Reference-Based OCT Angiogram Super-Resolution With Learnable Texture Generation.

Optical coherence tomography angiography (OCTA) can visualize retinal microvasculature and is important to qualitatively and quantitatively identify potential biomarkers for different retinal diseases. However, the resolution of optical coherence tomography (OCT) angiograms inevitably decreases when increasing the field-of-view (FOV) given a fixed acquisition time. To address this issue, we propose a novel reference-based super-resolution (RefSR) framework to preserve the resolution of the OCT angiograms while increasing the scanning area. Specifically, textures from the normal RefSR pipeline are used to train a learnable texture generator (LTG), which is designed to generate textures according to the input. The key difference between the proposed method and traditional RefSR models is that the textures used during inference are generated by the LTG instead of being searched from a single reference (Ref) image. Since the LTG is optimized throughout the whole training process, the available texture space is significantly enlarged and no longer limited to a single Ref image, but extends to all textures contained in the training samples. Moreover, our proposed LTGNet does not require an Ref image at the inference phase, thereby becoming invulnerable to the selection of the Ref image. Both experimental and visual results show that LTGNet has competitive performance and robustness over state-of-the-art methods, indicating good reliability and promise in real-life deployment. The source code is available at https://github.com/RYY0722/LTGNet.

A Competition for the Diagnosis of Myopic Maculopathy by Artificial Intelligence Algorithms.

Importance: Myopic maculopathy (MM) is a major cause of vision impairment globally. Artificial intelligence (AI) and deep learning (DL) algorithms for detecting MM from fundus images could potentially improve diagnosis and assist screening in a variety of health care settings. Objectives: To evaluate DL algorithms for MM classification and segmentation and compare their performance with that of ophthalmologists. Design, Setting, and Participants: The Myopic Maculopathy Analysis Challenge (MMAC) was an international competition to develop automated solutions for 3 tasks: (1) MM classification, (2) segmentation of MM plus lesions, and (3) spherical equivalent (SE) prediction. Participants were provided 3 subdatasets containing 2306, 294, and 2003 fundus images, respectively, with which to build algorithms. A group of 5 ophthalmologists evaluated the same test sets for tasks 1 and 2 to ascertain performance. Results from model ensembles, which combined outcomes from multiple algorithms submitted by MMAC participants, were compared with each individual submitted algorithm. This study was conducted from March 1, 2023, to March 30, 2024, and data were analyzed from January 15, 2024, to March 30, 2024. Exposure: DL algorithms submitted as part of the MMAC competition or ophthalmologist interpretation. Main Outcomes and Measures: MM classification was evaluated by quadratic-weighted κ (QWK), F1 score, sensitivity, and specificity. MM plus lesions segmentation was evaluated by dice similarity coefficient (DSC), and SE prediction was evaluated by R2 and mean absolute error (MAE). Results: The 3 tasks were completed by 7, 4, and 4 teams, respectively. MM classification algorithms achieved a QWK range of 0.866 to 0.901, an F1 score range of 0.675 to 0.781, a sensitivity range of 0.667 to 0.778, and a specificity range of 0.931 to 0.945. MM plus lesions segmentation algorithms achieved a DSC range of 0.664 to 0.687 for lacquer cracks (LC), 0.579 to 0.673 for choroidal neovascularization, and 0.768 to 0.841 for Fuchs spot (FS). SE prediction algorithms achieved an R2 range of 0.791 to 0.874 and an MAE range of 0.708 to 0.943. Model ensemble results achieved the best performance compared to each submitted algorithms, and the model ensemble outperformed ophthalmologists at MM classification in sensitivity (0.801; 95% CI, 0.764-0.840 vs 0.727; 95% CI, 0.684-0.768; P = .006) and specificity (0.946; 95% CI, 0.939-0.954 vs 0.933; 95% CI, 0.925-0.941; P = .009), LC segmentation (DSC, 0.698; 95% CI, 0.649-0.745 vs DSC, 0.570; 95% CI, 0.515-0.625; P < .001), and FS segmentation (DSC, 0.863; 95% CI, 0.831-0.888 vs DSC, 0.790; 95% CI, 0.742-0.830; P < .001). Conclusions and Relevance: In this diagnostic study, 15 AI models for MM classification and segmentation on a public dataset made available for the MMAC competition were validated and evaluated, with some models achieving better diagnostic performance than ophthalmologists.

