The ophthalmology field was among the first to adopt artificial intelligence (AI) in medicine. The availability of digitized ocular images and substantial data have made deep learning (DL) a popular topic.
Main text
At the moment, AI in ophthalmology is mostly used to improve disease diagnosis and assist decision-making aiming at ophthalmic diseases like diabetic retinopathy (DR), glaucoma, age-related macular degeneration (AMD), cataract and other anterior segment diseases. However, most of the AI systems developed to date are still in the experimental stages, with only a few having achieved clinical applications. There are a number of reasons for this phenomenon, including security, privacy, poor pervasiveness, trust and explainability concerns.
Conclusions
This review summarizes AI applications in ophthalmology, highlighting significant clinical considerations for adopting AI techniques and discussing the potential challenges and future directions.
The effect of interventional time for 3% Diquafosol reatment in post-cataract surgery has not been well established. A meta-analysis was performed to evaluate the improvement of ocular surface condition in post-cataract surgery patients who received 3% DQS for various treatment durations.
Methods
Studies were performed based on 5 databases: PubMed, Cochrane Library, Web of Science, Embase, and China National Knowledge Infrastructure. Data on changes in Schirmer’s test, tear breakup time (TBUT), corneal staining score, and OSDI score were collected for meta-analysis.
Results
A total of 621 affected eyes from 9 independent clinical studies were included. 6 studies conducted Schirmer's test after the application of 3% DQS. Meta-analysis showed that the difference between 3% DQS and control groups was not statistically significant for short-term application (less than or equal to 1 month) (WMD = 0.14, P = 0.27, 95% CI:-0.11 to 0.39), but was statistically different for long-term application (longer than or equal to 3 months) (WMD = 0.76, P = 0.03, 95% CI:0.08 to 1.45). For the corneal fluorescence staining score, the data from 6 studies indicated that the improvement was statistically significant for short-term application (WMD = -0.40, P <0.00001, 95% CI:-0.72 to -0.08) and but not long-term application (WMD = -0.21, P = 0.26, 95% CI:-0.57 to 0.15). For TBUT, the data from 9 studies indicated that both short-term and long-term application showed significant improvement (WMD = 1.70, P <0.00001, 95% CI:1.38 to 2.03; WMD = 1.52, P <0.00001, 95% CI:1.09 to 1.95). Similar results were observed in data from 5 studies with OSDI scores, where both short-term and long-term application showed statistically significant improvements (WMD = -5.41, P <0.00001, 95% CI: -7.02 to -3.81; WMD = -6.10, P <0.00001, 95% CI:-8.52 to -3.67).
Conclusions
The application of 3% DQS in post-operative cataract patients has a positive effect on improving the ocular surface conditions. Short-term application resulted in lower corneal staining scores, prolonged TBUT, and improved OSDI scores. Long-term application improved Schirmer's test results, TBUT, and subjective symptoms.
Key messages
The updated article suggests that 3% Diquafosol is less effective in the short term after cataract surgery, and that application over three months can improve the patient's ocular surface condition.
Monoclonal antibodies such as rituximab (RTX), eculizumab, inebilizumab, satralizumab, and tocilizumab have been found to be effective therapies for neuromyelitis optica spectrum disease (NMOSD) in several clinical randomized controlled trials.
Objective
The purpose of this meta-analysis of randomized controlled trials was to assess the efficacy and safety of monoclonal antibodies in the treatment of NMOSD.
Methods
We searched the following databases for relevant English language literature from the establishment of the database to June 2021: PubMed, Embase, Cohorane Library, the Central Register of Controlled Trials (CENTRAL), and Web of Science. Randomized controlled trials of monoclonal antibodies were the targets of the review.
Results
We included seven trials containing 775 patients (485 in the monoclonal antibody group and 290 in the control group). Patients in the monoclonal group (HR 0.24, 95% CI: 0.14 to 0.40, P < 0.00001), as well as patients with seropositive AQP4-IgG (HR 0.18, 95% CI: 0.11 to 0.29, P < 0.00001), both had a higher free recurrence rate than that in the control group. In the first year (HR 0.25, 95% CI: 0.09 to 0.71, P = 0.009) and the second year (HR 0.32, 95% CI: 0.13 to 0.81, P = 0.02), no relapses were documented. The average changes of the expanded disability status scale (EDSS) score decreased by 0.29 (95% CI: −0.09 to 0.51, P = 0.005). Upper respiratory tract infection (OR 1.52, 95% CI: 0.76 to 3.04, P = 0.24), urinary tract infection(OR 0.79, 95% CI: 0.51 to 1.21, P = 0.27), and headache (OR 1.30, 95% CI: 0.78 to 2.17, P = 0.31) were three most frequent adverse reactions.
Conclusions
Monoclonal antibodies are particularly effective treatments in avoiding recurrence for NMOSD patients, according to this meta-analysis. The associated adverse responses are not significantly different from those seen with traditional immunosuppressants.
To observe the expression differences and potential effects of autophagy-related Beclin1 (mammalian Atg6) and Uncoordinated-51 like kinase 1 (ULK1) in the oxygen-induced retinopathy (OIR) model.
Materials and methods
Thirty-three C57BL/6 mice in OIR model group were exposed to 75 ± 0.5% oxygen from postnatal day-of-life 7 (P7) to P12, and were then brought into normal room environment (relative hypoxia stage) and raised to P17. Thirty-three control mice were kept in a normal room environment. The expression of autophagy in the retina tissue was assessed by Western blot analysis. The thickness and ultrastructural of retina were observed by light microscopy and transmission electron microscope (TEM) on P17.
