London 2014 Registration Visa Letters Programme Satellite Meetings Glaucoma Day 2014 Exhibition Hotel Booking Virtual Exhibition Star Alliance
london escrs

Course handouts are now available
Click here

Come to London


WATCH to find out why

Site updates:

Programme Updates. Programme Overview and - Video Symposium on Challenging Cases now available.

Can you see what you're doing? A study of microsurgical microscope control

Search Abstracts by author or title
(results will display both Free Papers & Poster)

Session Details

Session Title: Interactive Poster Session 02: Cataract

Session Date/Time: Sunday 14/09/2014 | 14:00-15:00

Paper Time: 14:30

Venue: Interactive Poster Terminal (Poster Village)

First Author: : N.Din UK

Co Author(s): :    J. Wawrzynski   S. Jones   P. Smith   P. Sullivan   S. Caputo   G. Saleh

Abstract Details


Studies from allied surgical specialities that use an operating camera have noted the important link between surgical outcome and operative visualisation. The aim of our study is to utilise feature extraction and the latest computer vision techniques to track the location of the pupil during a recorded cataract surgery. We aim to compare the relationship between pupil centration and level of operative experience, and evaluate its potential impact.


This study was carried out at Moorfields Eye Hospital (Bedford Hospital and City Road), in affiliation with the NIHR biomedical research centre and UCL Institute of Ophthalmology. Visualisation of the operative field is an essential component to any successful surgical procedure. This aspect is even more critical during ophthalmic surgery when the operating microscope requires continuous adjustments to ensure optimal eye centration, focus and zoom to assess different depths of tissue or tracking patients’ eye movements. However, a number of barriers exist in acquiring these skills for successful operative visualisation of cataract surgery. Firstly, mastering the microscope pedal is an essential but a difficult skill requiring continuous practise. Secondly, phacoemulsification cataract surgery is further complicated by the need to independently use all four limbs simultaneously which is an unnatural state for the novice trainee. Finally, the trainee ophthalmic surgeon must overcome a steep learning curve of not being able to directly visualise the operative field but rely on an indirect operative field of view governed by the microscope. Hence, these microscope skills are difficult to master and appraise.


“Phaco-tracking” is where tracking methodology is used in phacoemulsification cataract surgery to analyse instrument and eye movement. A histogram of gradient (Hog) detector was applied to tracking the pupil position over time for all videos in our study. The inclusion criteria for this study were adult patients with informed consent, pupils dilated fully; 1+ nuclear sclerosis or cortical lens opacity only; able to fully lie flat and still for the duration of surgery; and no ocular comorbidity (eg, glaucoma or pseudoexfoliation). Exclusion criteria included unable to give informed consent, non-routine cataract (eg secondary to previous trauma or intraocular surgery); and concurrent pathology impeding a clear view (eg, corneal pathology). A prospective cohort of novices and expert surgeons were used in this study. All cataract cases were deemed suitable for the junior surgeons to allow for fairer comparison. Videos were recorded from the microscope platform. From this tracking result, the average distance of the centre of the pupil to the centre of the frame, in addition to the total path length travelled by the pupil was calculated for all frames. An independent t-test was applied to compare the relationship between novices and expert surgeons.


A total of 26 participants were recruited in this trial- 9 novices and 17 experts. The average pupil centre distance from frame centre and total path length travelled by the pupil were analysed and calculated. The p-values for an approximate t-test between the 2 groups were 0.0003 and 0.0277 for pupil centration and total pupil distance travelled. The results show that novice surgeons have their camera off the pupil centre at a far greater distance than expert surgeons. However, there was less variation in the pupil centration distance amongst the novice group. The average pupil centre distance from frame centre for novices had a mean of 6.899, with a standard deviation of 1.497, compared with the experts who had a mean of 3.686 and sd 1.930. The expert surgeons maintained good camera control and limited the total pupil path length travelled. Expert surgeons had a mean (SD) of total pupil length travelled of 17616 (7565) whereas novices had a mean (SD) of 25424 (8135).


This present study is the first in which control of the operating microscope in live cataract surgery has been objectively analysed using motion analysis technology. Our results demonstrate that surrogate markers for camera control- a small pupil centre to frame centre demonstrate construct validity. In addition, shorter microscope path length as evidenced by the distance travelled by the pupil is also an indicator for good camera control. An efficient and accurate ophthalmic surgeon aims to keep the surgical focus in their field of view at all times, making minor continuous camera adjustments throughout the procedure to ensure it lies within the designated frame. Novice surgeons often do not centralise their view as well as experts (pupil distance from frame centre) and adjust their view more frequently over the course of a procedure (path length).

Financial Interest:


Back to previous