Minimally invasive surgery—endoscopic retinal detachment repair in patients with media opacities

Feature Tracking and Segmentation in Real Time via Deep Learning in Vitreoretinal Surgery: A Platform for Artificial Intelligence-Mediated Surgical Guidance

PURPOSE

This study investigated whether a deep-learning neural network can detect and segment surgical instrumentation and relevant tissue boundaries and landmarks within the retina using imaging acquired from a surgical microscope in real time, with the goal of providing image-guided vitreoretinal (VR) microsurgery.

DESIGN

Retrospective analysis via a prospective, single-center study.

PARTICIPANTS

One hundred and one patients undergoing VR surgery, inclusive of core vitrectomy, membrane peeling, and endolaser application, in a university-based ophthalmology department between July 1, 2020, and September 1, 2021.

METHODS

A dataset composed of 606 surgical image frames was annotated by 3 VR surgeons. Annotation consisted of identifying the location and area of the following features, when present in-frame: vitrector-, forceps-, and endolaser tooltips, optic disc, fovea, retinal tears, retinal detachment, fibrovascular proliferation, endolaser spots, area where endolaser was applied, and macular hole. An instance segmentation fully convolutional neural network (YOLACT++) was adapted and trained, and fivefold cross-validation was employed to generate metrics for accuracy.

MAIN OUTCOME MEASURES

Area under the precision-recall curve (AUPR) for the detection of elements tracked and segmented in the final test dataset; the frames per second (FPS) for the assessment of suitability for real-time performance of the model.

RESULTS

The platform detected and classified the vitrector tooltip with a mean AUPR of 0.972 ± 0.009. The segmentation of target tissues, such as the optic disc, fovea, and macular hole reached mean AUPR values of 0.928 ± 0.013, 0.844 ± 0.039, and 0.916 ± 0.021, respectively. The postprocessed image was rendered at a full high-definition resolution of 1920 × 1080 pixels at 38.77 ± 1.52 FPS when attached to a surgical visualization system, reaching up to 87.44 ± 3.8 FPS.

CONCLUSIONS

Neural Networks can localize, classify, and segment tissues and instruments during VR procedures in real time. We propose a framework for developing surgical guidance and assessment platform that may guide surgical decision-making and help in formulating tools for systematic analyses of VR surgery. Potential applications include collision avoidance to prevent unintended instrument-tissue interactions and the extraction of spatial localization and movement of surgical instruments for surgical data science research.

FINANCIAL DISCLOSURE(S)

Proprietary or commercial disclosure may be found after the references.

VISUAL AND ANATOMIC OUTCOMES OF PEDIATRIC ENDOSCOPIC VITRECTOMY IN 326 CASES

PURPOSE

To report on the indications, outcomes, and complications of endoscopic vitrectomy in a large cohort of pediatric vitreoretinal patients.

METHODS

This is a retrospective interventional case series consisting of 244 eyes of 211 patients aged 18 years or younger undergoing a total of 326 endoscopic vitrectomies from 2008 to 2017. A 23-gauge vitrectomy was performed with use of a 19-gauge endoscope.

RESULTS

Two hundred and eleven patients with a mean age of 7.5 years (range: 0-18 years) and median follow-up since last surgery of 28 months (range: 3 months-8.7 years) were included. The most common indication for endoscopic vitrectomy was retinal detachment (234/326; 72%) with proliferative vitreoretinopathy (162/234; 69%). Other diagnoses included trauma (25%), retinopathy of prematurity (15%), and glaucoma (9%). Twenty-five percent of surgeries (80/326) were performed on eyes with significant corneal opacities. Retinal reattachment was achieved in 67% of eyes with retinal detachment (119/178). Visual acuity improved in 26% of retinal detachment eyes versus 53% of nonretinal detachment eyes (P = 0.005). Surgical complications included band keratopathy (15%), hypotony (8%), cataract (7%), and elevated intraocular pressure (3%).

CONCLUSION

In this large series of pediatric endoscopic vitreoretinal surgeries, anatomic outcomes and complication rates were comparable with previous studies.