A deep learning system for myopia onset prediction and intervention effectiveness evaluation in children.

The increasing prevalence of myopia worldwide presents a significant public health challenge. A key strategy to combat myopia is with early detection and prediction in children as such examination allows for effective intervention using readily accessible imaging technique. To this end, we introduced DeepMyopia, an artificial intelligence (AI)-enabled decision support system to detect and predict myopia onset and facilitate targeted interventions for children at risk using routine retinal fundus images. Based on deep learning architecture, DeepMyopia had been trained and internally validated on a large cohort of retinal fundus images (n = 1,638,315) and then externally tested on datasets from seven sites in China (n = 22,060). Our results demonstrated robustness of DeepMyopia, with AUCs of 0.908, 0.813, and 0.810 for 1-, 2-, and 3-year myopia onset prediction with the internal test set, and AUCs of 0.796, 0.808, and 0.767 with the external test set. DeepMyopia also effectively stratified children into low- and high-risk groups (p < 0.001) in both test sets. In an emulated randomized controlled trial (eRCT) on the Shanghai outdoor cohort (n = 3303) where DeepMyopia showed effectiveness in myopia prevention compared to NonCyc-based model, with an adjusted relative reduction (ARR) of -17.8%, 95% CI: -29.4%, -6.4%. DeepMyopia-assisted interventions attained quality-adjusted life years (QALYs) of 0.75 (95% CI: 0.53, 1.04) per person and avoided blindness years of 13.54 (95% CI: 9.57, 18.83) per 1 million persons compared to natural lifestyle with no active intervention. Our findings demonstrated DeepMyopia as a reliable and efficient AI-based decision support system for intervention guidance for children.

Value proposition of retinal imaging in Alzheimer's disease screening: A review of eight evolving trends.

Alzheimer's disease (AD) is the leading cause of dementia worldwide. Current diagnostic modalities of AD generally focus on detecting the presence of amyloid ß and tau protein in the brain (for example, positron emission tomography [PET] and cerebrospinal fluid testing), but these are limited by their high cost, invasiveness, and lack of expertise. Retinal imaging exhibits potential in AD screening and risk stratification, as the retina provides a platform for the optical visualization of the central nervous system in vivo, with vascular and neuronal changes that mirror brain pathology. Given the paradigm shift brought by advances in artificial intelligence and the emergence of disease-modifying therapies, this article aims to summarize and review the current literature to highlight 8 trends in an evolving landscape regarding the role and potential value of retinal imaging in AD screening.

Deep learning-based image quality assessment for optical coherence tomography macular scans: a multicentre study.

AIMS: To develop and externally test deep learning (DL) models for assessing the image quality of three-dimensional (3D) macular scans from Cirrus and Spectralis optical coherence tomography devices. METHODS: We retrospectively collected two data sets including 2277 Cirrus 3D scans and 1557 Spectralis 3D scans, respectively, for training (70%), fine-tuning (10%) and internal validation (20%) from electronic medical and research records at The Chinese University of Hong Kong Eye Centre and the Hong Kong Eye Hospital. Scans with various eye diseases (eg, diabetic macular oedema, age-related macular degeneration, polypoidal choroidal vasculopathy and pathological myopia), and scans of normal eyes from adults and children were included. Two graders labelled each 3D scan as gradable or ungradable, according to standardised criteria. We used a 3D version of the residual network (ResNet)-18 for Cirrus 3D scans and a multiple-instance learning pipline with ResNet-18 for Spectralis 3D scans. Two deep learning (DL) models were further tested via three unseen Cirrus data sets from Singapore and five unseen Spectralis data sets from India, Australia and Hong Kong, respectively. RESULTS: In the internal validation, the models achieved the area under curves (AUCs) of 0.930 (0.885-0.976) and 0.906 (0.863-0.948) for assessing the Cirrus 3D scans and Spectralis 3D scans, respectively. In the external testing, the models showed robust performance with AUCs ranging from 0.832 (0.730-0.934) to 0.930 (0.906-0.953) and 0.891 (0.836-0.945) to 0.962 (0.918-1.000), respectively. CONCLUSIONS: Our models could be used for filtering out ungradable 3D scans and further incorporated with a disease-detection DL model, allowing a fully automated eye disease detection workflow.

Frequency and clinicopathologic features of renal low-grade oncocytic tumour and eosinophilic vacuolated tumour: reclassification of 605 eosinophilic tumours including patients managed with active surveillance.

