Deep Learning-Based Automated Detection of Retinal Breaks and Detachments on Fundus Photography.

Purpose: The purpose of this study was to develop a deep learning algorithm, to detect retinal breaks and retinal detachments on ultra-widefield fundus (UWF) optos images using artificial intelligence (AI). Methods: Optomap UWF images of the database were annotated to four groups by two retina specialists: (1) retinal breaks without detachment, (2) retinal breaks with retinal detachment, (3) retinal detachment without visible retinal breaks, and (4) a combination of groups 1 to 3. The fundus image data set was split into a training set and an independent test set following an 80% to 20% ratio. Image preprocessing methods were applied. An EfficientNet classification model was trained with the training set and evaluated with the test set. Results: A total of 2489 UWF images were included into the dataset, resulting in a training set size of 2008 UWF images and a test set size of 481 images. The classification models achieved an area under the receiver operating characteristic curve (AUC) on the testing set of 0.975 regarding lesion detection, an AUC of 0.972 for retinal detachment and an AUC of 0.913 for retinal breaks. Conclusions: A deep learning system to detect retinal breaks and retinal detachment using UWF images is feasible and has a good specificity. This is relevant for clinical routine as there can be a high rate of missed breaks in clinics. Future clinical studies will be necessary to evaluate the cost-effectiveness of applying such an algorithm as an automated auxiliary tool in a large practices or tertiary referral centers. Translational Relevance: This study demonstrates the relevance of applying AI in diagnosing peripheral retinal breaks in clinical routine in UWF fundus images.

Performance of different automatic photographic identification software for larvae and adults of the European fire salamander.

For many species, population sizes are unknown despite their importance for conservation. For population size estimation, capture-mark-recapture (CMR) studies are often used, which include the necessity to identify each individual, mostly through individual markings or genetic characters. Invasive marking techniques, however, can negatively affect the individual fitness. Alternatives are low-impact techniques such as the use of photos for individual identification, for species with stable distinctive phenotypic traits. For the individual identification of photos, a variety of different software, with different requirements, is available. The European fire salamander (Salamandra salamandra) is a species in which individuals, both at the larval stage and as adults, have individual specific patterns that allow for individual identification. In this study, we compared the performance of five different software for the use of photographic identification for the European fire salamander: Amphibian & Reptile Wildbook (ARW), AmphIdent, I3S pattern+, ManderMatcher and Wild-ID. While adults can be identified by all five software, European fire salamander larvae can currently only be identified by two of the five (ARW and Wild-ID). We used one dataset of European fire salamander larval pictures taken in the laboratory and tested this dataset in two of the five software (ARW and Wild-ID). We used another dataset of European fire salamander adult pictures taken in the field and tested this using all five software. We compared the requirements of all software on the pictures used and calculated the False Rejection Rate (FRR) and the Recognition Rate (RR). For the larval dataset (421 pictures) we found that the ARW and Wild-ID performed equally well for individual identification (99.6% and 100% Recognition Rate, respectively). For the adult dataset (377 pictures), we found the best False Rejection Rate in ManderMatcher and the highest Recognition Rate in the ARW. Additionally, the ARW is the only program that requires no image pre-processing. In times of amphibian declines, non-invasive photo identification software allowing capture-mark-recapture studies help to gain knowledge on population sizes, distribution, movement and demography of a population and can thus help to support species conservation.

ImageAssist: A Novel Smartphone Application for Standardized Clinical Photography That Implements and Automates American Society of Plastic Surgery/Plastic Surgery Foundation Photographic Guidelines With Background Deletion.

