Automatic inference of ICD-10 codes from German ophthalmologic physicians' letters using natural language processing.

Physicians' letters are the optimal source of diagnoses for registries. However, most registries demand for diagnosis codes such as ICD-10. We herein describe an algorithm that infers ICD-10 codes from German ophthalmologic physicians' letters. We assess the method in three German eye hospitals. Our algorithm is based on the nearest-neighbor method as well as on a large thesaurus for ICD-10 codes. This thesaurus was embedded into a Word2Vec space created from anonymized physicians' reports of the first hospital. For evaluation, each of the three hospitals sent all diagnoses taken from 100 letters. The inferred ICD-10 codes were evaluated for correctness by the senders. A total of 3332 natural language terms had been sent in (812 hospital one, 1473 hospital two, 1047 hospital three). A total of 526 non-diagnoses were excluded upfront. 2806 ICD-10 codes were inferred (771 hospital one, 1226 hospital two, 809 hospital three). In the first hospital, 98% were fully correct and 99% correct at the level of the superordinate disease concept. The percentages in hospital two were 69% and 86%. The respective numbers for hospital three were 69% and 91%. Our simple method is capable of inferring ICD-10 codes for German natural language diagnoses, especially when the embedding space has been built with physicians' letters from the same hospital. The method may yield sufficient accuracy for many tasks in the multi-centric setting and can easily be adapted to other languages/specialities.

Light conditions influence optic nerve OCT angiography parameter in healthy subjects with neutral pupils.

Optical coherence tomography angiography measurements are influenced by a range of environmental factors as blood pressure and physical fitness. The present study aimed to evaluate the effects of light and dark exposure in eyes with neutral and mydriatic pupils on vessel density in the macular and optic nerve head regions, as measured using optical coherence tomography angiography (OCTA). 55 eyes of 55 healthy volunteers (28 patients with neutral pupils; 27.18 ± 4.33 years) were examined using a high-speed and high-resolution spectral-domain OCT XR Avanti system with a split-spectrum amplitude de-correlation angiography algorithm. OCTA imaging was performed after dark adaptation and after exposure to light. The vessel density data of the superficial and deep retinal macular and optic nerve head region OCT-angiogram were analyzed for these two light conditions. Through Bonferroni correction for multiple testing, the p- value was adapted from 0.05 to 0.017. In eyes with neutral pupils, a significant increase was found in the capillary region of the optic nerve head region (p = 0.002), comparing dark- and light-adaptation. In the macular region of eyes with neutral (p = 0.718) and mydriatic pupils (p = 0.043), no significant differences were observed, as were any in the optic nerve head region of the mydriatic eyes (p = 0.797). This observation suggests that light conditions could be a possible factor influencing OCTA measurements. After dark exposure, vessel density data were significantly different between eyes with neutral and mydriatic pupils (nerve head region: p < 0.0001, superficial macula: p < 0.0001, deep macula: p = 0.0025). These data warn for the effect of mydriatic drops on vessel density measurements.

[Current situation of electronic medical records in ophthalmological departments in Germany]. / Status der elektronischen Dokumentation in ophthalmologischen Hauptabteilungen in Deutschland.

BACKGROUND: Electronic medical records are required in ophthalmology clinics to be integrated into digital care networks and efficient clinical registry databases. AIM OF THE WORK: To assess the prevalence and methods of electronic medical recordkeeping in inpatient ophthalmological care in Germany. MATERIAL AND METHODS: An online questionnaire was sent to all German university eye hospitals and ophthalmology departments in June 2021. It included 13 open and closed option questions concerning current practices of digital recordkeeping, including the structure of data storage and the recording of billing-relevant codes in the departments. RESULTS: A total of 44 (44%) out of 100 clinics responded. Patient documentation was completely digital in 15 (34%) clinics and partly digital and paper-based in the remaining 29 (66%). A total of 16 different constellations of documentation programs were specified. The most frequently used programs were Orbis (27%) (Dedalus HealthCare, Bonn, Germany), FIDUS (18%) (Arztservice Wente, Darmstadt, Germany), and SAP/i.s.h.med (16%) (SAP Deutschland, Walldorf, Germany; Cerner Deutschland, Berlin, Germany) and 3 clinics indicated primary use of paper records. Structured documentation of findings was performed in 61% of the departments, while 23% used a semistructured manner and 15% used a nonstructured format. Electronic documents are stored as DICOM (Digital Imaging and Communications in Medicine) documents 20% of the clinics and as PDF (Portable Document Format) files in 34% of the clinics while 23% store scanned printouts. DISCUSSION: Methods of medical record keeping in German eye clinics are heterogeneous, with paper-based documentation continuing to play an important role. This, as well as the high number of different electronic medical record software pose important challenges in terms of interoperability and secondary use of clinical data.