AIMS: Low-grade oncocytic tumour (LOT) and eosinophilic vacuolated tumour (EVT) are recently described emerging entities, which demonstrate distinct features but are not yet recognised as separate neoplasms in the fifth WHO classification. Published series to date have been largely multi-institutional and based on surgically resected tumours. This study aims to determine the frequency, clinicopathologic features and outcome of LOT and EVT in a single institutional series of oncocytic/eosinophilic renal neoplasms, including patients managed with active surveillance and non-surgical intervention. METHODS AND RESULTS: Cases were identified from a consecutive institutional series of in-house renal tumours diagnosed on biopsy and/or nephrectomy (2003-2023). Tumours with a diagnosis or differential diagnosis of oncocytoma, chromophobe renal cell carcinoma or oncocytic neoplasm not otherwise specified (including LOT, EVT and tumours with overlapping hybrid features) were retrospectively reviewed and classified/reclassified.In total, 605 oncocytic/eosinophilic renal neoplasms were reviewed, among which 33 LOT (5.5%) and 5 EVT (0.8%) were identified. LOT were CK7+, CD117- and GATA3+ (94%). EVT were CD117+, CK7 focal+ (80%) and cathepsin K+ (80%). At the median follow-up of 34 months (range 2-253) and 56 months (range 8-90) for LOT and EVT, respectively, there was no evidence of recurrence following ablation/surgical resection, metastasis or death from disease for all patients, including the 22 managed with active surveillance (20 LOT and 2 EVT). CONCLUSIONS: LOT and EVT comprised a minority of oncocytic renal neoplasms in this series. We report a large institutional series including patients managed non-surgically, with no adverse outcome, adding to the existing literature indicating a benign outcome.

Randomized 20-year infancy-onset dietary intervention, life-long cardiovascular risk factors and retinal microvasculature.

BACKGROUND AND AIMS: Retinal microvasculature characteristics predict cardiovascular morbidity and mortality. This study investigated associations of lifelong cardiovascular risk factors and effects of dietary intervention on retinal microvasculature in young adulthood. METHODS: The cohort is derived from the longitudinal Special Turku Coronary Risk Factor Intervention Project study. The Special Turku Coronary Risk Factor Intervention Project is a 20-year infancy-onset randomized controlled dietary intervention study with frequent study visits and follow-up extending to age 26 years. The dietary intervention aimed at a heart-healthy diet. Fundus photographs were taken at the 26-year follow-up, and microvascular measures [arteriolar and venular diameters, tortuosity (simple and curvature) and fractal dimensions] were derived (n = 486). Cumulative exposure as the area under the curve for cardiovascular risk factors and dietary components was determined for the longest available time period (e.g. from age 7 months to 26 years). RESULTS: The dietary intervention had a favourable effect on retinal microvasculature resulting in less tortuous arterioles and venules and increased arteriolar fractal dimension in the intervention group when compared with the control group. The intervention effects were found even when controlled for the cumulative cardiovascular risk factors. Reduced lifelong cumulative intake of saturated fats, main target of the intervention, was also associated with less tortuous venules. Several lifelong cumulative risk factors were independently associated with the retinal microvascular measures, e.g. cumulative systolic blood pressure with narrower arterioles. CONCLUSIONS: Infancy-onset 20-year dietary intervention had favourable effects on the retinal microvasculature in young adulthood. Several lifelong cumulative cardiovascular risk factors were independently associated with retinal microvascular structure.

Integrated image-based deep learning and language models for primary diabetes care.

Primary diabetes care and diabetic retinopathy (DR) screening persist as major public health challenges due to a shortage of trained primary care physicians (PCPs), particularly in low-resource settings. Here, to bridge the gaps, we developed an integrated image-language system (DeepDR-LLM), combining a large language model (LLM module) and image-based deep learning (DeepDR-Transformer), to provide individualized diabetes management recommendations to PCPs. In a retrospective evaluation, the LLM module demonstrated comparable performance to PCPs and endocrinology residents when tested in English and outperformed PCPs and had comparable performance to endocrinology residents in Chinese. For identifying referable DR, the average PCP's accuracy was 81.0% unassisted and 92.3% assisted by DeepDR-Transformer. Furthermore, we performed a single-center real-world prospective study, deploying DeepDR-LLM. We compared diabetes management adherence of patients under the unassisted PCP arm (n = 397) with those under the PCP+DeepDR-LLM arm (n = 372). Patients with newly diagnosed diabetes in the PCP+DeepDR-LLM arm showed better self-management behaviors throughout follow-up (P < 0.05). For patients with referral DR, those in the PCP+DeepDR-LLM arm were more likely to adhere to DR referrals (P < 0.01). Additionally, DeepDR-LLM deployment improved the quality and empathy level of management recommendations. Given its multifaceted performance, DeepDR-LLM holds promise as a digital solution for enhancing primary diabetes care and DR screening.

