AI-based support for optical coherence tomography in age-related macular degeneration

Artificial intelligence (AI) has emerged as a transformative technology across various fields, and its applications in the medical domain, particularly in ophthalmology, has gained significant attention. The vast amount of high-resolution image data, such as optical coherence tomography (OCT) images, has been a driving force behind AI growth in this field. Age-related macular degeneration (AMD) is one of the leading causes for blindness in the world, affecting approximately 196 million people worldwide in 2020. Multimodal imaging has been for a long time the gold standard for diagnosing patients with AMD, however, currently treatment and follow-up in routine disease management are mainly driven by OCT imaging. AI-based algorithms have by their precision, reproducibility and speed, the potential to reliably quantify biomarkers, predict disease progression and assist treatment decisions in clinical routine as well as academic studies. This review paper aims to provide a summary of the current state of AI in AMD, focusing on its applications, challenges, and prospects.

Surgical Management of Full-Thickness Macular Holes in Macular Telangiectasia Type 2: A Global Multicenter Study

PURPOSE

To report on macular hole repair in macular telangiectasia type 2 (MacTel2).

DESIGN

Global, multicenter, retrospective case series.

PARTICIPANTS

Patients undergoing surgery for MacTel2-associated full-thickness macular hole (MTMH).

METHODS

Standardized data collection sheet distributed to all surgeons.

MAIN OUTCOME MEASURES

Anatomic closure and visual outcomes of MTMH.

RESULTS

Sixty-three surgeries in 47 patients with MTMH were included from 30 surgeons. Mean age was 68.1 years, with 62% female, 72% White, 21% East or South Asian, 2% African American, and 2% Hispanic or Latino. Procedures included 34 internal limiting membrane (ILM) peeling alone, 22 ILM flaps, 5 autologous retinal transplantations (ARTs), 1 retinotomy, and 1 subretinal bleb. For ILM peeling, preoperative visual acuity (VA) was 0.667 ± 0.423 logarithm of the minimum angle of resolution (logMAR). Minimum hole diameter (MHD) was 305.5 ± 159.4 μm (range, 34-573 μm). Sixteen of 34 ILM peels (47%) resulted in MTMH closure. At postoperative month 6, VA was stable at 0.602 ± 0.516 logMAR (P = 0.65). VA improved by at least 2 lines in 43% and at least 4 lines in 24%. For ILM flaps, preoperative VA was 0.878 ± 0.552 logMAR. MHD was 440.8 ± 175.5 μm (range, 97-697 μm), which was significantly larger than for ILM peels (P < 0.01). Twenty of 22 ILM flaps (90%) resulted in MTMH closure, which was significantly higher than for ILM peels (P < 0.01). At postoperative month 6, VA improved to 0.555 ± 0.405 logMAR (P < 0.05). VA improved by at least 2 lines in 56% and at least 4 lines in 28%. For ARTs, preoperative VA was 1.460 ± 0.391 logMAR. MHD was 390.2 ± 203.7 μm (range, 132-687 μm). All 5 ARTs (100%) resulted in MTMH closure. At postoperative month 6, VA was stable at 1.000 ± 0.246 logMAR (P = 0.08). Visual acuity improved at least 2 lines in 25%.

CONCLUSIONS

Surgical closure of macular holes improved VA in 57% of MTMHs. Internal limiting membrane flaps achieved better anatomic and functional outcomes than ILM peeling alone. Autologous retinal transplantation may be an option for refractory MTMHs.

FINANCIAL DISCLOSURE(S)

Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Approved AI-based fluid monitoring to identify morphological and functional treatment outcomes in neovascular age-related macular degeneration in real-world routine (FRB!)

AIM

To predict antivascular endothelial growth factor (VEGF) treatment requirements, visual acuity and morphological outcomes in neovascular age-related macular degeneration (nAMD) using fluid quantification by artificial intelligence (AI) in a real-world cohort.

METHODS

Spectral-domain optical coherence tomography data of 158 treatment-naïve patients with nAMD from the Fight Retinal Blindness! registry in Zurich were processed at baseline, and after initial treatment using intravitreal anti-VEGF to predict subsequent 1-year and 4-year outcomes. Intraretinal and subretinal fluid and pigment epithelial detachment volumes were segmented using a deep learning algorithm (Vienna Fluid Monitor, RetInSight, Vienna, Austria). A predictive machine learning model for future treatment requirements and morphological outcomes was built using the computed set of quantitative features.

RESULTS

Two hundred and two eyes from 158 patients were evaluated. 107 eyes had a lower median (≤7) and 95 eyes had an upper median (≥8) number of injections in the first year, with a mean accuracy of prediction of 0.77 (95% CI 0.71 to 0.83) area under the curve (AUC). Best-corrected visual acuity at baseline was the most relevant predictive factor determining final visual outcomes after 1 year. Over 4 years, half of the eyes had progressed to macular atrophy (MA) with the model being able to distinguish MA from non-MA eyes with a mean AUC of 0.70 (95% CI 0.61 to 0.79). Prediction for subretinal fibrosis reached an AUC of 0.74 (95% CI 0.63 to 0.81).

CONCLUSIONS

The regulatory approved AI-based fluid monitoring allows clinicians to use automated algorithms in prospectively guided patient treatment in AMD. Furthermore, retinal fluid localisation and quantification can predict long-term morphological outcomes.