[Diagnostics of diseases of the optic nerve head in times of artificial intelligence and big data]. / Diagnostik von Erkrankungen des Sehnervenkopfes in Zeiten von künstlicher Intelligenz und Big Data.

BACKGROUND: The use of artificial intelligence (AI) interesting for automated image segmentation, analysis and classification, among others and has already been described for various fields of ophthalmology. OBJECTIVE: This manuscript provides an overview of current approaches and advances in the application of big data and AI in various diseases of the optic nerve head. MATERIAL AND METHODS: A PubMed search was performed. Studies were searched for that answered clinical questions using big data approaches or classical machine learning methods in the analysis of multimodal imaging of the optic nerve head. RESULTS: Big data can help to answer clinical questions in common diseases such as glaucoma. The AI is applied for the segmentation of multimodal imaging of the optic nerve head as well as for the classification of diseases, such as glaucoma or optic disc edema on this imaging data. CONCLUSION: With the help of big data and AI, relationships can be recognized more easily and the diagnostics and course assessment of diseases of the optic nerve head can be facilitated or automated. A prerequisite for clinical application is a CE marking as a medical device in Europe and approval by the Food and Drug Administration in the USA.

["Flipped classroom"-A future concept for student teaching in ophthalmology?] / "Flipped classroom" ­ Ein Zukunftskonzept für die studentische Lehre in der Augenheilkunde?

BACKGROUND: "Flipped classroom" is a didactic teaching concept in which learning contents are prepared by self-study with arranged tools before the classroom session. The concept offers the advantage of a uniform knowledge base for the students at the beginning of the course and also the advantage of a greater theoretical knowledge, which creates more opportunities for practical exercises, application and consolidation in the subsequent joint teaching units. This study describes the establishment and application of such a model in student teaching in ophthalmology and analyzes the student's evaluation. METHODOLOGY: For the winter term 2018/2019, a new teaching module was designed and established in a cooperation between the department of ophthalmology and the Institute for Education and Study Affairs (IfAS) at the medical faculty of the University of Münster. A uniform training of the lecturers as well as a preparation of the students for the restructuring took place. After the course the evaluation of the students was recorded and evaluated using a standardized online evaluation. RESULTS: Between the winter semester 2018/2019 and the winter semester 2019/2020, an average of 112.3 ± 4.0 students were taught with the "flipped classroom" model. Of these 93.7% were able to give an assessment. In the previous semesters with the old teaching concept (summer semester 2015 to summer semester 2018), the average number of students was 115.4 ± 15.1 with an assessment rate of 93.3%. The new teaching concept achieved on average a better assessment than the old module. CONCLUSION: With a "flipped classroom" space and flexibility can be generated for a more individual course preparation and at the same time a higher practical part. Further studies are needed to analyze whether this also enables a sustainable transfer of knowledge.

[Artificial intelligence in management of macular edema : Opportunities and challenges]. / Künstliche Intelligenz zum Management von Makulaödemen : Chancen und Herausforderungen.

Macular edema occurs in a wide variety of ophthalmological diseases. The diagnostics and treatment are an important part of modern ophthalmology. Due to the continuous development, artificial intelligence (AI) offers many opportunities to improve the management of macular edema. This article provides the readership with an overview of this interesting topic.

[Analysis of the importance of e-learning in ophthalmology and evaluation of an e-learning app]. / Analyse des Stellenwertes von "eLearning" in der Augenheilkunde und Evaluierung einer "eLearning-App".

BACKGROUND: The aim of the study was to analyze the importance of e­learning in the learning and training behavior of ophthalmologists in Germany and to evaluate the acceptance of a new e­learning user software (app). MATERIAL AND METHODS: Ophthalmological residents and specialists were asked about continuing education activities by means of a questionnaire during continuing education events. Furthermore, a structured evaluation was carried out after the presentation and application of an e­learning app. RESULTS: A total of 149 ophthalmologists took part in the survey. While the majority of colleagues (74.3%) used specialist journals weekly or monthly for further education, 45.9% of ophthalmologists used digital print media (books, journals, articles) and 46.5% used specialist books in printed form. Only 35% of the interviewees used online training platforms, e.g. digital courses (CME courses) or portals for retrieving recorded lectures. The use of the offered e­learning app was generally accompanied by a positive acceptance. Of the interviewed colleagues 91.7% would recommend this form of interactive learning. DISCUSSION: Despite a progressive digitalization in all areas of life, e­learning continues to play a minor role as a learning medium in ophthalmological advanced training. Interestingly, the evaluation of app users showed a high level of acceptance, regardless of age or field of work.

