Towards 'automated gonioscopy': a deep learning algorithm for 360° angle assessment by swept-source optical coherence tomography.

AIMS: To validate a deep learning (DL) algorithm (DLA) for 360° angle assessment on swept-source optical coherence tomography (SS-OCT) (CASIA SS-1000, Tomey Corporation, Nagoya, Japan). METHODS: This was a reliability analysis from a cross-sectional study. An independent test set of 39 936 SS-OCT scans from 312 phakic subjects (128 SS-OCT meridional scans per eye) was analysed. Participants above 50 years with no previous history of intraocular surgery were consecutively recruited from glaucoma clinics. Indentation gonioscopy and dark room SS-OCT were performed. Gonioscopic angle closure was defined as non-visibility of the posterior trabecular meshwork in ≥180° of the angle. For each subject, all images were analysed by a DL-based network based on the VGG-16 architecture, for gonioscopic angle-closure detection. Area under receiver operating characteristic curves (AUCs) and other diagnostic performance indicators were calculated for the DLA (index test) against gonioscopy (reference standard). RESULTS: Approximately 80% of the participants were Chinese, and more than half were women (57.4%). The prevalence of gonioscopic angle closure in this hospital-based sample was 20.2%. After analysing a total of 39 936 SS-OCT scans, the AUC of the DLA was 0.85 (95% CI:0.80 to 0.90, with sensitivity of 83% and a specificity of 87%) to classify gonioscopic angle closure with the optimal cut-off value of >35% of circumferential angle closure. CONCLUSIONS: The DLA exhibited good diagnostic performance for detection of gonioscopic angle closure on 360° SS-OCT scans in a glaucoma clinic setting. Such an algorithm, independent of the identification of the scleral spur, may be the foundation for a non-contact, fast and reproducible 'automated gonioscopy' in future.

Evaluation of meridional scans for angle closure assessment with anterior segment swept-source optical coherence tomography.

BACKGROUND/AIMS: As swept-source optical coherence tomography (SS-OCT) simultaneously obtains 128 meridional scans, it is important to identify which scans are playing the main role in classifying gonioscopic angle closure to simplify the analysis. We aimed to evaluate the diagnostic performance of every meridional scan in its ability to detect gonioscopic angle closure. METHODS: Observational study with 2027 phakic subjects consecutively recruited from a community polyclinic. Gonioscopy and SS-OCT were performed. Gonioscopic angle closure was defined as non-visibility of the posterior trabecular meshwork in ≥180° of the angle, while SS-OCT was defined as iridotrabecular contact anterior to the scleral spur. The area under the receiver operating characteristic curve (AUC) was calculated to assess the diagnostic performance of each single scan, the sequential anticlockwise cumulative effect of those single scans and different combinations of them. RESULTS: The AUCs of each scan ranged from 0.73 to 0.82. The single scan at 80°-260° had the highest AUC (0.82, 95% CI 0.79 to 0.84) and performed significantly better than most of the temporonasal scans (from 0° to 52° and from 153° to 179°). The superoinferior scans achieved higher AUCs compared with the temporonasal ones. When assessing the cumulative effect of adding individual scans consecutively, the peak AUC (0.80) was obtained when considering the superoinferior scans closer to 80°-85°, but no further positive cumulative effect was seen when adding the rest of the temporonasal scans of the circumference. CONCLUSIONS: In conclusion, the single SS-OCT scan at 80°-260° had the highest diagnostic performance. Our study suggests that the 360° evaluation may not translate to better clinical utility for detection of gonioscopic angle closure.

Understanding diagnostic disagreement in angle closure assessment between anterior segment optical coherence tomography and gonioscopy.

BACKGROUND/AIMS: Although being a more objective tool for assessment and follow-up of angle closure, reliability studies have reported a moderate diagnostic performance for anterior segment optical coherence tomography (OCT) technologies when comparing with gonioscopy as the reference standard. We aim to determine factors associated with diagnostic disagreement in angle closure when assessed by anterior segment swept source OCT (SS-OCT, CASIA SS-1000; Tomey, Nagoya, Japan) and gonioscopy. METHODS: Cross-sectional study. A total of 2027 phakic subjects aged ≥50 years, with no relevant previous ophthalmic history, were consecutively recruited from a community polyclinic in Singapore. Gonioscopy and SS-OCT (128 radial scans) for the entire circumference of the angle were performed for each subject. A two-quadrant closed gonioscopic definition was used. On SS-OCT images, angle closure was defined as iridotrabecular contact (ITC) to the extent of ≥35%, ≥50% and ≥75% of the circumferential angle. Diagnostic disagreements between both methods, that is, false positives or overcalls and false negatives or undercalls were defined, respectively, as gonioscopic open/closed angles inversely assessed as closed/open by SS-OCT. RESULTS: Two hundred and seventy-two (14.7%) resulted in overcall results (false positives) when ≥50% of the angle circumference was closed using SS-OCT. These eyes had significantly wider (anterior chamber width, 11.7 vs 11.6 mm, p<0.001) and deeper (anterior chamber depth (ACD), 2.4 vs 2.2 mm, p<0.001) anterior chambers than eyes assessed by both methods as closed (true positives). Deeper ACD (OR 9.31) and lower lens vault (LV) (OR 0.04) were significantly associated with a false positive diagnosis in the multivariate analysis. Most of these cases had short (52.6%) or irregular (39%) ITC in SS-OCT images. CONCLUSIONS: We found that anterior chamber dimensions, determined by ACD and LV, were factors significantly associated with diagnostic disagreement between anterior segment SS-OCT and gonioscopy in angle closure assessment.

