AI may help assess long-term risk in pulmonary hypertension, study finds
Right atrial size and coronary calcium were linked to mortality
Written by |
Artificial intelligence may help researchers analyze CT scans to assess long-term mortality risk in people with pulmonary hypertension. (Photo from iStock)
Artificial intelligence (AI) analysis of CT scans may help predict the long-term risk of death in people with different forms of pulmonary hypertension (PH), according to a new study.
“Although further work is required, these findings suggest that incorporating AI-derived imaging variables into established clinical risk score assessment may provide complementary information for risk stratification and improve prognostic assessment in PH patients,” the researchers wrote.
The study, “AI-derived cardiothoracic structures and coronary artery calcification volumes on non-contrast CT as predictors of mortality in pulmonary hypertension,” was published in European Radiology Cardiovascular and Thoracic.
AI could add to existing PH risk assessments
PH is abnormally high pressure in the vessels that carry blood from the heart to the lungs. There are several different types of PH with distinct underlying causes, but in all forms, this high pressure strains the heart and limits oxygen delivery throughout the body. This can cause PH symptoms such as shortness of breath.
CT scans use a series of X-rays to create detailed images of structures inside the body and have long helped doctors assess and monitor PH. Measuring heart and blood vessel structures on these scans by hand can be time-consuming, and measurements may vary between clinicians. Researchers are exploring whether AI can automate these measurements, making the process faster and more consistent.
Researchers in the U.K. analyzed data from hundreds of people evaluated for suspected PH at a center in England. They used AI programs to measure features on CT scans taken during the evaluation, then examined whether those measurements were associated with long-term survival.
“This study aimed to evaluate the prognostic value of AI-derived [measures from CT scans] in PH by assessing their association with mortality and prognostic significance after adjustment for demographic factors, PH subgroups, and established clinical risk scores,” the team wrote.
CT measurements linked to mortality risk
The study included 620 people: 202 with pulmonary arterial hypertension (PAH), 84 with PH associated with left heart disease, 112 with PH associated with lung disease, 65 with chronic thromboembolic PH, nine with PH of unclear type, and 148 who did not have PH. The AI measurements showed that right atrial and pulmonary artery volumes were significantly larger in the four main PH groups than in people without PH.
Statistical models showed that several AI-derived imaging features were significantly associated with mortality. In particular, patients with calcifications (calcium deposits) in certain arteries supplying the heart — the right coronary artery, the mainstem and left anterior descending artery, and the left circumflex artery — had a significantly higher risk of death. A larger right atrium — the part of the heart that collects oxygen-depleted blood before it is pumped toward the lungs — was also linked to a higher risk of death in the models.
Among people with PAH, AI-based measures of coronary artery calcification and right atrial volume remained significantly associated with mortality even after the researchers adjusted for REVEAL scores, a standard tool used to assess risk in this form of PH.
“AI-derived volumetric measurements of the [right atrium], and coronary artery calcifications … on routine [CT scans] independently predict mortality in PH and demonstrate an independent and significant prognostic value even after adjustment for established risk scores,” the researchers concluded.
The researchers noted that the study was retrospective and involved patients at a single center, and that they did not assess the impact of using AI in clinical practice. They said further research is needed to validate the findings, but that their data support the potential use of AI-derived CT measurements to provide additional information for assessing mortality risk in people with PH.

Leave a comment
Fill in the required fields to post. Your email address will not be published.