Researchers have developed the first automated phase detection system to use AI in small incision cataract surgery (SICS).
Their findings have been published in Scientific Reports.
Although algorithms for AI-assisted video analysis in cataract surgery exist in high-income countries, there are currently no datasets or algorithms available for SICS.
The international team at the Sankara Eye Foundation India, led by the University Hospital Bonn (UKB) and the University of Bonn, is now launching a global AI competition at the MICCAI 2025 conference in Daejeon, South Korea, where AI algorithms for surgical phase detection will compete against one another.
Cataracts are the leading cause of blindness globally, particularly impacting individuals in low- and middle-income countries like India.
The low-cost and effective SICS surgical method is preferred in these countries; however, it is often associated with poorer outcomes due to limited resources and training opportunities.
Dr Maximilian Wintergerst, head of a working group at the UKB Eye Clinic and principal investigator of the study, said: ‘In addition, the application of AI on this technique has not yet been researched enough.’
The new study now makes videos of manual small-incision cataract operations publicly available for the first time, using the SICS-105 data set. This is based on operations conducted on 105 patients at Sankara Eye Hospitals in India.
The study shows that the innovative deep learning model MS-TCN++, developed by Professor Dr Jürgen Gall’s group at the University of Bonn, can recognise different surgical phases, such as preparation of the surgical approaches to the eye and various surgical steps on the lens, with over 85% accuracy.
Dr Kaushik Murali, president of medical administration at the Sankara Eye Foundation India, said: ‘The analysis of surgical phases is important because it enables a quantitative comparison between different surgeons, feedback on identified critical steps and the detection of deviations from surgical protocols. It is therefore the first step towards automatic assessment of surgical quality.’
The research consortium is now calling for an AI competition to analyse surgical videos. To this end, the research team has expanded the study's first public data set for SICS, including surgical videos and hand-marked (annotated) surgical phases, to encompass a total of 155 operations with 18 different phases. Researchers at Microsoft Research India and Sankara Eye Foundation have developed the software for annotation. Subsequently, the annotations have been created by ophthalmologists at the Sankara Eye Foundation.
The research team anticipates that participants will submit an algorithm for predicting surgical phases based on the provided video data, as well as write a brief paper on their approach.
- Submission deadline is 15 August 2025. For further information, visit https://medvisbonn.github.io
Photo caption - Sankara Eye Foundation India & University Hospital Bonn
