Implementing and evaluating a fully functional AI-enabled model for chronic eye disease screening in a real clinical environment.

BACKGROUND: Artificial intelligence (AI) has the potential to increase the affordability and accessibility of eye disease screening, especially with the recent approval of AI-based diabetic retinopathy (DR) screening programs in several countries. METHODS: This study investigated the performance, feasibility, and user experience of a seamless hardware and software solution for screening chronic eye diseases in a real-world clinical environment in Germany. The solution integrated AI grading for DR, age-related macular degeneration (AMD), and glaucoma, along with specialist auditing and patient referral decision. The study comprised several components: (1) evaluating the entire system solution from recruitment to eye image capture and AI grading for DR, AMD, and glaucoma; (2) comparing specialist's grading results with AI grading results; (3) gathering user feedback on the solution. RESULTS: A total of 231 patients were recruited, and their consent forms were obtained. The sensitivity, specificity, and area under the curve for DR grading were 100.00%, 80.10%, and 90.00%, respectively. For AMD grading, the values were 90.91%, 78.79%, and 85.00%, and for glaucoma grading, the values were 93.26%, 76.76%, and 85.00%. The analysis of all false positive cases across the three diseases and their comparison with the final referral decisions revealed that only 17 patients were falsely referred among the 231 patients. The efficacy analysis of the system demonstrated the effectiveness of the AI grading process in the study's testing environment. Clinical staff involved in using the system provided positive feedback on the disease screening process, particularly praising the seamless workflow from patient registration to image transmission and obtaining the final result. Results from a questionnaire completed by 12 participants indicated that most found the system easy, quick, and highly satisfactory. The study also revealed room for improvement in the AMD model, suggesting the need to enhance its training data. Furthermore, the performance of the glaucoma model grading could be improved by incorporating additional measures such as intraocular pressure. CONCLUSIONS: The implementation of the AI-based approach for screening three chronic eye diseases proved effective in real-world settings, earning positive feedback on the usability of the integrated platform from both the screening staff and auditors. The auditing function has proven valuable for obtaining efficient second opinions from experts, pointing to its potential for enhancing remote screening capabilities. TRIAL REGISTRATION: Institutional Review Board of the Hamburg Medical Chamber (Ethik-Kommission der Ärztekammer Hamburg): 2021-10574-BO-ff.

Diurnal functional and anatomical changes in X-linked retinoschisis.

BACKGROUND: To investigate the changes in macular cystic schisis (MCS) and sensitivity during the day in X-linked retinoschisis (XLRS) patients. METHODS: Treatment-naïve patients with genetically verified XLRS underwent best-correlated visual acuity (BCVA) testing with ETDRS charts, spectral domain optical coherence tomography, and microperimetry (MP) twice a day, at 9 a.m. and 4 p.m., to measure changes in central retinal thickness (CRT), macular volume (MV), average threshold (AT), and fixation stability parameters (P1 and P2). RESULTS: At baseline, the BCVA of the 14 eyes of 8 patients amounted 0.73 (± 0.23) LogMAR. Between timepoints, the BCVA increased in 3.21 letters (p = .021), the AV improved in 1.84 dB (p = .03, 9.73%), the CRT decreased in 24.43 µm (p = .007, - 4.05%), and the MV dropped in 0.27 µm3 (p = .016, - 2.68%). P1 and P2 did not variate. The collapse of the MCS led to the reduction of macula thickness. CRT at baseline correlated with the decrease of CRT (Spearman's ρ: - 0.83 [p = .001]). Age and change of BCVA, CRT, and AV did not correlate among one another. Eyes with disrupted ellipsoid zone showed a more prominent change in CRT (p = .050). Photoreceptor outer segment length and integrity of the external limiting membrane and cone outer segment tips were not associated with BCVA, AT, or CRT variation. CONCLUSION: Eyes of treatment-naïve XLRS patients show diurnal macular thickness and function changes. Eyes with pronounced macular thickness show a greater reduction of the MCS. These results should be taken into consideration in upcoming clinical trials in XLRS. TRIAL REGISTRATION NUMBER: Institutional Review Board of the Hamburg Medical Chamber (Ethik-Kommission der Ärztekammer Hamburg): 2020-10,328.

X-Linked Retinoschisis Masquerading Uveitis.