Endoscopic vitreoretinal surgery: Review of current applications and future trends

Endoscopy provides unique optical properties to circumvent anterior segment opacities and visualize difficult-to-access anatomical regions, including retroirideal, retrolental, ciliary body, and anterior retinal structures. We summarize the basic principles and utilization of endoscopic vitreoretinal surgery, along with recent technological advances in the field base on a structured literature search in Pubmed, Embase, and Google Scholar database up to February, 2020. Endoscopy has been used in the management of retinal detachment, ischemic retinopathies with neovascular glaucoma, severe ocular trauma, endophthalmitis, lens-related disorders in the posterior segment, pediatric vitreoretinal diseases, and implantation of retinal prostheses. Ongoing development of endoscopic technology aims to provide higher resolution images with endoscopes of smaller diameter. New surgical techniques supported by the adoption of endoscopy are available to manage challenging surgical scenarios. Endoscopy can be a useful adjunct to microscope wide-angle viewing systems in the management of complex vitreoretinal diseases.

B-scan ultrasound, visual electrophysiology and perioperative videoendoscopy for predicting functional results in keratoprosthesis candidates

BACKGROUND/AIMS

We analysed the ability of B-scan ultrasound, ocular electrophysiology testing and videoendoscopic examination for predicting visual prognosis in Boston Type 1 keratoprosthesis (KPro-1) candidates. Indirect anatomical and electrophysiological findings and results from direct endoscopic evaluations were correlated with postoperative functional data.

METHODS

In this prospective and interventional study, we included 13 individuals who had previously been indicated for Kpro-1 surgery. All subjects underwent preoperative screening, including ophthalmic evaluation, B-scan ultrasound, electrophysiological testing, and perioperative intraocular videoendoscopic evaluation (VE). B-scan ultrasound, electrophysiological testing, and VE evaluation results were categorised as favourable or unfavourable predictors of postoperative functional results according to predefined criteria. The predictability values of B-scan ultrasound, electrophysiological testing, and VE prognostication were calculated based on the visual acuity level achieved.

RESULTS

All surgeries and perioperative VEs were uneventful. Preoperative best-corrected visual acuity (BCVA) ranged from light perception to counting fingers. The 1-year postoperative BCVA was better than 20/200 (satisfactory visual acuity result) in 10 eyes (76.9%) and 20/40 or better in 5 eyes (38.5%). B-scan ultrasound presented a positive predictive value (PPV) of 85.7% for satisfactory postoperative visual acuity, electroretinography showed a PPV of 66.7%, and visual evoked potential presented a PPV of 66.7%. The perioperative VE PPV of a negative finding for satisfactory visual acuity was 100%.

CONCLUSIONS

Fundoscopic visualisation by intraocular VE is a minimally invasive procedure that can be used to predict functional outcomes in keratoprosthesis candidates. This technique demonstrated better prognostication in keratoprosthesis candidates than B-scan ultrasound and electrophysiological testing.

Consolidation of Imaging Modalities Utilizing Digitally Assisted Visualization Systems: The Development of a Surgical Information Handling Cockpit

The management of vitreoretinal cases is ever-evolving, paralleled by rapid advancements in operative imaging modalities. In this article, we describe an advanced application of digitally assisted vitreoretinal surgery (DAVS) that involves the consolidation of pre-existing ancillary imaging technology into a single same-screen viewing platform. Forty-four eyes of 44 patients were operated using same screen simultaneous viewing of the primary three-dimensional high definition (3DHD) surgical field and simultaneous auxiliary video feed viewing of all currently approved ocular endoscopy (n=12), intraoperative optical coherence tomography (iOCT) units (n=24), or computer feeds from the EHR/image management software (n=8). All surgeries were successful with excellent functional and anatomic outcomes. DAVS facilitated same screen viewing of multiple video/information feeds was notable for improved ergonomics, surgical efficiency, and precision when compared to viewing the surgical field and auxiliary video feeds separately. We describe a new concept for the vitreoretinal operating room – a DAVS-based surgical information handling cockpit – integrating FDA approved ocular endoscopy (n=1), microscope-integrated iOCT units (n=3), and one EHR/Image management solution with the primary surgical field 3DHD feed. We suggest same screen viewing of multiple video and other clinical information feeds is a promising modality that may be considered in the management of patients with surgical vitreoretinal disease and should be purposefully incorporated into future iterations of DAVS technology platforms.

Endoscopy-assisted vitrectomy in the anterior vitreous

Endoscopy-assisted ocular surgery is a relatively old technique that is increasingly being recognized for its application in cases of vitreoretinal disease. This technique is especially useful when both the vitreous and retina are difficult to access because of media opacity, a small pupil, or a microcornea. In this context, the anterior vitreous is often difficult to dissect because of its complex pathological changes. This article reviews the common anatomical features and pathologies that are observed in the anterior vitreous, as well as the applications and indications of endoscopy-assisted vitrectomy in the anterior vitreous.

