Using Serious Games for Research in Keyhole Surgery

Authors: Phoebe Whitley, Connor Creasey, Aayushi Oza, Matthew J. Clarkson, and Stephen Thompson

Keyhole surgery benefits the patient by reducing scarring, shortening the hospital stay, and reducing post operative pain. However, keyhole surgery is more difficult to perform partly because the surgeon can only see a small part of the anatomy at any one time. This limited field of view is commonly cited by scientists and engineers developing systems to improve key hole surgery. Such systems may use image mosaicking, augmented reality, or novel optics to increase the effective field of view. Understanding how effective these systems may be is made difficult by our own lack of quantitative knowledge about the specific impacts of limited field of view. For example, does a limited field of view increase the time required to perform and operation, or does it increase the number of mistakes made? Studying these questions clinically is very difficult due to the need to perform large numbers of trials weighed against the ethical issues with recruiting sufficient patients and surgeons. This post describes our recent efforts to answer these questions using abstract simulations.

We set out three years ago to design a browser based serious game to study the impact of reducing the field of view on operation time and mistakes made by unskilled users performing a very abstract version of surgery. The idea is that because the game is easily accessible and can be played by unskilled users, it can support the rapid collection of lots of data from diverse users. Our initial results have just been published in JMIR Serious Games.

We successfully demonstrated that a serious game could be used to examine the impact of reducing the field of view on task completion time. The first two versions of our game were built using the Phaser game engine. Game development was performed by MSc students from UCL’s Department of Medical Physics and Biomedical Engineering, with no prior JavaScript experience. Phaser provided the ideal combination of ease of use, tutorials, and community support to enable simple game development by novice users.

Version 1: Comparing player performance with and without limited field of view.

The initial version of the game was written in JavaScript and HTML, importing game functions from a stand alone copy of phaser.min.js (v3.60.0). The game consisted of 8 levels (only 7 of  which were used on our study). The game was played under the supervision of the MSc student (Phoebe Whitley) who recorded task completion time and the number of mistakes. The study was approved by UCL’s research ethics committee (Ethics ID 24249_001).

The game levels get progressively more difficult as the user learns more mechanics; how to cut a single vessel, understanding interwoven vessels, and how to execute the game when there is only a limited field of view. Paired comparison of task completion time for levels with and without limited field of view enabled us to prove the hypothesis that reducing the field of view will increase task completion time.

Version 2: Plotting task performance time versus field of view.

The second iteration of the game introduced a variable field of view, to see if we could measure a quantitative relationship between task completion time and the size of the field of view. Getting enough data to show a numerical relationship required automated measurement of task completion time, cursor movement, and the number of mistakes. To do this we created a web application using the Flask web application framework hosted using the Google app engine. All of the game logic was implemented in the frontend in JavaScript, once again using the Phaser game engine, whilst a back end written in Python handled communication with a Google hosted database.

The data collection protocol was approved by UCL’s research ethics committee (Ethics ID 24249/005) and all users had to give informed consent via a web form prior to playing the game. Using this approach we were able to rapidly recruit a large number of participants (~60) and gather significant amounts of data at different fields of view. We were able to show statistically significant relationships between the field of view and task completion time and cursor movement. These relationships will inform ongoing research in ways to enlarge the field of view for key hole surgery. You can play the second version of the game here.

Version 3: Measuring the effects of distractions on task performance.

The third iteration of the game introduced various audio and visual distractions to the game environment, whilst maintaining the limited field of view and basic cutting task. A typical surgical environment contains many distractions and unexpected events, so we wanted to test the players’ ability to respond to these whilst still performing a basic surgical task. Data collection (under UCL Ethics ID 24249/005b) and analysis of this iteration is still ongoing. You can play the game here.

Conclusion

Using open source tools and the Phaser game engine we have created engaging simulations to answer research questions relevant to minimally invasive (keyhole) surgery. Our work is helping to establish a new way of performing research in this important area.