X-linked retinoschisis (XLRS) shows features also seen in patients with uveitis and is recognized as an uveitis masquerade syndrome. This retrospective study aimed to describe characteristics of XLRS patients with an initial uveitis diagnosis and to contrast these to patients with an initial XLRS diagnosis. Patients referred to a uveitis clinic, which turned out to have XLRS (n = 4), and patients referred to a clinic for inherited retinal diseases (n = 18) were included. All patients underwent comprehensive ophthalmic examinations, including retinal imaging with fundus photography, ultra-widefield fundus imaging, and optical coherence tomography (OCT). In patients with an initial diagnosis of uveitis, a macular cystoid schisis was always interpreted as an inflammatory macular edema; vitreous hemorrhages were commonly interpreted as intraocular inflammation. Patients with an initial diagnosis of XLRS rarely (2/18; p = 0.02) showed vitreous hemorrhages. No additional demographic, anamnestic, and anatomical differences were found. An increased awareness of XLRS as a uveitis masquerade syndrome may facilitate early diagnosis and may prevent unnecessary therapies.

PORES OF THE INTERNAL LIMITING MEMBRANE: A Common Finding in Vitreomaculopathies.

PURPOSE: To describe presence and distribution of pores of the inner limiting membrane (ILM) in eyes with vitreomaculopathies. METHODS: Inner limiting membrane specimens were harvested from 117 eyes of 117 patients during vitrectomy with membrane peeling from eyes with vitreomacular traction syndrome, idiopathic and secondary epiretinal gliosis, and idiopathic full-thickness macular hole. All specimens were processed as flat-mounts for immunocytochemistry and examined by phase-contrast, interference, and fluorescence microscopy. Demographic and clinical data were correlated. RESULTS: Inner limiting membrane pores were found in all vitreomaculopathies. They were identified in 47 (40.2%) of 117 eyes being most evident with antilaminin. In eyes with full-thickness macular hole >400 µ m, pores were seen in more than half of all eyes. They occur as numerous and uniformly distributed defects of the flat-mounted ILM with a mean diameter of 9.5 ± 2.4 µ m. Edges of ILM pores are round with an irregular contour and no specific cellular pattern. Pores were distinguished from retinal vessel thinning and iatrogenic artefacts. CONCLUSION: Contrary to previous reports, ILM pores are a common finding in vitreomaculopathies easily visible with antilaminin staining. Further studies are needed to clarify whether their presence correlates with differences in disease progression or imaging before and after vitrectomy with ILM peeling.

[Immunohistochemical and Molecular Genetic Profile of Mantle Cell Lymphoma of the Lacrimal Gland: A Case Series of an Uncommon Tumour of the Lacrimal Gland]. / Immunhistochemisches und molekulargenetisches Profil von Mantelzelllymphomen der Glandula lacrimalis: eine detaillierte Beschreibung einer seltenen Tumorentität der Tränendrüse.

BACKGROUND: Mantle cell lymphomas (MCL) represent a rare subclass of Non-Hodgkin Lymphoma affecting the lacrimal gland (GL). AIM: To extensively describe the immunohistochemical profile of GL-MCL. MATERIAL UND METHODS: Single center, retrospective electronic records review of 3 patients with biopsy proven LG-MCL. RESULTS: The herein presented case series of three patients comprises a focal case involving solely the lacrimal gland, a symptomatic LG-MCL manifesting as the first sign of a systemic disease as well as a case of LG-MCL presenting as a relapsed systemic lymphoma. The three patients presented positive CD19 and CD20, negative CD10 and CD23. One patient showed an uncommon negativity for CD5. The increased expression of cyclin D1 caused by the classical translocation t(11;14) (q13;q32) in the fluorescence-in-situ-hybridisation were observed in all cases. B-cell-lymphoma-2 protein (BCL-2) and transcription factor SOX-11 (SOX-11) were also overexpressed. DISCUSSION: LG-MCL show an immunohistochemical profile corresponding to the classical profile of MCL. Overexpression of molecules for target therapies was found in all cases (CD20 for rituximab, BCL2 for Bruton-kinase-inhibitors and CD19 for CAR-T cell therapy). The removal of the GL can potentially drive to severe complications, even if aimed to confirm diagnosis. Therefore, the choice between GL-biopsy and exstirpation should be carefully evaluated, especially in cases of suspected lymphoma.

