Compression of two-dimensional echocardiographic images: How far can we go?

The effect of the level of compression of two-dimensional echocardiographic images was studied on individual cardiac cycles from 283 patient examinations. The images were selected without attention to the original image quality and were digitized from video tape. A subjective assessment of image quality (scores of 1 = poor to 5 = excellent) by two experienced blinded observers was used to assess the effect of compression levels. Interobserver variability in terms of the image quality score increased at compression levels > 7:1 and a logistic regression model showed that image quality was related to the level of compression with fair to poor quality associated with compression levels of > 8.5:1. It is concluded that increased compression levels affect the quality of digital echocardiographic Images that were originally acquired on video tape. These results may have important bearing on the level of compression that is prudent to perform on echocardiographic images which are not selected based on the original image quality.