AI Enhanced Pulmonary Hypertension Care: A Roadmap to Precision Medicine, AHJP December 2025 Review Reveals

This state-of-the-art review is published in the American Heart Journal Plus: Cardiology Research and Practice in December 2025.

The Challenge of Delayed PH Diagnosis

All-cause pulmonary hypertension (PH) is a major public health challenge linked to increased mortality. As a complex hemodynamic condition involving the heart, lungs, and pulmonary vasculature, PH—most commonly due to left heart disease (WHO Group 2)—often goes unrecognized because of its varied causes and diagnostic difficulty. Delayed or missed diagnosis allows progression to right heart failure and early death. Early identification and timely disease-modifying treatment are therefore essential to reduce hospitalizations and improve quality of life. The burden of disease is particularly high in individuals living in remote and disadvantaged regions, who also have the least access to the current specialized, tertiary-centric PH model of care.

Reviewing AI’s Role in Clinical Practice

This state-of-the-art review explored how emerging technologies, including artificial intelligence, can enhance the diagnosis and management of pulmonary hypertension (PH). Drawing on literature from 2015–2025 across major databases, the authors focused on clinically relevant and methodologically transparent studies applying AI to PH. Rather than a systematic review, the goal was to illustrate how AI can support multiple steps in the PH care pathway, offering more efficient and effective triage than current manual approaches.

Demonstrable Accuracy in AI-Enhanced Imaging

• Smarter Triage: AI systems can extract clinical information from Electronic Medical Records (EMR), such as diagnostic codes and major health issues, to create a "breathlessness triage" system that identifies individuals requiring further PH investigation.

• Echocardiography Enhancement: Echocardiography is the central investigation for many forms of PH. AI is poised to improve its efficiency, accuracy, and reproducibility. AI can guide non-expert operators, such as community health workers, to obtain diagnostic-quality images, which is critical for democratizing access in rural and inaccessible locations.

• Phenotype Prediction: Proof-of-concept studies using deep neural networks have demonstrated AI's ability to predict all-cause PH (defined as TRV >2.8 m/s) with reasonable accuracy, even when key PH-related measurements were withheld. Moreover, simple algorithms, such as the echocardiographic pulmonary to left atrial ratio (ePLAR), demonstrate good capacity to differentiate pre-capillary (non-left heart) PH from post-capillary (left heart) PH.

Advancing Prognosis and Risk Stratification

Beyond diagnosis, AI models offer significant promise in risk assessment and prognostication:

• Mortality Prediction: AI models trained on paired electrocardiogram (ECG) and echocardiogram data have demonstrated strong prognostic value in predicting mortality in PH patients.

• Personalized Risk: AI can integrate various sources of clinical data (demographics, comorbidities, lab results) to stratify patients into different risk categories. This enables healthcare providers to tailor treatment plans and interventions accordingly.

• Objective Outputs: AI algorithms can analyze multiple parameters simultaneously, and their outputs are free of the subjectivity typical of human interpretation and are not subject to fatigue.

Optimal Care Delivery and Future Strategy

Optimal care for complex PH patients requires a dedicated multidisciplinary team, including specialist physicians, nurses, allied health members, and community care providers. While AI enhances technological capabilities, the sources emphasize that compassion and human interaction are essential attributes which only human clinicians can provide.

• Individualized Treatment: Treatment must be individualized based on the background and expectations of each patient, with shared decision making facilitated by well-informed patients and families. PH clinicians must use their clinical acumen to avoid unnecessary, costly, and invasive tests.

• Administrative Efficiency: AI-based speech recognition can provide automated generation of medical notes and correspondence, which reduces the administrative burden on clinicians. This frees up more clinician time to spend with patients.

• Monitoring and Follow-up: AI-based wearable technologies (such as smartwatches) have potential utility for the efficient monitoring and follow-up of PH patients.

The collective integration of these AI systems holds the potential to improve diagnostic accuracy, efficiency, and speed, leading to more timely and targeted interventions that improve patient outcomes.

Clinical Inference for Cardiologists & Cardiology Teams

AI offers a powerful opportunity to enhance and personalize Pulmonary Hypertension (PH) care by improving early detection and risk stratification. Because PH diagnosis is often delayed and expertise is unevenly distributed, AI can strengthen the care pathway by increasing the efficiency, accuracy, and reproducibility of echocardiography—the key initial test. AI tools can guide non-expert operators to obtain high-quality images and help distinguish PH due to left heart disease from other forms using metrics such as ePLAR. In addition, AI-driven models show strong potential for prognostication, enabling more precise risk stratification and individualized treatment planning. By automating labor-intensive tasks like measurements and report generation, AI frees clinicians to focus on compassionate, patient-centered decision making within multidisciplinary PH teams.

Reference: Naing P, Scalia GM, Murdoch D, Ranasinghe I, Forrester DL, Strange G, Playford D. Artificial intelligence enhanced contemporary pulmonary hypertension care. American Heart Journal Plus: Cardiology Research and Practice. 2025 Nov 12:100673.

For regular cardiology updates from recent journals, kindly follow our WhatsApp group