Morphological and functional parameters in X-linked retinoschisis patients-A multicentre retrospective cohort study.

Introduction: X-linked retinoschisis (XLRS) is a potential target for gene supplementation approaches. To establish potential structural and functional endpoints for clinical trials, a comprehensive understanding of the inter-eye symmetry, relationship between structural and functional parameters, and disease progression is vital. Methods: In this retrospective multicentre study, 118 eyes of 59 XLRS patients with RS1 mutations were assessed. Information from center databases included: RS1 variant; age at presentation; best-corrected visual acuity (BCVA), central retinal thickness (CRT), macular volume (MV) at presentation and at the last follow up; full-field electroretinogram (ERG) findings; presence of peripheral retinoschisis and complications (vitreous hemorrhage, retinal detachment); treatment with systemic or topical carbonic anhydrase inhibitors (CAI). Results: Inter-eye symmetry revealed strong correlation in CRT (r = 0.77; p < 0.0001) and moderate correlations in MV (r = 0.51, p < 0.0001) and BCVA (r = 0.49; p < 0.0001). Weak or no correlations were observed between BCVA and structural parameters (CRT, MV). Peripheral retinoschisis was observed in 40 (68%), retinal detachment in 9 (15%), and vitreous hemorrhage in 5 (8%) patients, respectively. Longitudinal examinations (mean, 4.3 years) showed no BCVA changes; however, a reduction of the CRT (p = 0.02), and MV (p = 0.01) was observed. Oral and/or topical CAI treatment did not significantly alter the CRT (p = 0.34). Discussion: The XLRS phenotype demonstrates a strong CRT symmetry between the eyes within individual patients and stable BCVA over several years. BCVA exhibits a weak correlation with the morphological parameters of retinal thickness (CRT MV). In our cohort, longitudinal functional changes were not significant, likely attributed to the short average follow-up period. Furthermore, CAI treatment didn't influence both morphological and functional outcomes.

Predicting speed of progression of lens opacification after pars plana vitrectomy with silicone oil.

PURPOSE: An increasing number of posterior segment disorders is routinely managed with pars plana vitrectomy (PPV). In older, phakic patients cataract formation is expected within the first two years after surgery. For younger patients its progression is individually fluctuating. This study uses an objective quantitative measurement for lens-status-monitoring after PPV with silicone oil to derive predictions for progression and severity of post-operative lens opacification evaluated in patients with rhegmatogenous retinal detachment (RRD). METHODS: Data acquisition was performed prospectively between March 2018 and March 2021. PentacamHR® Nucleus Staging mode (PNS) was used to objectively gather data about nuclear cataracts after PPV at different time points. Data was grouped into training and test sets for a mathematical prediction model. Via backward variable selection method a mathematical formula was set up by means of which predictions about lens densitometry (LD) can be calculated. RESULTS: 20 males [58.8%] and 14 females [41.2%] matched the inclusion criteria (mean age 50.6 years [23-75; ±12.3]). Average follow-up was 8.1 months (3,4-17.4; ±3.4). Mean baseline LD of the treated and fellow eye before surgery was 11.1% (7.7%-17.6%; ±2.0) and 11.2% (7.7%-14.8%; ±1.5), respectively. Predicted LD values by the model for five pre-selected patients closely match the observed data with an average deviation of 1.06%. CONCLUSIONS: Using an objective parameter like LD delivered by the PentacamHR® PNS mode additionally to the patient's age allows us to make an individual prediction for any time after PPV with silicone oil due to RRD for all ages. The accuracy of the model was stronger influenced by baseline LD as cofactor in the equation than patient's age. The application for the prediction lens opacification [which can be accessed for free under the following link (https://statisticarium.com/apps/sample-apps/LensDensityOil/)] can help vitreoretinal surgeons for patient consultation on the possibility to combine PPV with cataract surgery.

Retinal Tacks for Complicated Retinal Detachment: Retinal Tacks in the Times of Modern Small-Gauge Vitrectomy.

