Takahiro Mizukami

Background: Narrowing of the trabecular–iris angle (TIA) is a key anatomical feature of primary angle-closure disease and contributes to impaired aqueous humor outflow; however, conventional angle assessment requires specialized expertise or costly imaging equipment. This study aimed to develop an artificial intelligence model for quantitative TIA estimation from anterior-segment images acquired with a smartphone-attached slit-lamp microscope. Methods: In this retrospective study, 127,512 candidate frames were extracted from anterior-segment videos obtained from 2654 eyes. After automated diagnosability filtering, 59,412 frames suitable for TIA500 estimation were retained and assigned to the training, validation, and test datasets using patient-level grouping. A ResNet50-based model was trained to estimate inferior, nasal, and temporal TIA500 values using the corresponding measurements obtained by anterior-segment optical coherence tomography (AS-OCT) as the ground-truth reference. Model performance was evaluated by comparing the AI-predicted values with the AS-OCT measurements using mean absolute error (MAE), root mean squared error (RMSE), Spearman’s rank correlation coefficient, and the intraclass correlation coefficient (ICC). Results: Per-eye MAEs were 7.01° ± 5.69°, 7.28° ± 5.80°, and 7.90° ± 6.37° for inferior, nasal, and temporal TIA, respectively. Corresponding correlations with reference measurements were 0.683, 0.696, and 0.646, while ICCs were 0.769, 0.764, and 0.743 (all p < 0.001). Conclusions: These findings demonstrate the feasibility of estimating TIA500 from smartphone-based slit-lamp images. However, the moderate correlations with AS-OCT measurements indicate that the present model is not yet sufficiently accurate for independent clinical diagnosis or screening, and further model improvement and external validation are required.

Background: Anemia is a common systemic condition associated with adverse maternal, perioperative, and cardiovascular outcomes. Although timely screening is clinically important, diagnosis still relies on invasive blood testing. Palpebral conjunctival pallor has traditionally been used as a noninvasive indicator of anemia, but its diagnostic accuracy remains limited. This study aimed to develop and validate a deep learning system to estimate hemoglobin (Hb) concentration and screen for anemia using palpebral conjunctiva images captured with a smartphone-compatible slit-lamp microscope. Methods: In this prospective observational study, 225 Japanese participants (20–92 years) underwent conjunctival imaging and blood testing. Palpebral conjunctiva videos were obtained using the Smart Eye Camera. Video frames were processed using automated anterior-segment segmentation and conjunctiva extraction. A ConvNeXt-based regression model was trained to predict Hb values. Anemia was defined using sex-specific Hb thresholds. Results: From 225 videos, 53,776 frames were extracted, yielding 9903 quality-filtered conjunctiva images (training: 8082; test: 1821). Video-level predicted Hb values moderately correlated with measured Hb (r = 0.42). For anemia screening, frame-level analysis achieved an AUC of 0.75, with accuracy of 0.76, sensitivity of 0.71, and specificity of 0.79. Video-level aggregation achieved 69% accuracy. Conclusions: Deep learning analysis of palpebral conjunctiva images acquired with a portable slit-lamp microscope demonstrated the feasibility of non-invasive hemoglobin estimation and anemia screening. Although the proposed approach achieved moderate performance, further improvements in model accuracy and prospective multi-center validation are required before clinical implementation.

Objectives: To develop an effective learning method for using a smartphone-based slit-lamp microscope (SBSL) and to identify key points to emphasize when coaching individuals with no prior SBSL experience. Methods: This study included 60 orthoptic students: 40 second-year students (control group: 20, training group 1: 20) and 20 first-year students (training group 2). Subjects were instructed to record the anterior eye segment of a patient-role subject using the Smart Eye Camera. The control group was given paper instruction and was shown the demonstration of the SBSL beforehand. In addition, training groups 1 and 2 watched a tutorial video, practiced using the SBSL for 30 min, and received guidance from an expert. Four ophthalmologists evaluated the recordings based on the eyelid, conjunctiva, cornea, pupil including iris, lens, and anterior chamber depth. Results: ANOVAs showed significant differences among groups for all items. The control group had significantly lower scores than both training groups, while no significant differences were found between training groups 1 and 2. Principal component analysis of training groups 1 and 2 showed that the first principal component accounted for 74.36% of the variance. The second principal component accounted for 10.71%, with a wide range of loadings (anterior chamber depth of 0.7780 to conjunctiva of −0.5585), implying the existence of different favorite focusing depths within subjects. Conclusions: A coaching program consisting of tutorial video learning, a 30 min hands-on trial, and feedback is effective in helping individuals without an ophthalmological background acquire anterior segment imaging skills using SBSL. Comprehensive focusing across the entire anterior segment should also be emphasized.

Background/Objectives: Surgery performed in the steep Trendelenburg position is considered potentially detrimental to ocular structures. This study aimed to evaluate morphological changes in the anterior segment during robot-assisted laparoscopic prostatectomy (RALP). Methods: This was a single-center, prospective observational study involving 60 eyes of 30 consecutive patients undergoing RALP between May and November 2021. Anterior segment images were obtained using a Smart Eye Camera before surgery (supine and awake), during surgery (supine and steep Trendelenburg under anesthesia), and after surgery (supine). Assessed parameters included the iridocorneal angle, the ratio of peripheral anterior chamber depth to peripheral corneal thickness based on Van Herick Plus grading, and pupil diameter. Each parameter in the steep Trendelenburg position under anesthesia was compared with measurements obtained in the supine position under anesthesia and in the awake condition. The primary outcome was the comparison of anterior segment morphological changes between the supine and steep Trendelenburg positions during RALP. Results: Upon transition to the steep Trendelenburg position, anterior chamber depth significantly decreased (p < 0.001), recovering after returning to the supine position. The proportion of eyes classified as having narrowed anterior chambers increased significantly (p < 0.001), with more than 60% showing shallower chambers relative to preoperative measurements. Extreme anterior chamber narrowing occurred in 3 out of 410 intraoperative assessments (1%). Pupils were uniformly constricted under anesthesia. Conclusions: Steep Trendelenburg positioning significantly reduces anterior chamber depth. This morphological alteration may contribute to the marked increase in intraocular pressure observed during RALP.

