The 2024 study: screening evaluated on 184 video clips ↗
This study provides the published foundation for using smartphone-recorded ultrasound video to classify left ventricular function with AI. Its 177 correct results among 184 test clips, or 96.2%, must be read alongside data-selection criteria and the Mildly Reduced EF limitation.
Study summary • peer review not verifiedThe Rayong study: 218 participants in clinical practice ↗
Rayong Hospital data from December 2025–January 2026 adds a view of EasyEF in pre-chemotherapy and general LV function assessment. The reported 93.6% overall result comes with a clear limitation: fair LV accuracy was 27.3%.
Evidence-reading guideHow to interpret EasyEF evidence ↗
A single percentage does not describe an entire system. This guide uses EasyEF sources to explain clips versus participants, training versus testing, overall versus subgroup results, and demonstrations versus clinical validation.
12-month project planRoadmap: building evidence for broader use ↗
The EasyEF brief sets a 12-month goal of moving from TRL 7 to TRL 9 through system development, evaluation in at least five hospitals, and SaMD-readiness documentation. The items below are source-based goals or directions, not a report that every milestone is complete.
01
Evidence map
The two datasets answer questions in different settings. The 2024 work evaluates classification of held-out clips; Rayong compares results with clinician assessment in a clinical-practice cohort. Presenting them together shows the development path without treating them as one study.
Source: 2024 paper, pp. 123–124; Rayong abstract, pp. 1–3; CNF 2026, pp. 55–59
| Evidence | Unit and size | Overall result | Material limitation |
|---|---|---|---|
| 2024 publication | 184 test clips; 739 training clips | 177/184 = 96.2% | Mildly Reduced EF: 12/17 ≈ 71% |
| Rayong study summary | 218 participants | 93.6% | fair LV: 27.3%; peer review not verified |
| Multisite evaluation plan | Target of at least five hospitals | No completed result reported | Planned in the project brief |
02
2024: a published foundation
The article in the Journal of Prapokklao Hospital Clinical Medical Education Center, volume 41, issue 2, pages 123–132, describes smartphone-recorded parasternal long-axis video. It includes 923 usable clips: 739 for training and 184 for testing.
The 96.2% figure concerns three-category classification against clinician reference assessment. It is not a trial result demonstrating lower mortality or replacement of standard examination. Data selection and limitations should be read alongside the abstract.
Source: 2024 paper, pp. 123–132
03
Rayong: evidence from a clinical workflow
The Rayong abstract describes a retrospective study of 218 participants. The slides date it to December 2025–January 2026, with 76 pre-chemotherapy assessments and 142 general LV function assessments. This adds workflow evidence in a setting different from a clip split used for model development.
The available sources are an abstract and the CNF slides, rather than a full article with verified review status. Results are therefore presented as reported by the study team, with the 27.3% fair LV limitation included whenever performance is summarized.
Source: Rayong abstract, pp. 1–3; CNF 2026, pp. 54–59
04
Why the intermediate category matters
Overall accuracy reflects both the number of examples and performance within each category. A large category can produce a strong overall result while another remains weak. The 2024 study reports only 12 correct classifications among 17 Mildly Reduced EF clips, and Rayong reports 27.3% for fair LV.
The category labels and evaluation settings differ, so the two percentages should not be treated as a direct trend measurement. What both sources support is the need for close examination of the intermediate category.
Source: 2024 paper, pp. 130–132; CNF 2026, p. 59
05
Questions for the next evaluation
Planned Multi-site Clinical Validation in at least five hospitals is intended to assess how results vary across personnel, equipment, and participant populations. The brief also includes User Acceptance and Deployment Feasibility work to examine readiness beyond classification percentages.
Future evidence should report unusable clips, subgroup results, system versions, and reference-assessment procedures. These explain the conditions behind the headline result without assuming that adding sites will automatically improve performance.
Source: 2024 paper, p. 132; EasyEF project brief, 22 June 2026
SOURCE NOTES
Sources for this page
- บทความวิจัยปี 2567 / 2024 original article — J Prapokklao Hosp Clin Med Educat Center 41(2):123–132 ↗
- บทคัดย่อการศึกษาระยอง ไทย–อังกฤษ หน้า 1–3 / Rayong study abstract, Thai and English, pp. 1–3; publication and peer-review status not verified / ยังไม่ยืนยันสถานะตีพิมพ์และการทบทวนโดยผู้ทรงคุณวุฒิ
- เอกสารนำเสนอ / Presentation: EasyEF Lunch Symposium, CNF 17th, 26 March 2026 — หน้า / pp. 54–59
- ข้อเสนอโครงการ EasyEF ฉบับ 22 มิถุนายน 2569: ขอบเขตและแผนงาน / EasyEF project brief, 22 June 2026: scope and planned work
Presentations reflect information at the time they were prepared. Explore context and evidence status in the resource library.
Explore the source library →