1University of Edinburgh, Edinburgh, United Kingdom2Western General Hospital, Edinburgh, United Kingdom
Checking for direct PDF access through Ovid
Echocardiography, while a prevalent tool for assessing cardiac morphology and function, suffers from a range of artefacts that reduce its diagnostic value.This work qualitatively and quantitatively evaluates 3 novel post-processing methods for enhancing echocardiographic images. Data enhancement is achieved by utilising multiple partially decorrelated instances of a cardiac cycle acquired through a single acoustic window. Such information has until now been largely disregarded during data post-processing. Moreover, unlike past approaches, data enhancement is achieved without filtering out information based on static or adaptive selection criteria.Qualitative assessment using 32 clinical datasets demonstrated (i) suppression of cavity noise, (ii) increase in tissue/cavity contrast, and (iii) visual enhancement of tissue structures previously masked-out by various artefacts (Figure 1). The effect of each post-processing method on the diagnostic value of cardiac ultrasound data was quantitatively assessed by examining the repeatability coefficient variations (via Bland-Altman plots) in clinical measurements performed by 2 echocardiographers. Routine clinical measurements such as wall thickness and cavity diameter were employed. The effect of each method varied depending on the echocardiographer and the measurement performed. Improvement of up to 57% on the repeatability coefficient and reduction by up to 47% on the number of omitted measurements due to heavy data corruption were achieved.All methods demonstrate potential in providing a powerful assistive tool to B-mode imaging during challenging cardiac ultrasound examinations. Since both echocardiographers were novices with the processed data, it is believed that there is further scope for improvement as their familiarity with such data increases.