Assessment of cardiac image quality from single-probe handheld ultrasound devices

This study aims to assess quality of cardiac ultrasound imaging from commercially available single-probe handheld devices.

Cardiac imaging from single-probe handheld ultrasound devices was prospectively acquired by registered cardiac sonographers. Videos were graded on a scale of 1 (uninterpretable) to 5 (diagnostic-quality imaging) by 2 sonographers. The average image quality score for each cine loop was analysed. Interpretable imaging was defined as an average IQ score ≥ 2. Good image quality was defined as an average IQ score ≥ 4. Imaging from single-probe devices representing 518 videos from 206 patients was assessed. The mean age was 72 (SD 11.1) years. Agreement on image quality was moderate (weighted kappa 0.50, 95% CI 0.45-0.55). The correlation between sonographer quality grades was 0.60. The median average IQ score was 2.5 (1.5). 454 (88%) of the videos were interpretable (average IQ ≥ 2). At least one interpretable video was obtained for 187 (91%) patients. 62/518 (12%) of videos had good image quality (average score ≥ 4). There was a trend towards lower image quality with increasing BMI categories [16-25, median 3.0 (2.0, 3.5), 25-29, median 3.0 (2.0, 3.5), > 29 median 2.5 (2.0, 3.0), P < 0.001].

Basic interpretable imaging from single-probe handheld ultrasound devices can be obtained for most patients under idealized conditions, suggesting possible use for enhanced cardiac disease detection. Since imaging from a significant minority of patients is uninterpretable and lower image quality is associated with higher BMI, rigorous assessment of image quality is essential prior to image interpretation.