Performance evaluation in cataract surgery with an ensemble of 2D-3D convolutional neural networks.

An important part of surgical training in ophthalmology is understanding how to proficiently perform cataract surgery. Operating skill in cataract surgery is typically assessed by real-time or video-based expert review using a rating scale. This is time-consuming, subjective and labour-intensive. A typical trainee graduates with over 100 complete surgeries, each of which requires review by the surgical educators. Due to the consistently repetitive nature of this task, it lends itself well to machine learning-based evaluation. Recent studies utilize deep learning models trained on tool motion trajectories obtained using additional equipment or robotic systems. However, the process of tool recognition by extracting frames from the videos to perform phase recognition followed by skill assessment is exhaustive. This project proposes a deep learning model for skill evaluation using raw surgery videos that is cost-effective and end-to-end trainable. An advanced ensemble of convolutional neural network models is leveraged to model technical skills in cataract surgeries and is evaluated using a large dataset comprising almost 200 surgical trials. The highest accuracy of 0.8494 is observed on the phacoemulsification step data. Our model yielded an average accuracy of 0.8200 and an average AUC score of 0.8800 for all four phase datasets of cataract surgery proving its robustness against different data. The proposed ensemble model with 2D and 3D convolutional neural networks demonstrated a promising result without using tool motion trajectories to evaluate surgery expertise.

Use of Machine Learning to Assess Cataract Surgery Skill Level With Tool Detection.

Purpose: To develop a method for objective analysis of the reproducible steps in routine cataract surgery. Design: Prospective study; machine learning. Participants: Deidentified faculty and trainee surgical videos. Methods: Consecutive cataract surgeries performed by a faculty or trainee surgeon in an ophthalmology residency program over 6 months were collected and labeled according to degrees of difficulty. An existing image classification network, ResNet 152, was fine-tuned for tool detection in cataract surgery to allow for automatic identification of each unique surgical instrument. Individual microscope video frame windows were subsequently encoded as a vector. The relation between vector encodings and perceived skill using k-fold user-out cross-validation was examined. Algorithms were evaluated using area under the receiver operating characteristic curve (AUC) and the classification accuracy. Main Outcome Measures: Accuracy of tool detection and skill assessment. Results: In total, 391 consecutive cataract procedures with 209 routine cases were used. Our model achieved an AUC ranging from 0.933 to 0.998 for tool detection. For skill classification, AUC was 0.550 (95% confidence interval [CI], 0.547-0.553) with an accuracy of 54.3% (95% CI, 53.9%-54.7%) for a single snippet, AUC was 0.570 (0.565-0.575) with an accuracy of 57.8% (56.8%-58.7%) for a single surgery, and AUC was 0.692 (0.659-0.758) with an accuracy of 63.3% (56.8%-69.8%) for a single user given all their trials. Conclusions: Our research shows that machine learning can accurately and independently identify distinct cataract surgery tools in videos, which is crucial for comparing the use of the tool in a step. However, it is more challenging for machine learning to accurately differentiate overall and specific step skill to assess the level of training or expertise. Financial Disclosures: The author(s) have no proprietary or commercial interest in any materials discussed in this article.

Ocular manifestations of disseminated Mycobacterium chimaera infection after cardiothoracic surgery.

OBJECTIVE: To describe the ocular manifestations of disseminated Mycobacterium chimaera infection after cardiothoracic surgery. DESIGN: Retrospective, observational case series. PARTICIPANTS: Patients with disseminated M. chimaera infection treated at the University of Alberta Hospital and Royal Alexandra Hospital in Edmonton, Alberta, Canada. METHODS: Seven patients (14 eyes) with systemic M. chimaera infection after aortic graft and/or valvular surgeries were reviewed for ocular involvement. Cases were identified based on histopathologic analysis of cardiac tissue, repeat positive mycobacterial blood cultures, and cerebrospinal fluid analysis. Clinical ocular findings, fundus autofluorescence, fluorescein angiography, spectral-domain optical coherence tomography (SD-OCT), and autopsy results are described. RESULTS: The mean age of our patients was 63 years (range, 22-76 years). Aortic graft and/or valvular surgeries were performed between June 2015 and April 2016. The mean duration from surgery to diagnosis of infection was 27 months (range, 19-36 months). All patients exhibited white-yellowish choroidal lesions. Bilateral, multifocal choroiditis was observed in 6 patients (12 eyes), who died of disseminated M. chimaera infection despite aggressive, multiagent antimicrobial therapy. One patient had a solitary, white-yellow choroidal lesion in the left eye only. Choroidal lesions were hyperfluorescent on late-phase fluorescein angiography and corresponded to localized choroidal thickening and retinal elevation on SD-OCT. CONCLUSIONS: M. chimaera infection after cardiothoracic surgery can cause life-threatening disease involving multiple organ systems. Ocular involvement is an important and possibly early clinical finding, with good correlation to systemic disease severity. Ophthalmological assessment and multimodal imaging may aid in early diagnosis and can be used to monitor disease progression.

Delayed Vision Recovery after Carotid Vascular Surgery for Branch Retinal Artery Occlusion.

Branch retinal artery occlusion (BRAO) is typically associated with irreversible vision and peripheral visual field loss. We report a case of a 62-year-old woman with a BRAO related to several cardiovascular risk factors. Our patient encountered gradual but significant vision recovery months following carotid artery endarterectomy for carotid stenosis.

Silicone oil migration into periorbital space surrounding an extraocular muscle after sutureless 23-gauge vitrectomy in a child.

We report the case of an 11-year-old girl with a rare finding of silicone oil migration into the periorbital space surrounding an extraocular muscle, discovered during strabismus surgery for a sensory exotropia that had developed after retinal detachment repair by pars plana vitrectomy with intraocular silicone oil.

Lid splinting eyelid retraction technique: a minimised sterile approach for intravitreal injections.

BACKGROUND/AIMS: To describe an alternative technique for avoiding contact with the lids and lashes, without the use of a lid speculum, during intravitreal anti-vascular endothelial growth factor injections. METHODS: Retrospective case series of all patients undergoing intravitreal injections of bevacizumab and ranibizumab, with the lid splinting retraction technique from January 2010 to December 2015. Injections performed by six vitreoretinal specialists were included. The key preinjection ocular surface preparation includes topical anaesthetic, 5% povidone-iodine and a subconjunctival injection of 2% lidocaine with epinephrine. A second instillation of 5% povidone-iodine is given and the intravitreal injection is then performed. No lid speculum is used. A search of the electronic medical records identified patients diagnosed with postinjection endophthalmitis and charts were reviewed to ensure inclusion criteria were met. The main outcome measure was incidence of postinjection endophthalmitis. RESULTS: A total of 78 009 consecutive intravitreal injections were performed, of which 22 207 were bevacizumab and 55 802 were ranibizumab. In this cohort of patients (n=6320), 12 cases of endophthalmitis developed, corresponding to a rate of 0.015%. CONCLUSIONS: The technique of eyelid retraction for intravitreal injection has a low rate of endophthalmitis, similar to the reported rates using a metal lid speculum. This is beneficial for both the physician and the patient as it minimises patient discomfort as well as the duration of the procedure. To our knowledge, this is one of the largest studies performed to date evaluating intravitreal injection-related endophthalmitis.