Radial wall strain for residual risk stratification after percutaneous coronary intervention.
Huang J, Tu S, et al. • EuroIntervention : journal of EuroPCR in collaboration with the Working Group on Interventional Cardiology of the European Society of Cardiology • 2026
Radial wall strain derived from routine angiography via a novel AI algorithm independently predicted long-term adverse cardiovascular events in non-target vessels after PCI, with an adjusted hazard ratio of 4.82 and adjusted AUC of 0.73 over 5 years.
Key Findings
Methods
A novel AI algorithm enabling real-time automated assessment of radial wall strain (RWS) from routine angiography was successfully evaluated in 1,384 non-target vessels from 802 patients.
The study was a blinded post hoc analysis of the TARGET All Comers randomised trial, which enrolled 1,551 patients total.
Angiographic image quality was independently assessed prior to blinded computation of RWS.
The algorithm achieves subpixel-level lumen delineation without requiring complex modelling or expensive intracoronary imaging.
A prespecified cutoff value of maximal RWS (RWSmax) ≥13% was used to define increased plaque vulnerability.
Results
Baseline RWSmax independently predicted the non-target vessel-oriented composite endpoint (NT-VOCE) over 5 years with an adjusted hazard ratio of 4.82.
Adjusted HR of 4.82 (95% CI: 3.14–7.40; p<0.0001).
Adjusted area under the curve (AUC) of 0.73 (95% CI: 0.69–0.78; p<0.0001).
The NT-VOCE was a composite of cardiac death, non-target vessel myocardial infarction, and non-target vessel revascularisation over 5 years.
Outcomes were prospectively adjudicated in the parent trial before being unblinded and linked to vessel-level outcomes.
Results
RWS demonstrated the highest predictive accuracy for non-target vessel revascularisation among all components of the composite endpoint.
Adjusted AUC for non-target vessel revascularisation was 0.92 (95% CI: 0.88–0.95; p<0.0001).
This was substantially higher than the overall NT-VOCE AUC of 0.73.
Non-target vessel revascularisation was one of three components of the NT-VOCE, along with cardiac death and non-target vessel myocardial infarction.
Results
In the subgroup with high-quality angiographic images, RWS demonstrated even greater predictive performance for the NT-VOCE.
Adjusted HR of 6.89 (95% CI: 3.15–15.07; p<0.0001) in the high-quality angiographic image subgroup.
This compares to an adjusted HR of 4.82 in the overall cohort.
Image quality was independently assessed prior to analysis, suggesting image quality is a meaningful moderator of predictive performance.
Results
Elevated baseline RWSmax was strongly and independently associated with long-term adverse events in deferred non-target vessels over a 5-year follow-up period.
The analysis was conducted in non-target vessels, meaning vessels not treated during the index PCI procedure.
Untreated non-target vessels with vulnerable plaques are described as 'a major contributor to future adverse cardiovascular events.'
The association remained significant after adjustment for other variables, confirming independent prognostic value.
Ongoing and planned prospective, randomised trials are evaluating the role of RWS-guided risk stratification.
What This Means
This research suggests that a new artificial intelligence tool can measure how much coronary artery walls stretch and compress with each heartbeat—a property called radial wall strain (RWS)—using standard X-ray images taken during heart procedures (angiography). High wall strain is thought to indicate unstable or 'vulnerable' plaques, which are fatty deposits in artery walls that are prone to rupturing and causing heart attacks. Previously, identifying these plaques required expensive and complex imaging tools inserted into arteries; this new approach works from images already routinely collected.
Using data from over 800 patients followed for 5 years after a heart stenting procedure, the researchers found that arteries with high RWS (≥13%) at the time of the procedure were dramatically more likely to later cause problems—even though those arteries were not treated during the original procedure. The risk of future adverse events (such as heart attack, cardiac death, or need for further procedures in those untreated arteries) was nearly 5 times higher in vessels with elevated RWS, and the tool was especially good at predicting which vessels would eventually need additional treatment, with a predictive accuracy (AUC) of 0.92 out of a possible 1.0. In patients whose X-ray images were of the highest quality, the tool's predictive power was even stronger.
This research suggests that RWS analysis from routine angiography could help doctors identify which untreated coronary arteries are most likely to cause future heart events, potentially allowing more targeted preventive treatment. Because the method works from standard images already collected during normal procedures, it would not require additional invasive tools or extra cost. The authors note that prospective clinical trials are underway to test whether using this information to guide treatment decisions actually improves patient outcomes.
Huang J, Tu S, Huang H, Fezzi S, Zhong J, Baumbach A, et al.. (2026). Radial wall strain for residual risk stratification after percutaneous coronary intervention.. EuroIntervention : journal of EuroPCR in collaboration with the Working Group on Interventional Cardiology of the European Society of Cardiology. https://doi.org/10.4244/EIJ-D-26-00126