Computer-aided evaluation of cataract surgery; a metric comparison of continuous circular capsulorhexis by trainee and specialist surgeons
Aniyath, Praseedha Krishnan
Seow, Kiam Tian
Kwok, Jian Wah
Fam, Han Bor
Heng, Wee Jin
Date of Issue2015
Investigative Ophthalmology & Visual Science
School of Computer Engineering
Purpose: There is a correlation between the centration and quality of the continuous circular capsulorhexis (CCC) and the subsequent refractive outcomes in cataract surgery. We developed a novel software evaluation tool based on video processing to assess the execution of CCC by comparing trainee and specialist surgeons from a teaching hospital. The software incorporates a novel performance metric that quantifies their performance. Methods: We first detected the limbus of the eye in each video frame using Hough circle detection. Next, the capsulorhexis forceps is detected based on its linearity and specularity. Then a visual tool-tracking function is invoked based on an image similarity measure which is illumination invariant and computationally inexpensive. The number of capsular grasps is then found from a functional plot of distance between the pair of forceps tips. Other parameters computed include surgical efficiency with respect to surgical time, circularity index and absolute decentration of the CCC with respect to the optical centre. These parameters are integrated into a single novel performance metric for each surgery (Fig 1). Results: The software was implemented in MATLAB and we evaluated 35 capsulorhexis videos of surgeries done by 19 specialist and 16 trainee surgeons. The quantitative parameters for all videos are listed in Fig 2. A student t-test comparison of the mean performance metric scores found that the trainee group scored 0.4244 (±0.2) which was significantly lower than the specialist group which scored 0.8676 (±0.1) (P=0.0001), indicating that the two groups could be differentiated. Conclusions: We developed a tool for evaluation of the performance of capsulorhexis during cataract surgery. The proposed performance metric computed by the software could differentiate the two groups of surgeons. Using quantitative parameters, we can have an objective and repeatable way for surgical assessment to identify areas for improvement.
Computer Science and Engineering
© 2015 Association for Research in Vision and Ophthalmology (ARVO). This is the author created version of a work that has been peer reviewed and accepted for publication by Investigative Ophthalmology & Visual Science, Association for Research in Vision and Ophthalmology (ARVO). It incorporates referee’s comments but changes resulting from the publishing process, such as copyediting, structural formatting, may not be reflected in this document. The published version is available at: [http://iovs.arvojournals.org/article.aspx?articleid=2331014&resultClick=1].