Rayong study: learning from 218 participants

01

Evaluation in a care setting

This retrospective study helps us understand a workflow combining healthcare professionals, smartphones, and cardiac assessment. Summary findings were presented at the 17th Cardiac Network Forum (CNF) in 2026.

Full publication and peer-review status remain unverified. The findings here are a summary of the Rayong study in its specific setting.

02

Participants and assessment indications

The 218 participants included 76 undergoing pre-chemotherapy assessment and 142 undergoing general LV function assessment. These indications provide two distinct clinical-service populations.

03

A workflow involving nurses and clinicians

Registered nurses recorded ultrasound displays with smartphone cameras. EasyEF classified LV function as poor, fair, or good, with comparison against assessments confirmed by a cardiologist.

Extending the findings to other services requires considering image quality, device and model versions, and user experience together.

04

Results by ventricular-function category

Overall accuracy was 93.6%, with 100% for poor LV, 96.9% for good LV, and 27.3% for fair LV.

Fair LV is the main limitation requiring further work. The 100% poor LV result applies to this sample; errors remain possible in new populations.

05

Performance differs between participant groups

Overall accuracy was 98.7% in pre-chemotherapy assessment and 90.8% in general assessment. Separating these results reveals differences hidden by a single overall figure.

Category mix, image quality, and examination conditions need further study to explain the difference. Standard assessment before treatment remains essential.

06

From findings to the next research questions

The Rayong findings sharpen our focus on fair LV and multisite testing, including how AI information might help care teams prioritize further assessment.

Effects on referral reduction need an appropriate study design. Statistical significance thresholds also need reconciliation between the two abstracts before reuse.