STATEMENT OF THE PROBLEM: Standardized medical photography of the face is a vital part of patient documentation, clinical evaluation, and scholarly dissemination. Because digital photography is a mainstay in clinical care, there is a critical need for an easy-to-use mobile device application that could assist users in taking a standardized clinical photograph. ImageAssist was developed to answer this need. The mobile application is integrated into the electronic medical record (EMR); it implements and automates American Society of Plastic Surgery/Plastic Surgery Research Foundation photographic guidelines with background deletion. INITIAL PRODUCT DEVELOPMENT: A team consisting of a craniofacial plastic surgeon and the Health Information Technology product group developed and implemented the pilot application of ImageAssist. The application launches directly from patients' chart in the mobile version of the EMR, EPIC Haiku (Verona, Wisconsin). Standard views of the face (90-degree, oblique left and right, front and basal view) were built into digital templates and are user selected. Red digital frames overlay the patients' face on the screen and turn green once standardized alignment is achieved, prompting the user to capture. The background is then digitally subtracted to a standard blue, and the photograph is not stored on the user's phone. EARLY USER EXPERIENCE: ImageAssist initial beta user group was limited to 13 providers across dermatology, ENT, and plastic surgery. A mix of physicians, advanced practice providers, and nurses was included to pilot the application in the outpatient clinic setting using Image Assist on their smart phone. After using the app, an internal survey was used to gain feedback on the user experience. In the first 2 years of use, 31 users have taken more than 3400 photographs in more than 800 clinical encounters. Since initial release, automated background deletion also has been functional for any anatomic area. CONCLUSIONS: ImageAssist is a novel smartphone application that standardizes clinical photography and integrated into the EMR, which could save both time and expense for clinicians seeking to take consistent clinical images. Future steps include continued refinement of current image capture functionality and development of a stand-alone mobile device application.

Occlusion development after premature loss of deciduous anterior teeth: preliminary results of a 24-month prospective cohort study.

OBJECTIVE: This study aimed to evaluate occlusion development after premature loss or extraction of deciduous anterior teeth, by means of a prospective cohort study. METHODS: Fifteen infants and children aged 1 to 5 years old were longitudinally assessed (with loss or extraction of deciduous anterior teeth [n = 9], and without tooth losses [n = 6]). Photographs and dental casts at the baseline and after 24 months of follow-up were performed. Dental casts were scanned, and linear measurements were made on the digitalized models (missing tooth space, arch perimeter, arch length, arch width, intercanine length and intercanine width). The t-test was used for groups comparisons (α = 0.05). RESULTS: Individuals' mean age at baseline was 2.93 (± 1.18) years. No statistically significant differences were observed in the missing tooth space in the group with tooth loss during the 24 months of follow-up (p > 0.05). Arch perimeter, arch length, arch width, intercanine length and intercanine width did not show differences between the groups (p > 0.05). Qualitative photographic evaluation revealed other changes in the dental arches and occlusion, such as exfoliation and eruption of deciduous teeth, eruption of permanent teeth, self-correction or establishment of malocclusion, among others. CONCLUSION: The results suggest that the premature loss of deciduous anterior teeth does not affect the perimeter, length and width of the dental arches; however, other alterations that lead to malocclusion could be established.

Evaluating imaging repeatability of fully self-service fundus photography within a community-based eye disease screening setting.

PURPOSE: This study aimed to investigate the imaging repeatability of self-service fundus photography compared to traditional fundus photography performed by experienced operators. DESIGN: Prospective cross-sectional study. METHODS: In a community-based eye diseases screening site, we recruited 65 eyes (65 participants) from the resident population of Shanghai, China. All participants were devoid of cataract or any other conditions that could potentially compromise the quality of fundus imaging. Participants were categorized into fully self-service fundus photography or traditional fundus photography group. Image quantitative analysis software was used to extract clinically relevant indicators from the fundus images. Finally, a statistical analysis was performed to depict the imaging repeatability of fully self-service fundus photography. RESULTS: There was no statistical difference in the absolute differences, or the extents of variation of the indicators between the two groups. The extents of variation of all the measurement indicators, with the exception of the optic cup area, were below 10% in both groups. The Bland-Altman plots and multivariate analysis results were consistent with results mentioned above. CONCLUSIONS: The image repeatability of fully self-service fundus photography is comparable to that of traditional fundus photography performed by professionals, demonstrating promise in large-scale eye disease screening programs.

Spot speed cameras in a series - Effects on speed and traffic safety.

