ECG-AI PH) may enable earlier diagnosis and reduce healthcare utilization, including hospitalizations and procedures. Methods: Retrospective analysis used a de-identified data platform of >7M clinical records from a multistate integrated health system. Adult precapillary PH patients (mPAP >20 mmHg, PVR >2 WU, PCWP ≤15 mmHg) were identified as PAH or CTEPH based on ICD codes, use of approved therapies, or surgical interventions (for CTEPH) between 2002 and 2024. ECG-AI PH was applied to ECGs within 30 days of diagnostic right heart catheterization, using a 5:1 randomly sampled PH-negative control cohort. Training set patients were excluded. Clinical event frequency was compared between two intervals: from first possible PH symptom (dyspnea, syncope, chest pain, fatigue, lower limb swelling) to diagnosis, and from symptom onset to first positive ECG-AI PH prediction. Results: A total of 1882 PAH and 359 CTEPH patients met inclusion criteria. Of these, 1340 PAH and 258 CTEPH patients had symptom codes prior to diagnosis. Both groups showed prolonged intervals from symptom onset to diagnosis, with multiple diagnostic procedures and hospitalizations (Figure). ECG-AI PH performance evaluation on the test set included 647 PAH and 152 CTEPH patients. ECG-AI PH achieved AUCs of 0.90 and 0.89 for PAH and CTEPH, sensitivities of 80.3% and 76.8%, and specificities of 83.4% and 82.4%. Among those tested, 576 PAH and 95 CTEPH patients had a positive ECG-AI PH prediction after symptom onset but before diagnosis. Compared to the current patient journey, the interval between initial symptoms and a positive output from ECG-AI PH was shorter and had fewer diagnostic tests/visits. Conclusion: ECG-AI PH demonstrated strong performance in detecting PAH and CTEPH. It may reduce diagnostic delays, support earlier PH-focused screening (e.g., echocardiograms evaluating the right heart), enable earlier intervention, and reduce pre-diagnosis healthcare burden, benefitting both patient outcomes and healthcare system efficiency.