Twenty five-gauge endoscopic vitrectomy for proliferative vitreoretinopathy with severe corneal opacity

PURPOSE

To report 4 cases undergoing 25-gauge endoscopic vitrectomy for the treatment of proliferative vitreoretinopathy with severe corneal opacity in which a transpupillary view of the fundus was not possible.

STUDY DESIGN

A retrospective interventional case series.

METHODS

The main outcomes measured were postoperative anatomic status of the retina and subjective improvement of vision.

RESULTS

Postoperative reattachment of the retina and subjective improvement of vision were achieved in all 4 eyes.

CONCLUSION

Twenty five-gauge endoscopic vitrectomy provides a clear view making it possible conduct pars plana vitrectomy in order to reattach the retina in cases of proliferative vitreoretinopathy with severe corneal opacity.

Treatment of Bacillus cereus endophthalmitis with endoscopy-assisted vitrectomy

To evaluate the use of endoscopy-assisted vitrectomy in patients with sight-threatening Bacillus cereus endophthalmitis.A retrospective analysis was conducted in 15 eyes with Bacillus cereus endophthalmitis. Patients were divided into 2 groups: endoscopy-assisted vitrectomy (5 eyes) and conventional vitrectomy (10 eyes). The following clinical data were recorded and analyzed: sex, age, latent period, symptom duration, follow-up time, visual acuity pre- and postsurgery, recurrence of endophthalmitis, incidence of phithisis bulbi, and incidence of enucleation.In the conventional vitrectomy group, postoperative visual acuity ranged from no light perception in 5 patients (50%), light perception in 3 patients (30%), 20/1000 in 1 patient (10%), and 20/50 in 1 patient (10%). In the endoscopy-assisted vitrectomy group, postoperative visual acuity ranged from no light perception in 2 patients (40%), light perception in 1 patient (20%), and hand movements in 2 patients (40%). There was no statistically significant difference between the 2 groups in terms of the final postoperative visual acuity (F = 0.006, P = .937). There is no difference between the 2 groups in terms of the incidence of enucleation. The median symptom duration was 4 hours (range: 2-6 hours) in the conventional group and 9 hours (range: 7-11 hours) in the endoscopy-assisted vitrectomy group. The difference in the symptom duration between the 2 groups was statistically significant (P = .002).There is no statistical significant difference between the 2 groups in terms of visual acuity and incidence of enucleation. Therefore, endoscopy-assisted vitrectomy can be considered as an alternative treatment for treatment of B cereus endophthalmitis particularly for cases when symptom duration was more than 6 hours.

Clinical outcomes of endoscope-assisted vitrectomy for treatment of rhegmatogenous retinal detachment

Summary

We evaluated the clinical outcomes for ophthalmic endoscope-assisted vitrectomy in consecutive patients with uncomplicated rhegmatogenous retinal detachment (RRD). The primary success rate was 98.4% (125/127) without performing a posterior drainage retinotomy or using perfluorocarbon liquids (PFCL) for subretinal fluid drainage.

Purpose

To investigate the clinical outcomes of endoscope-assisted vitrectomy in patients with uncomplicated RRD.

Methods

We examined 127 eyes from consecutive patients who underwent repair of RRD by 23- or 25-gauge endoscope-assisted vitrectomy, with a minimum follow-up of 3 months. Eyes with the following criteria were excluded: Giant retinal tears, grade C proliferative vitreoretinopathy, dense vitreous hemorrhage, retinal detachment secondary to other ocular diseases, and prior retinal or vitreous surgery. All cases underwent subretinal fluid drainage, endolaser photocoagulation and fundus inspection were performed under ophthalmic endoscopic observation. Success rate, visual acuity, surgery time and complications were evaluated.

Results

Primary and final success rate was 98.4% (125/127) and 100% (127/127), respectively, Surgery time was 59.6±26.3 minutes. The best-corrected visual acuity significantly improved from 20/100 to 20/20 (P<0.0001). There were 2 cases (1.6%) of creation of a peripheral drainage retinotomy and 4 cases (3.1%) of using PFCL to suppress movement of the detached retina, but there were no cases of creation of a posterior drainage retinotomy or using PFCL for subretinal fluid drainage. There was 1 case of presumed endophthalmitis after surgery. There were 12 hypotonous cases at postoperative day 1 and one of them needed additional scleral sutures at postoperative day 4 for prolonged hypotony.