Prospective PED-study of intravitreal aflibercept for refractory vascularized pigment epithelium detachment due to age-related macular degeneration: morphologic characteristics of non-responders in optical coherence tomography.

PURPOSE: The aim of this study was to investigate the outcomes of a fixed intravitreal aflibercept regimen in patients with vascular pigment epithelium detachment (vPED) secondary to age-related macular degeneration with refractory subretinal fluid. METHODS: A prospective, interventional case series involved 20 eyes of 20 patients with refractory subretinal fluid and vPED treated with at least three injections of intravitreal anti-VEGF prior to study inclusion. After study inclusion, patients were treated with three injections of intravitreal aflibercept 2 mg/0.05 mL monthly followed by injections every 8 weeks. Best-corrected visual acuity (BCVA) and spectral-domain optical coherence tomography (SD-OCT) were evaluated at all visits. Fluorescein angiography and indocyanine green angiography were performed at baseline and quarterly. Primary outcomes were effectivity of a fixed treatment as measured in change in BCVA, PED greatest linear diameter (GLD), and PED height from baseline to month 12. In an additional post hoc analysis, vPED patients were differentiated into two groups: (1) vPED lesions that showed persistence of subretinal fluid throughout 1 year of treatment and (2) vPED lesions that showed complete resolution of subretinal fluid at least at one of the monthly performed OCT volume scans. Reflectivity values were determined in the subretinal pigment epithelium (RPE) compartment in OCT scans at baseline, month 6 and 12. RESULTS: A total of 18 patients completed the study protocol. The mean age was 74.8 ± 10.6 years, and six patients were female. The median BCVA of all patients was 72.0 ± 8.0 EDTRS letters at baseline and 72.5 ± 9.5 EDTRS letters at 12-month follow-up (p = 0.7420). The median PED height in all patients as measured in the OCT images significantly decreased from 372.0 ± 140.0 µm to 149.0 ± 142.0 µm after 12 months of treatment (p = 0.0020). Persistent subretinal fluid was present at every OCT control in six patients (group 1). Twelve patients showed resolution of subretinal fluid at least at one OCT control (group 2). Reflectivity values in the sub-RPE compartment in OCT scans were 41.48 ± 4.48 (group 1) and 42.62 ± 12.34 (group 2) at baseline (p = 0.854) and 65.88 ± 6.74 and 50.87 ± 14.11 at month 12 (p = 0.038). CONCLUSIONS: Intravitreal aflibercept in refractory vPED leads to a significant reduction in PED height and disease activity as well as preservation of BCVA over 1 year. Persistent subretinal fluid was present in PED lesions with high values of reflectivity under the RPE, suggesting both a diffusion barrier and an increasing fibrovascular maturization of the choroidal neovascularization. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT03370380.

[Status and perspectives of registry-based studies in German ophthalmology]. / Stand und Perspektiven registerbasierter Studien in der deutschen Augenheilkunde.

BACKGROUND AND OBJECTIVE: Registry studies provide insights into real-life diagnostic and treatment data outside of clinical trials. Registry studies are interesting for quality management and explorative analyses can lead to the identification of possible risk or prognostically relevant factors and generate hypotheses. There are currently relatively few active registry studies in German ophthalmology. The aim of this survey was to collate the different infrastructure and in particular the potential hurdles in the establishment and performance of registry studies in German ophthalmology departments. METHODS: An online questionnaire collected data on participation in registry studies in German ophthalmology departments between September and December 2018. The survey was addressed to all hospital management and medical staff involved in registry studies in German ophthalmology. RESULTS: Out of 45 participants 18 were head of the department, the remaining 27 were consultants (15), medical specialists (2) and residents (10). According to the department head an average of 2.5 (2.0-3.5) employees per clinic participate in 2.0 (1.7-3.0) registry studies. The amount of reimbursement recommended by the hospital management differed significantly from the amount suggested by the staff (0.0 (0; 75) € (0-100) vs. 100.0 (50.0; 150.0) €; p = 0.0012). The most frequent hurdles to conducting a registry study were bureaucracy and limited human resources. Half of the surveyed centers profited from a separate study center and good clinical practice (GCP) courses for employees involved in studies were mandatory. A quarter of these centers received support from an IT department and/or a statistician. CONCLUSION: Registry studies are an important instrument in ophthalmology research and their importance is increasing in Germany. An agreement on national standards would make the establishment of further registry studies easier.

