In this study, we present a post-operative energy dashboard to teach surgeons about electrosurgical device use. We showed that the energy generator and laparoscopic video data can accurately be transformed into a post-operative report. In addition, the pilot study showed significant differences in the application by the different surgeons. Therefore, the study results indicated that the proposed methodology can provide new insights for surgical skills assessment.

The comparison in Fig. 1 shows the potential insights energy monitoring can add to the current surgical skills assessment. Personal scores should be compared to large datasets on comparable surgeries to obtain well-founded insights. Therefore, we plan to expand to a multi-center study with a larger sample of surgeons and surgical procedures. For this, the platform must be deployed on a large scale. The proposed method of obtaining data on the application of electrosurgery is independent of the type of diathermy or the type of procedure, which offers the opportunity to expand the use of the energy dashboard easily. However, several post-processing and manual steps are still required to acquire the knowledge. A direct, automated data integration from the energy generator and the laparoscopic video would facilitate the broader dashboard application. Ultimately, it should be investigated whether the energy dashboard and this novel form of skills assessment led to an improved application of electrosurgery, which means less energy while maintaining adequate hemostasis and dissection. Clinical studies will have to determine the ideal energy use for effective hemostasis with minimal tissue coagulation. This future work will correlate energy metrics with hemostasis quality to establish optimal energy use.

Assessing the application throughout the different procedural phases provided further in-depth insight into this dynamic use. Figure 6 showed how much energy device use differs between the surgical phases. This distinction showed that the large differences in the total applied energy of Surgeon 2 with respect to Surgeon 1 and 3 (Table 1) mainly resulted from its application in the gastrosplenic phase. In addition, the applied energy in the preamble and reconstruction phase resulting from adhesiolysis could be distinguished from the energy applied in other surgical phases. These insights highlight the importance of integrating the video and energy data sources. In certain surgical phases, it may be preferable to provide less energy as minimizing thermal spread to neighboring tissues could be more important than ensuring perfect hemostasis, e.g., in the vicinity of Vagal nerve branches. Therefore, surgeons should adopt a dynamic approach when using a sealing tool. A phase-level evaluation of the device application enabled the assessment of this dynamic approach. In the next phase, we aim to automate this analysis by computer vision-based surgical phase detection [12].

In addition to monitoring electrosurgery use, we have developed algorithms to quantify blood loss as an indicator for the effectiveness of hemostasis. The objective of the dashboard is to perform surgery with minimal energy and tissue coagulation while maintaining optimal hemostasis. Therefore, the extent to which the activations of the device lead to bleeding should also be quantified in the dashboard. Device-induced bleeding should be identified by linking the activation moments in the generator data to the onset of bleeding in the laparoscopic video. Our current work focuses on AI-based blood detection algorithms near device activation on the timeline of generator-derived tool activation information [13, 14].

Future studies could focus on the broader application of the energy dashboard. This study only focused on a vessel sealing device with fixed generator settings, but electrosurgical devices with variable settings (e.g., surgical pencils) could also be assessed. The methodology presented can be adapted to acquire the relevant metrics of this device type. Generator settings (e.g., coagulate vs. cutting mode) can be derived based on the manufacturer to enable a more detailed analysis of electrosurgical practices. In addition, the current study provides insights into energy use postoperatively but does not offer real-time feedback. Future studies could potentially evaluate the impact of real-time dashboard feedback on energy use, although we need to consider the generalizability of the feedback for individual cases.