Background/Objectives: The relative efficacy of 8 mg aflibercept compared to 2 mg in treating neovascular age-related macular degeneration (nAMD) has not been fully established. This study aims to compare the visual and anatomical outcomes of aflibercept 8 mg versus 2 mg in patients with nAMD in both treatment-naïve individuals with no history of anti-VEGF treatment and those previously treated with intravitreal injections. Methods: This retrospective study included 13 eyes treated with aflibercept 8 mg and 14 eyes with aflibercept 2 mg in treatment-naïve patients, along with 15 eyes switched to aflibercept 8 mg previously treated with other intravitreal injections and 15 eyes continued on aflibercept 2 mg in patients. Baseline and one-month post-injection changes in best-corrected visual acuity (BCVA) and central macular thickness (CMT) were assessed. Results: In treatment-naïve patients, the aflibercept 8 mg group showed a significant improvement in BCVA (logMAR 0.19 ± 0.23 to 0.13 ± 0.20, p = 0.0156), while the 2 mg group did not. Both doses reduced CMT significantly, with a greater reduction in the 8 mg group (dCMT 28.60% vs. 24.08%, p = 0.0220). In previously treated patients, no significant changes in BCVA were noted in either group; however, both groups showed significant reductions in CMT. Conclusions: Real-world data demonstrated that aflibercept 8 mg led to substantial improvements in anatomical outcomes one month after injection, irrespective of previous intravitreal injection history. However, significant improvements in visual outcomes were observed exclusively in treatment-naïve patients. Further large-scale, long-term studies are required to determine the proportion of patients who experience improvement and to assess whether these improvements are maintained over time.

Background and aim: Rural and remote areas often face a shortage of ophthalmic services, particularly in aging populations. This study aimed to assess the prevalence of anterior segment diseases using the Smart Eye Camera (SEC), a portable device enabling remote ophthalmic screenings. Materials and methods: Ophthalmic screenings were conducted using the SEC to capture anterior segment videos of 158 eyes from 79 residents of Yoron Island. These videos were analyzed remotely by ophthalmologists, focusing on eyelid conditions, conjunctival abnormalities, anterior chamber depth, and lens status. Results: The study revealed high prevalence rates of ptosis in 24 eyes (15.1%) and cataracts in 108 eyes (68.4%). Among patients over 60 years, cataracts were observed in 71 eyes (84.5%). Additionally, five eyes (3.2%) showed evidence of pterygium, all of which occurred in patients over 60 years of age. Conclusions: Telemedicine using the SEC effectively identified age-related anterior segment diseases in this remote island population. The findings underscore the need for continued use of portable imaging devices in underserved areas to bridge gaps in eye care access.

Laser speckle flowgraphy (LSFG) offers a non-invasive, accessible means to evaluate ocular blood flow. A case of a woman in her 70s presenting with 6 months of ocular injection following cataract surgery is presented in this report. LSFG measurements indicated lower mean blur rate (MBR) values in the affected eye compared with the healthy eye, though a decrease was also noted in the healthy eye. Subsequent MRI/magnetic resonance angiography revealed an abnormally high signal in the right cavernous sinus, along with a dilated right superior ophthalmic vein, suggesting a carotid-cavernous fistula (CCF). The patient underwent transvenous embolisation under general anaesthesia at the neurosurgery department, successfully occluding the shunt point and eliminating abnormal venous flow. After successful transvenous embolisation, MBR values improved bilaterally. This case highlights LSFG’s potential as a simple, non-invasive screening tool for CCF, underscoring the need for further research to confirm its diagnostic utility.

Background: This study aimed to evaluate the impact of image enhancement techniques on the interobserver reliability of tear break-up time (TBUT), tear meniscus height (TMH), corneal fluorescein staining (CFS) scoring, and conjunctivochalasis detection using the Smart Eye Camera (SEC), a portable device for anterior segment examination. Methods: A retrospective analysis was conducted on video recordings captured by the SEC from 46 patients with dry eye disease (DED). Separate sets of images were created for each level of enhancement: unenhanced (G0), mildly enhanced (G3), and strongly enhanced (G7). These sets were not intermixed, ensuring that each enhancement level was assessed independently. Three observers—two DED specialists and one general ophthalmologist—assessed TBUT, TMH, CFS scores, and conjunctivochalasis. Interobserver reliability was evaluated using intraclass correlation coefficients (ICCs) for each image set. Results: Interobserver reliability for CFS scores significantly improved with G3, yielding an ICC of 0.8413. In contrast, G7 improved reliability for TBUT measurements (ICC = 0.7381), but led to a notable decrease in reliability for both CFS scoring (ICC = 0.2259) and conjunctivochalasis detection (ICC = 0.0786). Furthermore, the assessment of TMH demonstrated a progressive decline in accuracy with increasing levels of image enhancement. Conclusions: Image enhancement using the SEC improved the diagnostic consistency of dry eye specialists and general ophthalmologists, especially for TBUT and CFS assessments. However, excessive enhancement may obscure key diagnostic features, indicating the need for careful optimization of image processing techniques depending on the diagnostic focus.