Purpose: To investigate the efficacy and safety profile of retinal tacks (RTs) in cases of retinal detachment (RD) with advanced proliferative vitreoretinopathy (PVR). Materials and Methods: In this single-center, retrospective study medical record, optical coherence tomography and ultra-widefield fundus images of patients with complex PVR-related and RT surgery were reviewed. All cases underwent 23G pars plana vitrectomy (PPV), RT implantation, retinectomy, circumferential intraoperative laser retinopexy, and silicone oil tamponade. Results: Fourteen eyes of 14 patients with complex rhegmatogenous RD with PVR were included: 7 cases showed PVR grade C type P and 7 combined grades A and P. RTs were positioned at contracted, stiffened retinal areas to achieve attachment of retinectomy borders after extensive PVR peeling. Patients underwent on an average of 1.3 PPVs (range 0-3) prior RT surgery. An average of 2.5 RTs (range 1-4) were implanted. Only in a single eye, a recurrent RD occurred. In 10 eyes, the silicone oil tamponade was still in place at the last follow-up. In 5 eyes, the silicone oil could be removed without redetachment in all of these cases (average of 31.3 weeks, range 11.4-53). No RT-related intraoperative or postoperative complications like dislocation or bleedings were observed. Conclusion: RTs have the potential to improve the treatment of complex PVR-associated RD. RT can be a useful surgical tool to reattach borders of retinectomies in advanced PVR. No RT-associated complication were observed in this study.

Lens densitometry for assessment and prediction of cataract progression after pars plana vitrectomy with C3F8-gas for retinal detachment.

PURPOSE: Lens opacification is a common complication after pars plana vitrectomy (PPV) and knowing its progression would facilitate consulting patients. The purpose of this study was to evaluate a quantitative model for lens-status-monitoring after PPV with C3F8 gas. Our model was evaluated in rhegmatogenous retinal detachment (RRD) patients of various age and lens densitometry (LD). METHODS: Data between March 2018 and March 2020 were evaluated retrospectively. LD measurements of the PentacamHR® Nucleus Staging mode (PNS) were used to quantify lens opacification over time. A mixed-effect regression model was designed, to enable LD predictions at any time postoperatively. Calculations were based on patient's age and baseline LD as dependent variables. Six patients were randomly excluded during model development, to be used for testing its power afterwards. RESULTS: 34 patients (male 19 [55.9%], female 15 [44.1%]) matched the inclusion criteria. Average age was 58.5 years (32-77;±4.3) and average follow-up was 7.2 months (3,4-23.1;±1,8). Mean baseline LD of the treated and fellow eye before surgery were 10.9% (8.7%-14.8%;±0.8) and 10.7% (8.5%-14.1%;±0.6), respectively. Using our prediction model, LD values for the six pre-selected patients closely match the observed data with an average deviation of 1.07%. CONCLUSIONS: Evaluation of age and baseline LD using a mixed-effect regression model might predict cataract progression in RRD patients treated with PPV and C3F8-gas. Such a tool could be considered during cataract surgery consultation in these patients.

Premacular Cells as Source of Neurotrophic Factors in Idiopathic Macular Holes.

Purpose: To describe the presence of neurotrophic growth factors and histopathologic characteristics of internal limiting membrane (ILM) specimens obtained from large idiopathic full-thickness macular holes (FTMH). Methods: In 24 eyes of 24 patients with FTMH of diameter >400 µm, ILM specimens were harvested directly at the edge surrounding the macular hole during vitrectomy with peeling. We performed interference and phase contrast microscopy of flat mounts followed by immunostaining and transmission electron microscopy. Primary antigens directed against neurotrophic growth factors as well as antigens to glial and ganglion cells were used. Topographic relationship of cells and collagen was demonstrated by serial ultrathin sectioning. Results: Immunofluorescence microscopy demonstrated the presence of glial-derived neurotrophic factor and ciliary neurotrophic factor. Expression of vimentin, glial fibrillary acidic protein (GFAP), neurofilament, calretinin, and melanopsin was seen positive too. Cellular retinaldehyde-binding protein was seen positive in half of the specimens. Co-localisation of anti-GFAP as well as anti-vimentin with neurotrophic factors was found. Electron microscopy revealed cells exclusively on the vitreal side of the ILM. Cell fragments on the retinal side were rarely seen. Conclusion: In large FTMH, ILM specimens present positive immunolabelling of neurotrophic factors. The co-localization with macroglial cell markers suggests a premacular cell composition as a source of the neurotrophic factors. Ultrastructurally, premacular cells were found on the vitreal side of the ILM and not within the collagen network of the ILM itself.