Reduced speeds and increased speed compliance are crucial for achieving increased road traffic safety, cutting across most Safe System interventions. Speed cameras have been shown to be effective in increasing speed compliance and reducing the number of fatalities and seriously injured. The speed cameras system in Sweden is different compared to many other countries, spot speed cameras are almost always placed in series along a road stretch. The aim of this study is to investigate the effects of this system on mean speeds, speed compliance, and on the number of fatalities and seriously injured. Including 20 years of data, the study applies before-after analysis to 361 speed measurement spots, and Empirical Bayes before-after analysis with control to crash outcomes on 202 road sections. The results show a mean speed decrease of 3.5 km/h for all vehicles and road sections, 7.9 km/h at cameras and 3.0 km/h between cameras. Furthermore, follow-up measurements showed that the effects were maintained long-term. Speed compliance increased 16 %-units, 42 %-units at cameras and 13 %-units between cameras. Though larger effects can be seen at cameras, there are still substantial effects on the enforced road sections between cameras. The cameras had an average effect of 38.6 % on decreasing fatalities and may also suggest a decrease for seriously injured, though not statistically significant. This study also shows that for roads that received both a decreased speed limit from 90 to 80 km/h and speed cameras, the mean speeds were reduced by additionally 3.6 km/h compared to roads with unchanged limits of 90 km/h. The combined effect on fatalities and seriously injured was a reduction by 61.6 % and 33.4 %. In conclusion, the Swedish strategy with spot speed cameras in a series led to an increased speed compliance and a comprehensive reduction in mean speeds and of the number of fatalities.

Evaluation of point-to-point speed cameras to control speeding behavior in Thailand.

INTRODUCTION: This research examines the effectiveness of point-to-point speed cameras in Thailand compared with spot speed camera enforcement, which is widely used in the country. METHOD: The evaluation includes the speed control's efficiency by observing the 85th percentile average speed of drivers on the road; the speed detection rate; conducting a comparison with spot speed camera enforcement; and a cost effectiveness analysis of this measure. RESULTS: After speed detection using point-to-point cameras were employed during which warning tickets were issued in many days of experiment, the 85th percentile average speed in the controlled area decreased by 10 km/hr, with a maximum decrease of 20 km/hr. When comparing the detection rate of the two camera models, point-to-point cameras could detect 95% of drivers' driving speed, with 45% of the detected drivers identified as speeding drivers, while spot speed cameras were only able to detect 10% to 20% of speeders. Drivers traveling through the area with point-to-point cameras were more likely to drive at a constant speed throughout the average speed detection area. PRACTICAL APPLICATIONS: The point-to-point camera system is determined to be quite cost-effective.

Professional Cleaners' and Healthcare Workers' Ability to Recognize Hand Eczema.

Timely intervention reduces the risk of a poor prognosis in hand eczema, making early recognition of symptoms important in high-risk professions. However, limited data exist regarding the ability of cleaners and healthcare workers to recognize hand eczema. The aim of this study was to examine cleaners' and healthcare workers' ability to recognize hand eczema in clinical photographs and to assess the severity of the disease. Cleaners and healthcare workers completed a questionnaire consisting of 16 questions and participated in a structured interview referring to a validated photographic severity guide for chronic hand eczema, which comprised clinical photographs of hand eczema at varying levels of severity. Eighty cleaners and 201 healthcare workers (total N = 281) participated in the study. The rates of correctly identified hand eczema in clinical photographs (cleaners/ healthcare workers) were: 41.2%/57.7% (mild hand eczema), 81.2%/92.0% (moderate hand eczema), 85.0%/94.5% (severe hand eczema) and 82.5%/97.0% (very severe hand eczema). The proficiency of healthcare workers in recognizing hand eczema was significantly higher than that of cleaners. The results indicate that a large proportion of cleaners and healthcare workers fail to recognize mild hand eczema in clinical photographs. Healthcare workers had higher success rates in recognizing hand eczema in all severity categories. Symptom underestimation may lead to under-reporting of the true prevalence of hand eczema, with consequent loss of opportunities for prevention.

Editing of 3D images in echocardiographic evaluation of mitral valve: Is there any correlation with photography?

Editing of 3D raw images acquired during transesophageal echocardiography could be similar to the post processing of raw images with digital software used in photography. 3D image editing in echocardiography is often underestimated in clinical practice, and people are satisfied with the first 3D image they are able to obtain during transesophageal examination. In fact, it is often believed to represent solely an aesthetic addition that does not change the information already obtained with 2D and baseline 3D. In reality, it represents a crucial moment to better understand the mechanisms of mitral pathology, avoiding artifacts and misjudgments. The importance of acquiring raw 3D images of the valve having all the necessary information (ring, the leaflet in toto, the right frame rate) allows us then to edit them making them more beautiful and clearer from the point of view of the information received. Nevertheless, by exclusively acquiring a raw 3D with all the necessary information, we can quickly finish the transesophageal examination, reducing its duration and discomfort for the patient, as well as the inherent risk of complications related to the procedure per se.