Repeatability, interocular correlation and agreement of optic nerve head vessel density in healthy eyes: a swept-source optical coherence tomographic angiography study.

AIM: To assess the repeatability, interocular correlation, and agreement of quantitative swept-source optical coherence tomography angiography (OCTA) optic nerve head (ONH) parameters in healthy subjects. METHODS: Thirty-three healthy subjects were enrolled. The ONH of both eyes were imaged four times by a swept-source-OCTA using a 3 mm ×3 mm scanning protocol. Images of the radial peripapillary capillary were analyzed by a customized Matlab program, and the vessel density, fractal dimension, and vessel diameter index were measured. The repeatability of the four scans was determined by the intraclass correlation coefficient (ICC). The most well-centered optic disc from the four repeated scans was then selected for the interocular correlation and agreement analysis using the Pearson correlation coefficient, ICC and Bland-Altman plots. RESULTS: All swept-source-OCTA ONH parameters exhibited certain repeatability, with ICC>0.760 and coefficient of variation (CoV)≤7.301%. The obvious interocular correlation was observed for papillary vessel density (ICC=0.857), vessel diameter index (ICC=0.857) and fractal dimension (ICC=0.906), while circumpapillary vessel density exhibited moderate interocular correlation (ICC=0.687). Bland-Altman plots revealed an agreement range of -5.26% to 6.21% for circumpapillary vessel density. CONCLUSION: OCTA ONH parameters demonstrate good repeatability in healthy subjects. The interocular correlations of papillary vessel density, fractal dimension and vessel diameter index are high, but the correlation for circumpapillary vessel density is moderate.

Fit-for-Purpose Ki-67 Immunohistochemistry Assays for Breast Cancer.

New therapies are being developed for breast cancer, and in this process, some "old" biomarkers are reutilized and given a new purpose. It is not always recognized that by changing a biomarker's intended use, a new biomarker assay is created. The Ki-67 biomarker is typically assessed by immunohistochemistry (IHC) to provide a proliferative index in breast cancer. Canadian laboratories assessed the analytical performance and diagnostic accuracy of their Ki-67 IHC laboratory-developed tests (LDTs) of relevance for the LDTs' clinical utility. Canadian clinical IHC laboratories enrolled in the Canadian Biomarker Quality Assurance Pilot Run for Ki-67 in breast cancer by invitation. The Dako Ki-67 IHC pharmDx assay was employed as a study reference assay. The Dako central laboratory was the reference laboratory. Participants received unstained slides of breast cancer tissue microarrays with 32 cases and performed their in-house Ki-67 assays. The results were assessed using QuPath, an open-source software application for bioimage analysis. Positive percent agreement (PPA, sensitivity) and negative percent agreement (NPA, specificity) were calculated against the Dako Ki-67 IHC pharmDx assay for 5%, 10%, 20%, and 30% cutoffs. Overall, PPA and NPA varied depending on the selected cutoff; participants were more successful with 5% and 10%, than with 20% and 30% cutoffs. Only 4 of 16 laboratories had robust IHC protocols with acceptable PPA for all cutoffs. The lowest PPA for the 5% cutoff was 85%, for 10% was 63%, for 20% was 14%, and for 30% was 13%. The lowest NPA for the 5% cutoff was 50%, for 10% was 33%, for 20% was 50%, and for 30% was 57%. Despite many years of international efforts to standardize IHC testing for Ki-67 in breast cancer, our results indicate that Canadian clinical LDTs have a wide analytical sensitivity range and poor agreement for 20% and 30% cutoffs. The poor agreement was not due to the readout but rather due to IHC protocol conditions. International Ki-67 in Breast Cancer Working Group (IKWG) recommendations related to Ki-67 IHC standardization cannot take full effect without reliable fit-for-purpose reference materials that are required for the initial assay calibration, assay performance monitoring, and proficiency testing.