Assessment of Circumferential Angle Closure with Swept-Source Optical Coherence Tomography: a Community Based Study.

PURPOSE: To evaluate the diagnostic performance of swept-source optical coherence tomography (SS-OCT, CASIA SS-1000; Tomey Corporation, Nagoya, Japan) for angle closure detection, in comparison with gonioscopy, in a community setting. DESIGN: Reliability analysis. METHODS: A total of 2027 phakic subjects aged ≥50 years, with no previous history of glaucoma, laser (including peripheral iridotomy), intraocular surgery, or ocular trauma, were consecutively recruited from a community polyclinic in Singapore. Gonioscopy was performed by a single trained ophthalmologist. SS-OCT angle scans, which obtain radial scans for the entire circumference of the angle, were analyzed by a single examiner, masked to the subject's clinical details. On SS-OCT images, angle closure was defined as contact between the iris and any part of the angle wall anterior to the scleral spur. Different cutoff values of the degree of circumferential angle closure (≥35%, ≥50%, and ≥75%) were taken for analysis to assess SS-OCT performance in detecting angle closure. RESULTS: A total of 1857 subjects (91.6%) were included in the final analysis after excluding poor-quality SS-OCT scans. Almost 90% of the subjects were Chinese, with a mean age of 61.8 ± 6.7 years, and more than half were women (63.5%). The overall AUC of SS-OCT manual grading against gonioscopy was 0.84 (95% confidence interval, 0.81-0.88). The prevalence of angle closure on SS-OCT was 26.1% for the ≥35% definition, with an area under the curve of 0.80 (0.77-0.84), sensitivity of 82.5% (75.3%-88.4%), and specificity of 78.5% (76.5%-80.4%). The first-order agreement coefficient statistics for the 2-quadrant gonioscopic definition of angle-closure with corresponding ≥35%, ≥50%, and ≥75% angle closure definitions for SS-OCT were good at 0.89 (0.83-0.93), 0.88 (0.842-0.93), and 0.88 (0.831-0.99), respectively. CONCLUSIONS: In this large community-based study, SS-OCT exhibited moderate performance for angle closure detection compared to gonioscopy as the reference standard.

Segmentation and Quantification for Angle-Closure Glaucoma Assessment in Anterior Segment OCT.

Angle-closure glaucoma is a major cause of irreversible visual impairment and can be identified by measuring the anterior chamber angle (ACA) of the eye. The ACA can be viewed clearly through anterior segment optical coherence tomography (AS-OCT), but the imaging characteristics and the shapes and locations of major ocular structures can vary significantly among different AS-OCT modalities, thus complicating image analysis. To address this problem, we propose a data-driven approach for automatic AS-OCT structure segmentation, measurement, and screening. Our technique first estimates initial markers in the eye through label transfer from a hand-labeled exemplar data set, whose images are collected over different patients and AS-OCT modalities. These initial markers are then refined by using a graph-based smoothing method that is guided by AS-OCT structural information. These markers facilitate segmentation of major clinical structures, which are used to recover standard clinical parameters. These parameters can be used not only to support clinicians in making anatomical assessments, but also to serve as features for detecting anterior angle closure in automatic glaucoma screening algorithms. Experiments on Visante AS-OCT and Cirrus high-definition-OCT data sets demonstrate the effectiveness of our approach.

Swept-source optical coherence tomography assessment of iris-trabecular contact after phacoemulsification with or without goniosynechialysis in eyes with primary angle closure glaucoma.