Development of a deep learning model to distinguish the cause of optic disc atrophy using retinal fundus photography.

The differential diagnosis for optic atrophy can be challenging and requires expensive, time-consuming ancillary testing to determine the cause. While Leber's hereditary optic neuropathy (LHON) and optic neuritis (ON) are both clinically significant causes for optic atrophy, both relatively rare in the general population, contributing to limitations in obtaining large imaging datasets. This study therefore aims to develop a deep learning (DL) model based on small datasets that could distinguish the cause of optic disc atrophy using only fundus photography. We retrospectively reviewed fundus photographs of 120 normal eyes, 30 eyes (15 patients) with genetically-confirmed LHON, and 30 eyes (26 patients) with ON. Images were split into a training dataset and a test dataset and used for model training with ResNet-18. To visualize the critical regions in retinal photographs that are highly associated with disease prediction, Gradient-Weighted Class Activation Map (Grad-CAM) was used to generate image-level attention heat maps and to enhance the interpretability of the DL system. In the 3-class classification of normal, LHON, and ON, the area under the receiver operating characteristic curve (AUROC) was 1.0 for normal, 0.988 for LHON, and 0.990 for ON, clearly differentiating each class from the others with an overall total accuracy of 0.93. Specifically, when distinguishing between normal and disease cases, the precision, recall, and F1 scores were perfect at 1.0. Furthermore, in the differentiation of LHON from other conditions, ON from others, and between LHON and ON, we consistently observed precision, recall, and F1 scores of 0.8. The model performance was maintained until only 10% of the pixel values of the image, identified as important by Grad-CAM, were preserved and the rest were masked, followed by retraining and evaluation.

Multispectral Fundus Photography of Choroidal Nevi With Trans-Palpebral Illumination.

Purpose: The purpose of this study was to investigate the spectral characteristics of choroidal nevi and assess the feasibility of quantifying the basal diameter of choroidal nevi using multispectral fundus images captured with trans-palpebral illumination. Methods: The study used a widefield fundus camera with multispectral (625 nm, 780 nm, 850 nm, and 970 nm) trans-palpebral illumination to examine eight subjects diagnosed with choroidal nevi. Geometric features of nevi, including border clarity, overlying drusen, and lesion basal diameter, were characterized. Clinical imagers, including scanning laser ophthalmoscopy (SLO), autofluorescence (AF), and optical coherence tomography (OCT), were utilized for comparative assessment. Results: Fundus images depicted nevi as dark regions with high contrast against the background. Near-infrared (NIR) fundus images provided enhanced visibility of lesion borders compared to visible fundus images and SLO images. Lesion-background contrast measurements revealed 635 nm SLO at 11% and 625 nm fundus at 42%. Significantly enhanced contrasts were observed in NIR fundus images at 780 nm (73%), 850 nm (63%), and 970 nm (67%). For quantifying the diameter of nevi, NIR fundus images at 780 nm and 850 nm yielded a deviation of less than 10% when compared to OCT measurements. Conclusions: NIR fundus photography with trans-palpebral illumination enhances nevi visibility and boundary definition compared to SLO. Agreement in diameter measurements with OCT validates the accuracy and reliability of this method for choroidal nevi assessment. Translational Relevance: Multispectral fundus imaging with trans-palpebral illumination improves choroidal nevi visibility and accurately measures basal diameter, promising to enhance clinical practices in screening, diagnosis, and monitoring of choroidal nevi.

Online images amplify gender bias.