[FIDUSweb version 2.0 : An electronic platform to establish a cooperative university eye network]. / FIDUSweb in der Version 2.0 : Möglichkeiten einer elektronischen Plattform zur Etablierung eines kooperativen, universitären Augennetzwerks.

BACKGROUND: Due to the increasing digitalization in ophthalmology, increasingly more documentation and diagnostic data are becoming electronically available. In order to improve patient care beyond sectoral boundaries, electronic cooperation portals were introduced in Münster and Homburg, which serve the networking with ophthalmologists in private practice. METHOD: This article reports on the implementation of an electronic cooperation portal at the University Eye Hospitals in Münster and Homburg in 2015 and 2016. RESULTS: The FIDUSweb represents a data protection-compliant cooperation portal that can be easily set up. By implementing it as a web portal all terminal equipment and also electronic patient files can be networked. Particularly important functions are the transmission of medical reports in real time, appointment booking functions, the transmission of imaging data and shared treatment documentation. CONCLUSION: Cross-sectoral care can greatly benefit from the use of the FIDUSweb, especially in digital and multimodal ophthalmology.

Impact of integrated multiple image averaging on OCT angiography image quality and quantitative parameters.

PURPOSE: Multiple image averaging (MIA) is a new approach to improve OCT angiography (OCTA) imaging. The aim of this work was to analyze the impact of MIA on image quality and quantitative OCTA parameters. METHODS: Twenty eyes from 20 healthy volunteers (55.65 ± 14.8 years) were prospectively enrolled. Imaging was performed using two commercially available OCTA devices (Canon OCT HS-100, Optovue AngioVue) using a uniform imaging protocol. Each participant had two single scans of the macula (3 × 3mm, Canon and Optovue) as well as five continuous single scan imaging procedures (3 × 3mm each) using the Canon device. Three out of five of these images with highest quality were manually chosen and then automatically processed by the Canon device using MIA. The superficial retinal plexus of the single scans and of MIA images was analyzed with regard to the device' own image quality scores (IQS), peak signal-to-noise ratio (PSNR), the size of the foveolar avascular zone (FAZ), and vessel density (VD). Image acquisition times were recorded. Parameters were compared between the devices and the different imaging protocols. RESULTS: Average acquisition time was significantly higher for the MIA compared with the single measurements (29.09 ± 10.19 seconds (s) (MIA) vs. 5.56 ± 2.17 s (Canon single scan) vs. 20.28 ± 6.81 s (Optovue) (p < 0.001)). IQS showed no significant differences between the devices and between the recording protocols. PSNR was 12.38 ± 0.20 (Canon single scan), 13.01 ± 0.36 (Canon MIA), and 14.34 ± 0.60 (Optovue) (p < 0.001 between the groups). Mean FAZ area in Canon single scans was 0.29 ± 0.06 mm2, 0.27 ± 0.07 mm2 using MIA, and 0.27 ± 0.08 mm2 using the Optovue device. There was no significant difference between mean FAZ measurements before and after averaging (Canon single scan vs. MIA, p = 0.168). VD of the parafoveal area using MIA was significantly lower compared with both single scans (p < 0.001). CONCLUSIONS: MIA can improve PSNR, but it also reduces imaging speed and significantly affects VD measurements. Therefore, when comparing OCTA data, the use of uniform imaging protocols is required.

Automated OCT angiography image quality assessment using a deep learning algorithm.