AIMS: To compare the change in iris-trabecular contact (ITC) area using swept-source optical coherence tomography (SS-OCT) in eyes with primary angle closure glaucoma (PACG) and cataract that underwent phacoemulsification (PE) with intraocular lens implantation alone compared with PE with goniosynechialysis (GSL). METHODS: One eye of 22 patients with PACG with peripheral anterior synechiae (PAS) detected by indentation gonioscopy was randomised into two groups (PE alone (n=11) and PE+GSL (n=11)). The anterior chamber angles were evaluated by SS-OCT under dark conditions before and 12 months after surgery using the three-dimensional angle analysis scan protocol that simultaneously obtains 128 cross-sectional radial scans across the anterior chamber at equal intervals (every 1.4°). The ITC area, defined as the area of extent of the circumferential contact of peripheral iris to the angle wall, was computed automatically by SS-OCT after an observer marked the scleral spurs of all 128 scans of each eye. RESULTS: The majority of the 22 subjects were women (77.3%) and the mean±SD age was 67.3±5.8 years. The ITC area was significantly reduced in the PE+GSL group compared with the PE alone group (10.2 mm2 vs. 4.6 mm2, ß=0.54, p=0.03) after adjusting for age, gender, intraocular pressure, extent of PAS and pupil diameter before surgery. Smaller iris volume at baseline was associated with greater ITC area reduction by PE+GSL (ß=-0.728, p=0.03). CONCLUSIONS: Eyes that undergo PE+GSL surgery have a greater reduction in circumferential ITC area than eyes that undergo PE alone.

Assessment of circumferential angle-closure by the iris-trabecular contact index with swept-source optical coherence tomography.

PURPOSE: To evaluate the diagnostic performance of the iris-trabecular contact (ITC) index, a measure of the degree of angle-closure, using swept-source optical coherence tomography (SSOCT, CASIA SS-1000, Tomey Corporation, Nagoya, Japan) in comparison with gonioscopy. DESIGN: Prospective observational study. PARTICIPANTS: A total of 108 normal subjects and 32 subjects with angle-closure. METHODS: The SSOCT 3-dimensional angle scans, which obtain radial scans for the entire circumference of the angle, were performed under dark conditions and analyzed using customized software by a single examiner masked to the subjects' clinical details. The ITC index was calculated as a percentage of the angle that was closed on SSOCT images. First-order agreement coefficient (AC1) statistics and area under the receiver operating characteristic curve (AUC) analyses were performed for angle-closure on the basis of the ITC index in comparison with gonioscopy. MAIN OUTCOME MEASURES: Angle-closure on gonioscopy was defined as nonvisibility of posterior trabecular meshwork for at least 2 quadrants. Agreement of the ITC index with gonioscopically defined angle-closure was assessed using the AC1 statistic. RESULTS: Study subjects were predominantly Chinese (95.7%) and female (70.7%), with a mean age of 59.2 (standard deviation, 8.9) years. The median ITC index was 15.24% for gonioscopically open-angle eyes (n = 108) and 48.5% for closed-angle eyes (n = 32) (P = 0.0001). The agreement for angle-closure based on ITC index cutoffs (>35% and ≥50%) and gonioscopic angle-closure was 0.699 and 0.718, respectively. The AUC for angle-closure detection using the ITC index was 0.83 (95% confidence interval, 0.76-0.89), with an ITC index >35% having a sensitivity of 71.9% and specificity of 84.3%. CONCLUSIONS: The ITC index is a summary measure of the circumferential extent of angle-closure as imaged with SSOCT. The index had moderate agreement and good diagnostic performance for angle-closure with gonioscopy as the reference standard.

Assessment of trabecular meshwork width using swept source optical coherence tomography.

PURPOSE: Measurements of the angle width by ultrasound biomicroscopy or anterior segment optical coherence tomography are usually performed 500 µm from the scleral spur, as the anterior part of trabecular meshwork (TM) is assumed to lie within this distance. The aim of this study was to measure TM width using swept source optical coherence tomography (SS-OCT, CASIA SS-1000, Tomey Corporation, Nagoya, Japan), and to investigate factors influencing this measurement. METHODS: Participants underwent gonioscopy and SS-OCT imaging in the dark. High-definition SS-OCT images were corrected for refractive distortion; and customized software (ImageJ; National Institutes of Health, Bethesda, MD, USA) was utilized to measure TM width (distance between the scleral spur and Schwalbe's line). Linear regression analysis was performed to assess the relationship between TM width with demographic and angle parameters. RESULTS: One hundred and forty eight Chinese subjects were analyzed. The majority was female (62.4 %); the mean age was 59.2 ± 8.68 years. Identification of the scleral spur and Schwalbe's line with SS-OCT was possible in 590 (99.7 %) and 585 angle quadrants (98.8 %) respectively. TM width was wider in the inferior and superior quadrants (mean 889 [SD 138] and 793 [136] µm), compared to the nasal and temporal quadrants (712 [137] and 724 [115] µm, P<0.001). There was a difference in average TM width between open (789 [100]) and closed angle eyes (753 [86]) (P=0.048). There was no significant association between TM width and angle parameters, laterality, or demographic factors. CONCLUSIONS: In SS-OCT HD images, the mean TM width varied from 710 to 890 µm in the different quadrants of the eye, and the inferior quadrant TM was the widest compared to other quadrants.