Each year, people spend less time reading and more time viewing images1, which are proliferating online2-4. Images from platforms such as Google and Wikipedia are downloaded by millions every day2,5,6, and millions more are interacting through social media, such as Instagram and TikTok, that primarily consist of exchanging visual content. In parallel, news agencies and digital advertisers are increasingly capturing attention online through the use of images7,8, which people process more quickly, implicitly and memorably than text9-12. Here we show that the rise of images online significantly exacerbates gender bias, both in its statistical prevalence and its psychological impact. We examine the gender associations of 3,495 social categories (such as 'nurse' or 'banker') in more than one million images from Google, Wikipedia and Internet Movie Database (IMDb), and in billions of words from these platforms. We find that gender bias is consistently more prevalent in images than text for both female- and male-typed categories. We also show that the documented underrepresentation of women online13-18 is substantially worse in images than in text, public opinion and US census data. Finally, we conducted a nationally representative, preregistered experiment that shows that googling for images rather than textual descriptions of occupations amplifies gender bias in participants' beliefs. Addressing the societal effect of this large-scale shift towards visual communication will be essential for developing a fair and inclusive future for the internet.

Clinical utility of handheld fundus and smartphone-based camera for monitoring diabetic retinal diseases: a review study.

Diabetic retinopathy (DR) is the leading global cause of vision loss, accounting for 4.8% of global blindness cases as estimated by the World Health Organization (WHO). Fundus photography is crucial in ophthalmology as a diagnostic tool for capturing retinal images. However, resource and infrastructure constraints limit access to traditional tabletop fundus cameras in developing countries. Additionally, these conventional cameras are expensive, bulky, and not easily transportable. In contrast, the newer generation of handheld and smartphone-based fundus cameras offers portability, user-friendliness, and affordability. Despite their potential, there is a lack of comprehensive review studies examining the clinical utilities of these handheld (e.g. Zeiss Visuscout 100, Volk Pictor Plus, Volk Pictor Prestige, Remidio NMFOP, FC161) and smartphone-based (e.g. D-EYE, iExaminer, Peek Retina, Volk iNview, Volk Vistaview, oDocs visoScope, oDocs Nun, oDocs Nun IR) fundus cameras. This review study aims to evaluate the feasibility and practicality of these available handheld and smartphone-based cameras in medical settings, emphasizing their advantages over traditional tabletop fundus cameras. By highlighting various clinical settings and use scenarios, this review aims to fill this gap by evaluating the efficiency, feasibility, cost-effectiveness, and remote capabilities of handheld and smartphone fundus cameras, ultimately enhancing the accessibility of ophthalmic services.

Comparison of three-dimensional imaging of the nose using three different 3D-photography systems: an observational study.

BACKGROUND: New 3D technologies for superficial soft tissue changes, especially in plastic and reconstructive surgical procedures, can improve the planning and documentation of facial surgeries. The purpose of this study was to compare and determine the applicability and feasibility of three different 3D-photography systems in clinical practice imaging the nose. METHODS: A total of 16 healthy non-operated noses were included in this prospective study. A plaster model of each nose was produced, digitized, and converted to a .stl mesh (= ground truth model). Three-dimensional images of each nose were then taken using Artec Space Spider (gold standard), Planmeca ProFace®, and the Bellus3D Dental Pro application. All resulting .stl files were aligned to the ground truth model using MeshLab software, and the root mean square error (RMSE), mean surface distance (MSD), and Hausdorff distance (HD) were calculated. RESULTS: The Artec Space Spider 3D-photography system showed significantly better results compared to the two other systems in regard to RMSE, MSD, and HD (each p < 0.001). There was no significant difference between Planmeca ProFace® and Bellus3D Dental Pro in terms of RMSE, MSD, and HD. Overall, all three camera systems showed a clinically acceptable deviation to the reference model (range: -1.23-1.57 mm). CONCLUSIONS: The three evaluated 3D-photography systems were suitable for nose imaging in the clinical routine. While Artec Space Spider showed the highest accuracy, the Bellus3D Dental Pro app may be the most feasible option for everyday clinical use due to its portability, ease of use, and low cost. This study presents three different systems, allowing readers to extrapolate to other systems when planning to introduce 3D photography in the clinical routine.

Performance of Artificial Intelligence in Detecting Diabetic Macular Edema From Fundus Photography and Optical Coherence Tomography Images: A Systematic Review and Meta-analysis.