PURPOSE: To expedite and to standardize the process of image quality assessment in optical coherence tomography angiography (OCTA) using a specialized deep learning algorithm (DLA). METHODS: Two hundred randomly chosen en-face macular OCTA images of the central 3 × 3 mm2 superficial vascular plexus were evaluated retrospectively by an OCTA experienced reader. Images were defined either as sufficient (group 1, n = 100) or insufficient image quality (group 2, n = 100) based on Motion Artifact Score (MAS) and Segmentation Accuracy Score (SAS). Subsequently, a pre-trained multi-layer deep convolutional neural network (DCNN) was trained and validated with 160 of these en-face OCTA scans (group 1: 80; group 2: 80). Training accuracy, validation accuracy, and cross-entropy were computed. The DLA was tested in detecting 40 untrained OCTA images (group 1: 20; group 2: 20). An insufficient image quality probability score (IPS) and a sufficient image quality probability score (SPS) were calculated. RESULTS: Training accuracy was 97%, validation accuracy 100%, and cross entropy 0.12. A total of 90% (18/20) of the OCTA images with insufficient image quality and 90% (18/20) with sufficient image quality were correctly classified by the DLA. Mean IPS was 0.88 ± 0.21, and mean SPS was 0.84 ± 0.19. Discrimination between both groups was highly significant (p < 0.001). Sensitivity of the DLA was 90.0%, specificity 90.0%, and accuracy 90.0%. Coefficients of variation were 0.96 ± 1.9% (insufficient quality) and 1.14 ± 1.6% (sufficient quality). CONCLUSIONS: Deep learning (DL) appears to be a potential approach to automatically distinguish between sufficient and insufficient OCTA image quality. DL may contribute to establish image quality standards in this recent imaging modality.

[Analysis of real-life data on treatment quality in patients with exudative age-related macular degeneration (AMD) and vein occlusion at a German university eye hospital]. / Real-Life-Daten-Analyse der Therapiequalität bei Patienten mit exsudativer altersabhängiger Makuladegeneration (AMD) und venösen Gefäßverschlüssen an einer deutschen Universitätsaugenklinik.

PURPOSE: Real-life data provide an insight into the reality of patient care under everyday conditions. Digitization in ophthalmology has led to a structured documentation of patient data. This makes it possible to automate the retrieval of a huge amount of real-life data by developing suitable query algorithms. METHODS: Using an automated query from fully electronic patient documentation, real-life therapy data were obtained in patients with intravitreal operative drug delivery (IVOM) in age-related macular degeneration (AMD) and vein occlusion (VO). Among other things, injection frequency, frequency of visits and best-corrected visual acuity over a total of 4 years of treatment were recorded. The influence of factors, such as patient age, distance to the clinic and type of insurance on the quality of care were also investigated. Treatment-naïve and pretreated patients were analyzed. RESULTS: In the first year AMD patients received an average of 4.59 ± 2.00 injections with 10.29 ± 4.47 visits, in the second year 3.83 ± 2.23 injections with 7.72 ± 5.78 visits, in the third year 3.40 ± 2.18 injections with 7.19 ± 5.44 visits and in the fourth year 3.11 ± 2.29 injections with 6.46 ± 6.20 visits. The VO patients received an average of 4.17 ± 2.15 injections with 9.60 ± 4.88 visits in the first year, 3.31 ± 2.03 injections with 7.75 ± 4.88 visits in the second year, 2.94 ± 2.00 injections with 6.55 ± 4.77 visits in the third year and 3.03 ± 1.94 injections with 7.18 ± 5.26 visits in the fourth year. The course of the visual acuity was relatively stable over 4 years. With repect to the quality of care, a younger patient age and a closer distance to the clinic seemed to have a positive influence. CONCLUSION: The study results show a positive trend towards an improved quality of care in intravitreal injection therapy in patients with AMD and VO. The age of the patient and the place of residence are factors influencing the therapy.

[Optical coherence tomography angiography: Value for glaucoma diagnostics]. / Optische Kohärenztomographie-Angiographie: Stellenwert in der Glaukomdiagnostik.

BACKGROUND: Optical coherence tomography angiography (OCTA) is a novel noninvasive method which enables a quantitative evaluation of retinal and optic nerve head (ONH) perfusion. In this article, we discuss the principles of the application of OCTA and give a summary of the knowledge gained by using this method in glaucoma patients. METHODS: This article is based on a selective literature search and the analysis of own data. RESULTS: Quantitative OCTA parameters have a good reproducibility in glaucoma patients. Glaucoma patients show a reduced flow density (FD) in the ONH and in the area of the macula compared with a healthy control group. The FD parameters show a good diagnostic discriminatory power but are not superior to the structural parameters used in routine diagnostics. The reduced FD measured using OCTA correlates with the extent of functional and structural glaucoma damage. CONCLUSION: The OCTA is noninvasive, fast and reproducible. Initial results from studies on glaucoma patients show the high diagnostic potential of this method. The OCTA could become a part of clinical glaucoma management in the future.

[Foveal avascular zone and OCT angiography. An overview of current knowledge]. / Foveale avaskuläre Zone und OCT-Angiographie. Eine Übersicht aktueller Erkenntnisse.

