Author + information
- Ronald B. Chin,
- Cassie Tighe,
- Timothy Beaver,
- Terrence Welch and
- Salvatore Costa
Background: Technological advances have made it feasible to obtain real-time 3D volume datasets using transthoracic echo; however, despite its widespread availability, only a few echocardiography labs utilize it in daily clinical practice. We hypothesized that an automated system would facilitate use in a real world setting and would compare favorably to expert review.
Methods: 50 consecutive patients were prospectively enrolled at an academic medical center. Analysis of each study was as follows: a sonographer (user 1) and cardiology fellow (user 2) applied HM2 with no editing (fully-automated) and then with regional edits (semi-automated). HM2 analysis was blinded to the LVEF as reported by an expert reader (ER). Comparisons were assessed using Pearson's correlation coefficient.
Results: The fully automated Heart Model demonstrates perfect correlation between users (r=1.0); however, when compared to the expert reader (ER) the correlation is poor (r=0.4). In contrast, the semi-automated method demonstrates good correlation in both settings, between users (r=0.87) and compared with the expert reader (r=0.6).
Conclusions: Some editing by the end-user remains necessary to maintain LVEF accuracy as compared to an expert reader; however, automated software that minimizes the initial learning curve and facilitates analysis continues to decrease barriers for adaptation of this quantitative tool into clinically busy echocardiography labs.
Poster Hall, Hall C
Sunday, March 19, 2017, 9:45 a.m.-10:30 a.m.
Session Title: Insights From Echocardiography Use
Abstract Category: 28. Non Invasive Imaging: Echo
Presentation Number: 1291-232
- 2017 American College of Cardiology Foundation