Background:

Echocardiography is the most important and most frequently performed modality of cardiac imaging. In addition to manual skills in transducer use, rapid and valid visual assessment is a critical skill for image interpretation. However, it is still unclear whether there are cardiac structures that influence eye tracking or how viewers assess echocardiographic images in general. Therefore, there is neither guidance, nor advice for learners how to achieve valid image interpretation.  The purpose of this study was, to determine differences in viewing patterns between echocardiography experts and novices. 

Methods 

Participants were shown 21 apical four-chamber views. Eye movement of students without echocardiography training (n=15) and of experts (n=15) were recorded over predefined areas, using a 1000Hz eye tracker (data collection 06.12.21 to 07.01.22). Participants were asked to evaluate left ventricular ejection fraction and image quality. The local Ethics Committee approved the study (E-01-20210325).

Results:

The direction of eye tracking differed significantly between the two groups. While the students fixated longer on the center of the cavum of the left ventricle (723.45ms ± 103.05ms vs. 501.76ms ± 27.47ms p<0.001), the experts observed more often the wall of the left ventricle (2.31 ± 0.17 vs. 1.87± 0.13; p=0.024). First fixation duration on the left ventricle lumen is significantly longer (723.45ms ± 103.05ms vs. 501.76ms ± 27.47ms in students’ group; p<0.001). The number of fixations on the left ventricle wall was significantly higher in experts’ group (2.31 ± 0.17 vs. 1.87± 0.13; p=0.024).  Interestingly, view of the students lasted significantly longer on the moving mitral valve than in the experts (2254.86ms ± 324.20ms vs. 694.24ms ± 103.13ms vs. p<0.001).

Conclusion:

Eye-tracking demonstrated significant differences in the viewing pattern between ultrasound experts and inexperienced observers. Inexperienced observers looked more clearly longer at the cavum of the left ventricle and the moving valve. In contrast, experts necessarily analyzed mainly the wall of the left ventricle for estimation of LVEF. 

Further research is needed to investigate whether these findings can be used to actively direct the attention of inexperienced examiners to enable a faster learning curve.