The use of anterior segment optical coherence tomography (ASOCT) in demonstrating recurrence of vitreoretinal lymphoma (VRL) in the anterior vitreous

Background

Primary vitreoretinal lymphoma (VRL) is a rare disease with 30-380 new cases in the United States per year. Its insidious process and spread to the central nervous system (CNS) leads to a mean 5-year survival rate from 41.4 to 71%. Medical treatment of VRL has been summarized extensively in the literature and involves intraocular rituximab and methotrexate as first line agents in unilateral VRL, with systemic chemotherapy to be considered in bilateral or CNS-involving disease. In addition, therapeutic "debulking" vitrectomy has been reported in the literature, with some limited success. Despite this, recurrence rate is high and should always be suspected in the setting of new inflammation. Anterior segment optical coherence tomography (ASOCT) has not been previously used to image VRL recurrence in the anterior vitreous.

Case presentation

A 63-year-old man, with VRL was found to have cells and debris in the anterior vitreous, 10 months after his first vitrectomy, intravitreal rituximab and methotrexate. Since the patient was phakic at the time of initial vitrectomy, the anterior vitreous had not been removed. ASOCT confirmed the findings. Subsequent surgery removed the lens and debris. Both the patient's vision and ASOCT improved.

Conclusions

We suggest that ASOCT of the anterior segment is a useful diagnostic tool to monitor for recurrence of VRL. In biopsy-proven VRL, phakic patients who undergo therapeutic vitrectomy, should also be considered for lens extraction and anterior vitrectomy to limit recurrences.

Multimodal Imaging and Histopathological Evaluation of Berger's Space

Objective

To demonstrate the multimodal imaging and histopathology of Berger's space.

Methods

We conducted a retrospective in vivo analysis of 4 patients demonstrating Berger's space with intraocular pathological conditions, documented by slit-lamp biomicroscopic photography and, in 2 patients, also by optical coherence tomography (OCT). Additionally, we carried out a retrospective histological study of 7 enucleated eyes with retinoblastoma demonstrating Berger's space. A review of the literature was also performed.

Results

Two eyes had slit-lamp photographs. One case showed Berger's space surrounded by vitreous hemorrhage. In the other case, amyloid was trapped within Berger's space. In another 2 eyes that were pseudophakic, Berger's space was visible on anterior segment OCT. One had amyloid trapped in Berger's space that could be seen with OCT. The histological review of the 7 enucleated eyes with advanced retinoblastoma demonstrated the presence of pyknotic cells in Berger's space.

Conclusions

Berger's space is an actual space in pathological conditions and can be an important site of pathology. Additionally, to our knowledge, this is the first time that Berger's space has been documented by anterior segment OCT in a clinical setting.

DAPI diffusion after intravitreal injection of mesenchymal stem cells in the injured retina of rats

To evaluate DAPI (4',6-diamidino-2-phenylindole) as a nuclear tracer of stem cell migration and incorporation it was observed the pattern of retinal integration and differentiation of mesenchymal stem cells (MSCs) injected into the vitreous cavity of rat eyes with retinal injury. For this purpose adult rat retinas were submitted to laser damage followed by transplantation of DAPI-labeled BM-MSCs grafts and double-labeled DAPI and quantum dot-labeled BM-MSCs. To assess a possible DAPI diffusion as well as the integration and differentiation of DAPI-labeled BM-MSCs in laser-injured retina, host retinas were evaluated 8 weeks after injury/transplantation. It was demonstrated that, 8 weeks after the transplant, most of the retinal cells in all neural retinal presented nuclear DAPI labeling, specifically in the outer nuclear layer (ONL), inner nuclear layer (INL), and ganglion cell layer (GCL). Meanwhile, at this point, most of the double-labeled BM-MSCs (DAPI and quantum dot) remained in the vitreous cavity and no retinal cells presented the quantum dot marker. Based on these evidences we concluded that DAPI diffused to adjacent retinal cells while the nanocrystals remained labeling only the transplanted BM-MSCs. Therefore, DAPI is not a useful marker for stem cells in vivo tracing experiments because the DAPI released from dying cells in moment of the transplant are taken up by host cells in the tissue.

Pattern reversal visual evoked response in retinitis pigmentosa

PURPOSE

To determine if the pattern reversal visual evoked response (PVER) can objectively evaluate visual function in retinitis pigmentosa (RP).

METHODS

We retrospectively reviewed the medical records of 29 patients (29 eyes) with RP, a visual acuity (VA) better than 20/100, and for whom an electroretinogram (ERG) and a PVER had been recorded. A steady-state PVER was measured using five check sizes: 160, 80, 40, 20 and 10 min of arc. The best-corrected VA was measured using the ETDRS chart. Visual fields (VFs) were measured using standard Goldmann perimetry. To quantify the VFs, we measured the field size (isopter IV-2) in four meridians and averaged them. Twenty-five eyes of 25 normal subjects served as controls.

RESULTS

The mean and median VAs were 20/40 (logMAR = 0.30 +/- 0.20). The average VF (isopter IV-2 ) was 20 degrees (median, 9 degrees: range. 2 degrees to 62 degrees). All patients had a recordable PVER. Only 12 (41 %) patients had a recordable ERGs, all subnormal. The PVER amplitudes showed a relatively good correlation with VA in the intermediate check sizes (40 min of arc, r = 0.611. P = 0.0004: 20 min of arc, r = 0.596. P = 0.0007). The PVER amplitude-check size function had a flattened inverted V shape in the RP patients and an inverted U shape in the normal subjects, with the mean amplitudes for the RP patients significantly smaller than the normals at all check sizes.

CONCLUSION

The PVER is a useful method for objectively evaluating the visual function in RP patients whose central vision is still preserved.