Results
In the hyperoxia stage (P8–P11), the expression of Beclin1, ULK1 and Autophagy 5 (Atg5) in retina showed no significant difference between the OIR model group and the control group. In the relatively hypoxia stage (P14 to P17), however, the protein level of Beclin1, ULK1, and Bcl-2-associated X protein (Bax) were upregulated in the retina of the OIR model group, whereas B-cell lymphoma 2 (Bcl-2) was downregulated. The autophagosomes in the photoreceptors of retina in the OIR mice were observed. The inner-segment/out-segment (IS/OS) layer in OIR model group was thinner than that the control group on P17.
Conclusions
The expression of Beclin-1 and ULK1 in retina has changed in the OIR model, and the change of Beclin-1 and ULK1 expression is related to the change of oxygen concentration.
关键词:Retinopathy of prematurity;Autophagy;Retina;Beclin1;Uncoordinated-51 like kinase 1 (ULK1)
Phantom eye pain (PEP) is a major clinical problem after eye removal with no standard treatment protocol to date. As pain is a multidimensional experience associated with emotional and cognitive components, this study aimed to explore the possible neuropsychological mechanisms of PEP in a perspective of emotional cognition, in order to provide a basis for clinical treatment.
Methods
Visual oddball event-related potentials (ERPs) under different external emotional stimuli (Disgust, Fear, Sadness, Happiness, Erotica and Neutral) were tested in 12 patients and 12 healthy volunteers. Participants' affective states were measured with the Mood Disorder Questionnaire (MDQ), the Hypomania Checklist-32 (HCL-32), and the Plutchik–van Praag Depression Inventory (PVP). The amplitudes and latencies of N1, P2, N2 and P3 components were analyzed by three-way ANOVA, i.e., group (2) × emotion (6) × electrode (3). Multiple comparisons were performed using Bonferroni's test.
Results
Longer N1 latencies, increased N1 amplitudes; shorter P2 latencies under Disgust and Happiness, decreased P2 amplitudes; shorter N2 latencies under Erotica, increased N2 amplitudes were found in patients compared with controls. There was no main effect of group or interaction effect on P3 latencies and P3 amplitudes. The MDQ and HCL-32 scores were lower, and the N1 latencies under Sadness were negatively correlated with PVP scores in patients.
Conclusions
PEP patients showed reversed patterns in exogenous attention allocation and enhanced involuntary attention to emotional stimuli compared with controls. This study demonstrated cortical processing of emotions in PEP patients and could provide a basis for developing emotional intervention therapy.
关键词:Phantom eye pain;Event-related potentials;External emotions;Affective states
To analyse the differences between 30° blue autofluorescence (BAF30), 55° blue autofluorescence (BAF55) and 200° green Ultra-Wide Field autofluorescence (UWF200) imaging to detect retinal displacement (RD) after vitrectomy surgery for rhegmatogenous retinal detachment (RRD).
Methods
This cross-sectional study considers forty-nine consecutive patients who underwent RRD surgery in the time period 4–8 months previous to image acquisition. The exclusion criteria contemplate previous retinal pathology, under 18-year-olds and non-assessable images in any of the 3 devices. These images were analysed by two masked graders that assessed either the presence or absence of retinal displacement. A third observer reviewed the images that presented discordance.
Results
A total of forty-nine patients were analysed. 7 eyes were excluded due to poor quality in either of the imaging modalities. The final analysis included 42 eyes of forty-two patients with a mean age of 60.3 ± 11.9 years. All patients underwent a 3 port 23-gauge pars plana vitrectomy as the technique of choice. Any grade of RD was detectable in 45.2% of images. It was similar between BAF30 and UWF200 (42.9% of eyes). BAF55 showed RD to a lesser extent (38.1%). Agreement index between BAF30 and BAF55 was 0.901, 0.903 between BAF30 and UWF200 and 0.803 between BAF55 and UWF200. Kappa agreement index between graders was 0.775 for BAF30, 0.798 for BAF50 and 0.808 for UWF200 images.
Conclusions
All imaging modalities were able to detect RD after vitrectomy for RRD, with no inferiority of BAF30 and BAF55 over UWF200.
Due to limited imaging conditions, the quality of fundus images is often unsatisfactory, especially for images photographed by handheld fundus cameras. Here, we have developed an automated method based on combining two mirror-symmetric generative adversarial networks (GANs) for image enhancement.
Methods
A total of 1047 retinal images were included. The raw images were enhanced by a GAN-based deep enhancer and another methods based on luminosity and contrast adjustment. All raw images and enhanced images were anonymously assessed and classified into 6 levels of quality classification by three experienced ophthalmologists. The quality classification and quality change of images were compared. In addition, image-detailed reading results for the number of dubiously pathological fundi were also compared.
Results
After GAN enhancement, 42.9% of images increased their quality, 37.5% remained stable, and 19.6% decreased. After excluding the images at the highest level (level 0) before enhancement, a large number (75.6%) of images showed an increase in quality classification, and only a minority (9.3%) showed a decrease. The GAN-enhanced method was superior for quality improvement over a luminosity and contrast adjustment method (P<0.001). In terms of image reading results, the consistency rate fluctuated from 86.6% to 95.6%, and for the specific disease subtypes, both discrepancy number and discrepancy rate were less than 15 and 15%, for two ophthalmologists.
Conclusions
Learning the style of high-quality retinal images based on the proposed deep enhancer may be an effective way to improve the quality of retinal images photographed by handheld fundus cameras.