Discussion
In this study, we showed that using an automated TEE protocol resulted in a shorter TEE exam duration and a reduced number of sonographer keystrokes when compared with manual TEE protocols. This reduced exam duration (by an overall average of 1:23 min, or 12%) and total keystrokes (overall average reduction of 77 sonographer keystrokes, or 28%) persisted when comparing novice and expert TEE operators. Notably, the total number of image acquisitions was not significantly different between the exams that used the automated and manual protocols, suggesting that a similar number of images can be acquired in a more timely and efficient manner when using an automated protocol. When image quality was objectively evaluated by an independent echocardiographer, there was no significant difference in image quality between the images acquired during the manual protocol versus the automated protocol. Overall, these findings suggest that an automated protocol can improve TEE exam efficiency without a reduction in image acquisition quality.
We found that the second complete TEE exam performed on each patient was on average shorter than the first complete exam, regardless of which protocol (automated or manual) was performed first. This finding is most likely due to the operator’s and sonographer’s familiarity with the views, images, and cardiac pathology of that specific patient when performing the second TEE image acquisition. However, we did find that the percentage reduction in the duration of the second exam was significantly greater when the automated protocol was performed second. This further suggests that automated protocols improve TEE exam efficiency and duration.
To our knowledge, this is the first study to investigate automated protocol use in transesophageal echocardiography, an invasive imaging modality requiring both procedural technique and image acquisition.
Previous investigations found that scanner-based protocols reduced the variation in the number and sequence of images obtained for different ultrasound examination types; implementing automated protocols has been proved to improve the consistency, efficacy, and reduce the time spent on image acquisition for carotid Doppler and abdominal ultrasound examinations with Doppler imaging 5. Such consistencies in automated protocols not only allow for completeness of the imaging studies but also provide a common baseline in the delivery of patient care.
While outside the scope of this study, automated protocol-aided reduction in examination duration per TEE case would allow an echocardiography laboratory to accommodate a higher volume of cases. Further, as we have shown that automated TEE protocols are more efficient then manual TEE protocols, efficiency and exam duration should not be barriers to exporting automated TEE protocols to other echocardiography labs within a health system. By doing this, consistency and quality of TEE exams across a healthcare delivery system may be improved. By reducing reliance on manual protocols for echocardiography, a more efficient, less time consuming, and more reproducible complete TEE exams can be performed. This may lead to lowering the need for additional image acquisition or even repeating a TEE examination. When applied broadly to an echocardiography laboratory, and given time for both operator (physician) and sonographer to become proficient in a specific automated protocol software, automated protocols can reduce the time needed for TEE exams, reproducibility of acquisitions and allow for accommodating a higher volume of patients.
Although TEE is considered an advanced imaging technique and acquired by an experienced echocardiographer, applying an automated protocol can ensure a complete review of the anatomy in a consistent manner and image delivery to the reader. Additionally, providers referring patients to an echocardiography lab or health system can expect all TEE examinations to contain complete and consistent imaging and information, thereby reducing operator variability. One area that is outside the scope of this study is evaluating the performance of automated vs manual TEE protocols in the use of training new TEE operators. It is important to note that even our “novice” observers were advanced echocardiography fellows who had completed general TEE training and had experience and comfort with TEE acquisitions in the presence of an experienced echocardiographer. We found that by using the automated protocol the time needed to complete the study was reduced even in novice operators. Applying this method to training TEE operators may improve efficiency and exam completeness/quality. Further, by defining the views needed for each protocol, novice operators can learn in a more systematic approach the adjustments of the TEE probe to achieve the subsequent image and reduce training variability. Automated protocol acquisitions may ensure a baseline level of quality, consistency, and decreasing variability between operators within the same echocardiography laboratory, and ideally, in between different echocardiography laboratories.
The strengths of this study include the prospective evaluation of the duration and keystrokes using both manual and automated TEE protocols. Additionally, each protocol type (manual and automated) was performed on each patient, it is reasonable to assume that exam duration will vary per patient based on factors outside of the control of the operator and sonographer. By performing image acquisition sequences of manual and automated protocols on each patient, we are able to control for differences between patients, patient cardiac anatomy, and patient TEE imaging windows. Finally, the same sonographer and echocardiography machine were used for each exam to eliminate variability in sonographer skill/technique and machine system characteristics.
Limitations of this study include a relatively small sample size of patients in one academic echocardiography laboratory. However, we were able to show significant differences in time and keystroke efficiency between automated and manual TEE exams. While we chose to control for patient-specific factors (by performing a manual and automated acquisition on each patient), we found that the exam performed second was always faster. This is likely due to the operator and sonographer knowing the patient’s cardiac anatomy and imaging windows as it relates to TEE image obtainment and acquisition. However, we did show a greater reduction in TEE duration when the automated protocol was performed second. While we showed improvement in individual TEE exam duration/efficiency, we did not assess the effect of automated protocols on the overall echocardiography laboratory TEE throughput (for example, the number of TEE able to be performed on a daily or weekly basis). Such an assessment is out of the scope of this study. Finally, all operators (physicians) and sonographer were familiar with the software and workflow. One may assume that efficiency in automated protocols would require a learning curve of image acquisition and workflow. Again, this assessment would be outside the scope of this study.