Conclusion

The present study demonstrated the efficacy of endoscope-assisted vitrectomy for patients with uncomplicated RRD. To perform endoscope-assisted vitrectomy safely, sufficient closure of sclerotomies is necessary at the end of surgery.

Role of Diagnostic Endoscopy in Posterior Segment Evaluation for Definitive Prognostication in Eyes With Corneal Opacification

PURPOSE

Direct visualization of the posterior segment is not possible with conventional imaging techniques in eyes with corneal opacification. This study evaluated the efficacy of videoendoscopy in visualizing and hence prognosticating the visual outcomes in such eyes.

DESIGN

Prospective, noncomparative, consecutive interventional case series.

METHODS

This study included 64 eyes of 64 patients with corneal opacification who underwent a diagnostic ophthalmic endoscopy procedure from March 2014 to May 2016. All eyes also underwent conventional B-scan ultrasonography as part of their comprehensive ocular examination. The findings that were noted during endoscopy included the optic disc characteristics, anatomic integrity of the retina, and condition of the retinal blood vessels. The principal outcome measure was the determination of whether the results of the videoendoscopy helped in establishment of or brought a change in the management plan; accordingly, they were classified as either "contributory" or "noncontributory."

RESULTS

The videoendoscopy findings were classified as contributory in 62 of 64 eyes. The remaining 2 eyes had vitreous hemorrhage; endoscopic vitrectomy was done in them to detect an inoperable retinal detachment in 1 eye. With the help of the stereoscopic picture of the fundus provided by the endoscope, a poor visual prognosis in 17 eyes with retinal detachment and 22 eyes with attached retina was determined. Additionally, in eyes that had an attached retina on ultrasound B-scan, the endoscope helped in detecting a glaucomatous disc in 10 of those eyes; 7 of them had glaucomatous optic atrophy. In 3 patients ultrasound did not detect a retinal detachment (false-negative). One eye with detached retina on ultrasound did not actually have a retinal detachment (false-positive). Thus ultrasound had a sensitivity of 88.4% and a specificity of 97.3% in detecting RD as compared with an endoscopic evaluation. Based on the endoscopic findings, 34 of 64 (53%) eyes were identified as having a better visual prognosis and underwent subsequent vision-restoring procedures. This included 10 (16%) patients who directly underwent a corneal procedure and 24 (37.5%) patients who had endoscopic vitreoretinal procedures initially before undergoing a corneal procedure. Endoscopy helped in identifying poor prognosis in 30 of 64 (46.8%) eyes; thus, further invasive procedures could be avoided. The endoscopy procedure in itself did not lead to any complications.

CONCLUSION

Diagnostic endoscopy provided invaluable supplemental information about the posterior segment in eyes with corneal opacification. This helped in better prognostication and in reserving corneal reconstructive procedures for eyes with a favorable visual prognosis.

Use of intraocular video endoscopic examination in the preoperative evaluation of keratoprosthesis surgery to assess visual potential

PURPOSE

To determine the clinical utility of intraocular videoendoscopy examination for the evaluation of the retina and optic nerve in patients being considered for a Boston type I keratoprosthesis (KPro).

DESIGN

Interventional case series study.

METHODS

Ten patients with a history of corneal blindness caused by failed penetrating keratoplasty (PK) and inability to accurately assess visual potential were included in this study. Ophthalmologic examination, B-scan ultrasonography, and pars plana videoendoscopy were carried out to assess the retina and optic nerve before KPro.

RESULTS

Posterior segment examination was successfully used to evaluate the retina and optic nerve of all patients with opaque corneas. Out of 10 patients that underwent endoscopic examination, 3 (30%) were considered to be adequate candidates for KPro surgery and 7 (70%) were not. This was based on visualized retinal disease and/or optic nerve pathology. Of the 3 patients that underwent KPro surgery, all of them had a significant improvement of vision, including counting fingers to 20/100, hand motion to 20/5, and light perception to 20/80, as suggested by the endoscopy preoperative examination. No complications of the endoscopy procedure were observed.

CONCLUSIONS

This report demonstrates the successful use of intraocular videoendoscopy to rule out threats to a good visual outcome for patients being considered as candidates for KPro. Direct visualization of the posterior segment can be part of the preoperative algorithm in the decision process of performing a KPro surgery in patients when visual potential is questionable.