What is the clinical evidence from this study?

Study design: Observational. Population: Pulmonary arterial hypertension (n=181). Intervention: Machine learning models. Primary outcome: Prediction of mean pulmonary arterial pressure (mPAP) and pulmonary vascular resistance (PVR) (r = 0.80 for mPAP; r = 0.71 for PVR).

What question did this study set out to answer?

To develop and evaluate machine learning models for predicting key haemodynamic parameters in pulmonary arterial hypertension using non-invasive data.

June 19, 2026Open Access

Machine learning for prediction of key haemodynamic parameters in pulmonary arterial hypertension

Key Result

Machine learning models can estimate mean pulmonary arterial pressure (r=0.80) and pulmonary vascular resistance (r=0.71) from routine clinical data obtained prior to right heart catheterization.

Key Points

To develop and evaluate machine learning models for predicting key haemodynamic parameters in pulmonary arterial hypertension using non-invasive data.
Analyzed data from 181 patients with confirmed pulmonary arterial hypertension using 56 predictive variables.
Employed an 80/20 train-test split and fivefold cross-validation across various machine learning models.
Models evaluated include lasso regression, ridge regression, k-nearest neighbours, decision trees, random forest, and gradient boosting.
Lasso regression achieved the best performance for predicting mean pulmonary arterial pressure (mPAP) with r = 0.80 and R² = 0.64.
Ridge regression performed best for predicting pulmonary vascular resistance (PVR) with r = 0.71 and R² = 0.51.
Random forest and gradient boosting showed modest performance for cardiac index prediction with r = 0.38 and 0.37 respectively.

Study Design

Type

Observational (n=181)

Structured PICO

Can machine learning models accurately predict key haemodynamic parameters in patients with pulmonary arterial hypertension using routinely available non-invasive data?

Population

181 patients with invasively confirmed pulmonary arterial hypertension.

Exposure

Machine learning models (lasso regression, ridge regression, k-nearest neighbours, decision trees, random forest, and gradient boosting machine) using 56 non-invasive variables collected within 8 weeks prior to right heart catheterization

Outcome

Prediction of mean pulmonary arterial pressure (mPAP) and pulmonary vascular resistance (PVR)surrogate

Machine learning models can estimate mean pulmonary arterial pressure and pulmonary vascular resistance from routine non-invasive clinical data in patients with confirmed PAH.

Main Result

Effect estimate: r = 0.80 for mPAP; r = 0.71 for PVR

Limitations

External validation is required to confirm generalizability and clinical applicability.

Abstract

Abstract Aims Machine learning (ML) is increasingly recognized for its ability to identify and structure variables for predictive tasks. Pulmonary arterial hypertension (PAH) is a progressive disease characterized by elevated mean pulmonary arterial pressure (mPAP) and pulmonary vascular resistance (PVR) with normal pulmonary arterial wedge pressure (PAWP), as assessed by right heart catheterization (RHC). Despite increased awareness, delays between onset of non-specific symptoms and diagnosis continue to hinder early initiation of targeted therapies, leading to poorer outcomes. To develop and evaluate ML models for predicting key haemodynamic parameters in PAH, based on routinely available non-invasive data collected within 8 weeks prior to RHC, as a proof of concept. Methods and results We analysed data from 181 patients with invasively confirmed PAH, incorporating 56 variables, including demographics, echocardiography, blood gas analyses, 6-min walk distances, laboratory tests, and WHO functional class. An 80/20 train-test split and fivefold cross-validation were applied across multiple ML models, including least absolute shrinkage and selection operator (lasso) regression, ridge regression, k-nearest neighbours, decision trees, random forest, and gradient boosting machine. Lasso achieved best performance for predicting mPAP (r = 0.80, R² = 0.64, RMSE = 8.49). For PVR, ridge performed best (r = 0.71, R² = 0.51, RMSE = 3.60). Random forest and gradient boosting machines achieved modest but consistent performance for cardiac index (r = 0.38 and 0.37), while PAWP prediction remained limited across all models. Conclusion Machine learning models can estimate mPAP and PVR from routine clinical data obtained prior to RHC in patients with confirmed PAH. External validation is required to confirm generalizability and clinical applicability.