BACKGROUND: Diabetic macular edema (DME) is the leading cause of vision loss in people with diabetes. Application of artificial intelligence (AI) in interpreting fundus photography (FP) and optical coherence tomography (OCT) images allows prompt detection and intervention. PURPOSE: To evaluate the performance of AI in detecting DME from FP or OCT images and identify potential factors affecting model performances. DATA SOURCES: We searched seven electronic libraries up to 12 February 2023. STUDY SELECTION: We included studies using AI to detect DME from FP or OCT images. DATA EXTRACTION: We extracted study characteristics and performance parameters. DATA SYNTHESIS: Fifty-three studies were included in the meta-analysis. FP-based algorithms of 25 studies yielded pooled area under the receiver operating characteristic curve (AUROC), sensitivity, and specificity of 0.964, 92.6%, and 91.1%, respectively. OCT-based algorithms of 28 studies yielded pooled AUROC, sensitivity, and specificity of 0.985, 95.9%, and 97.9%, respectively. Potential factors improving model performance included deep learning techniques, larger size, and more diversity in training data sets. Models demonstrated better performance when validated internally than externally, and those trained with multiple data sets showed better results upon external validation. LIMITATIONS: Analyses were limited by unstandardized algorithm outcomes and insufficient data in patient demographics, OCT volumetric scans, and external validation. CONCLUSIONS: This meta-analysis demonstrates satisfactory performance of AI in detecting DME from FP or OCT images. External validation is warranted for future studies to evaluate model generalizability. Further investigations may estimate optimal sample size, effect of class balance, patient demographics, and additional benefits of OCT volumetric scans.

Comparison of quantification of intraretinal hard exudates between optical coherence tomography en face image versus fundus photography.

PURPOSE: To compare the quantification of intraretinal hard exudate (HE) using en face optical coherence tomography (OCT) and fundus photography. METHODS: Consecutive en face images and corresponding fundus photographs from 13 eyes of 10 patients with macular edema associated with diabetic retinopathy or Coats' disease were analyzed using the machine-learning-based image analysis tool, "ilastik." RESULTS: The overall measured HE area was greater with en face images than with fundus photos (en face: 0.49 ± 0.35 mm2 vs. fundus photo: 0.34 ± 0.34 mm2, P < 0.001). However, there was an excellent correlation between the two measurements (intraclass correlation coefficient [ICC] = 0.844). There was a negative correlation between HE area and central macular thickness (CMT) (r = -0.292, P = 0.001). However, HE area showed a positive correlation with CMT in the previous several months, especially in eyes treated with anti-vascular endothelial growth factor (VEGF) therapy (CMT 3 months before: r = 0.349, P = 0.001; CMT 4 months before: r = 0.287, P = 0.012). CONCLUSION: Intraretinal HE can be reliably quantified from either en face OCT images or fundus photography with the aid of an interactive machine learning-based image analysis tool. HE area changes lagged several months behind CMT changes, especially in eyes treated with anti-VEGF injections.

Development of a generative deep learning model to improve epiretinal membrane detection in fundus photography.

BACKGROUND: The epiretinal membrane (ERM) is a common retinal disorder characterized by abnormal fibrocellular tissue at the vitreomacular interface. Most patients with ERM are asymptomatic at early stages. Therefore, screening for ERM will become increasingly important. Despite the high prevalence of ERM, few deep learning studies have investigated ERM detection in the color fundus photography (CFP) domain. In this study, we built a generative model to enhance ERM detection performance in the CFP. METHODS: This deep learning study retrospectively collected 302 ERM and 1,250 healthy CFP data points from a healthcare center. The generative model using StyleGAN2 was trained using single-center data. EfficientNetB0 with StyleGAN2-based augmentation was validated using independent internal single-center data and external datasets. We randomly assigned healthcare center data to the development (80%) and internal validation (20%) datasets. Data from two publicly accessible sources were used as external validation datasets. RESULTS: StyleGAN2 facilitated realistic CFP synthesis with the characteristic cellophane reflex features of the ERM. The proposed method with StyleGAN2-based augmentation outperformed the typical transfer learning without a generative adversarial network. The proposed model achieved an area under the receiver operating characteristic (AUC) curve of 0.926 for internal validation. AUCs of 0.951 and 0.914 were obtained for the two external validation datasets. Compared with the deep learning model without augmentation, StyleGAN2-based augmentation improved the detection performance and contributed to the focus on the location of the ERM. CONCLUSIONS: We proposed an ERM detection model by synthesizing realistic CFP images with the pathological features of ERM through generative deep learning. We believe that our deep learning framework will help achieve a more accurate detection of ERM in a limited data setting.