BACKGROUND: The foveal avascular zone (FAZ) has often been analyzed under different aspects in the last decades, mainly by fluorescence angiography (FA); however, the novel technology of optical coherence tomography angiography (OCTA) enables a non-invasive examination, visualization and quantitative analysis of the FAZ, which has recently led to many new findings, especially in a multidisciplinary manner. This article provides an overview of the investigation of the FAZ using OCTA and the new findings that have been obtained using OCTA in recent years. METHODS: This article is based on a comprehensive literature review. RESULTS: In many studies a good reproducibility and repeatability of the FAZ measurements by OCTA could be proven, also by comparing different OCTA devices. In patients with ocular pathologies and systemic diseases, e. g. after retinal vein occlusion or retinal surgery and in patients with diabetes mellitus without diabetic retinopathy, differences to healthy control groups and correlations to visual function could be shown. Moreover, in patients with neurological diseases, such as Alzheimer's dementia, changes of the FAZ could be identified. CONCLUSION: The OCTA is a non-invasive technology, which enables a reliable visualization and reproducible quantification of the FAZ. The size of the FAZ seems to be altered in different retinal and systemic diseases, which also may correlate with visual function; however, long-term studies evaluating the diagnostic value of these changes in the course of the disease are currently lacking.

Prevalences of segmentation errors and motion artifacts in OCT-angiography differ among retinal diseases.

PURPOSE: To assess the prevalences of segmentation errors and motion artifacts in optical coherence tomography angiography (OCT-A) in different retinal diseases METHODS: In a retrospective analysis, multimodal retinal imaging including OCT-A was performed in one eye of 57 healthy controls (50.96 ± 22.4 years) and 149 patients (66.42 ± 14.1 years) affected by different chorioretinal diseases: early/intermediate age-related macular degeneration (AMD; n = 26), neovascular AMD (nAMD; n = 22), geographic atrophy due to AMD (GA; n = 6), glaucoma (n = 28), central serous chorioretinopathy (CSC; n = 14), epiretinal membrane (EM; n = 26), retinal vein occlusion (RVO; n = 11), and retinitis pigmentosa (RP; n = 16). Central 3 × 3 mm2 OCT-A imaging was performed with active eye-tracking (AngioVue, Optovue). Best-corrected visual acuity (BCVA) and signal strength index (SSI) were recorded. Images were independently evaluated by two graders using the OCT-A motion artifact score (MAS; scores I-IV) as well as a newly introduced segmentation accuracy score (SAS; score I-IIB). RESULTS: Mean SSI was 63.67 ± 9.2 showing a negative correlation with increasing age (rSp = - 0.42, p < 0.001, n = 206). In the healthy cohort, mean MAS was 1.45 ± 0.8 and segmentation was accurate (SAS I) in all eyes. In eyes with retinal pathologies, mean MAS was 2.1 ± 0.9 (p < 0.001). Lowest MAS was observed in GA (2.67 ± 0.5) and RVO (2.45 ± 1.1). Compared to an accurate segmentation in 100% in healthy subjects, 34.2% (n = 51) of all patients showed highest segmentation quality (p < 0.001). 63.8% showed segmentation errors in more than 5% of all single b-scans in one (SAS IIA, n = 58) or at least two (SAS IIB, n = 40) segmentation boundaries. Highest percentages of inaccurate segmentation (SAS IIA or IIB) were observed in the nAMD group (90.1%). The inner plexiform layer was the segmentation boundary most prone to inaccurate segmentation in all pathologies compared to the inner limiting membrane (ILM) and retinal pigment epithelium (RPE) segmentation layer. Incorrect ILM segmentation was only seen in patients with EM. CONCLUSIONS: Prior to both qualitative and quantitative analysis, OCT-A images must be carefully reviewed as motion artifacts and segmentation errors in current OCT-A technology are frequent particularly in pathologically altered maculae.

[Deep learning and neuronal networks in ophthalmology : Applications in the field of optical coherence tomography]. / "Deep Learning" und neuronale Netzwerke in der Augenheilkunde : Einsatzmöglichkeiten im Bereich der optischen Kohärenztomographie.

Deep learning is increasingly becoming the focus of various imaging methods in medicine. Due to the large number of different imaging modalities, ophthalmology is particularly suitable for this field of application. This article gives a general overview on the topic of deep learning and its current applications in the field of optical coherence tomography. For the benefit of the reader it focuses on the clinical rather than the technical aspects.