Comparison of two spectral domain optical coherence tomography devices for angle-closure assessment.

PURPOSE: To compare two spectral domain optical coherence tomography (SD-OCT) devices for the identification of angle structures and the presence of angle closure. METHODS: This was a prospective comparative study. Consecutive patients underwent gonioscopy and anterior segment imaging using two SD-OCT devices (iVue and Cirrus). Images were evaluated for the ability to detect angle structures such as Schwalbe's line (SL), trabecular meshwork (TM), Schlemm's canal (SC), and scleral spur (SS), and the presence of angle closure. Angle closure was defined as iris contact with the angle wall anterior to the SS on SD-OCT, and nonvisibility of the posterior TM on gonioscopy. Angle closure in an eye was defined as ≥two quadrants of closed angles. AC1 statistic was used to assess the agreement between devices. RESULTS: Of the 69 subjects studied (46.4% male, 84.1% Chinese, mean age 64.0 ± 10.5 years), 40 subjects (40 eyes, 58.0%) had angle closure on gonioscopy. The most identifiable structure on Cirrus SD-OCT was the SS (82.2%) and SL on iVue SD-OCT (74.5%). Angle closure was indeterminable in 14.5% and 50.7% of Cirrus and iVue scans (P < 0.001), respectively. Interdevice agreement for angle closure was moderately strong (AC1 = 0.67), but agreement with gonioscopy was only fair (AC1 = 0.35 and 0.50 for Cirrus and iVue, respectively). CONCLUSIONS: It was more difficult to determine angle closure status with iVue compared with Cirrus SD-OCT. There was fair agreement between both devices with gonioscopy for identifying angle closure.

Angle assessment by EyeCam, goniophotography, and gonioscopy.

PURPOSE: To compare EyeCam (Clarity Medical Systems, Pleasanton, CA) and goniophotography in detecting angle closure, using gonioscopy as the reference standard. METHODS: In this hospital-based, prospective, cross-sectional study, participants underwent gonioscopy by a single observer, and EyeCam imaging and goniophotography by different operators. The anterior chamber angle in a quadrant was classified as closed if the posterior trabecular meshwork could not be seen. A masked observer categorized the eyes as per the number of closed quadrants, and an eye was classified as having angle closure if there were 2 or more quadrants of closure. Agreement between the methods was analyzed by κ statistic and comparison of area under receiver operating characteristic curves (AUC). RESULTS: Eighty-five participants (85 eyes) were included, the majority of whom were Chinese. Angle closure was detected in 38 eyes (45%) with gonioscopy, 40 eyes (47%) using EyeCam, and 40 eyes (47%) with goniophotography (P=0.69 in both comparisons, McNemar test). The agreement for angle closure diagnosis (by eye) between gonioscopy and the 2 imaging modalities was high (κ=0.86; 95% Confidence Interval (CI), 0.75-0.97), whereas the agreement between EyeCam and goniophotography was not as good (κ=0.72; 95% CI, 0.57-0.87); largely due to lack of agreement in the nasal and temporal quadrants (κ=0.55 to 0.67). The AUC for detecting eyes with gonioscopic angle closure was similar for goniophotography and EyeCam (AUC 0.93, sensitivity=94.7%, specificity=91.5%; P>0.95). CONCLUSIONS: EyeCam and goniophotography have similarly high sensitivity and specificity for the detection of gonioscopic angle closure.

Automatic anterior chamber angle assessment for HD-OCT images.

Angle-closure glaucoma is a major blinding eye disease and could be detected by measuring the anterior chamber angle in the human eyes. High-definition OCT (Cirrus HD-OCT) is an emerging noninvasive, high-speed, and high-resolution imaging modality for the anterior segment of the eye. Here, we propose a novel algorithm which automatically detects a new landmark, Schwalbe's line, and measures the anterior chamber angle in the HD-OCT images. The distortion caused by refraction is corrected by dewarping the HD-OCT images, and three biometric measurements are defined to quantitatively assess the anterior chamber angle. The proposed algorithm was tested on 40 HD-OCT images of the eye and provided accurate measurements in about 1 second.

National survey of ophthalmologists in Singapore for the assessment and management of asymptomatic angle closure.

PURPOSE: Approximately 10% of Chinese people older than 50 years in Singapore have asymptomatic narrow angles [also termed primary angle closure suspects (PACS)]. The aim of this study was to determine the attitudes and practices of ophthalmologists in Singapore regarding the management of these patients. METHODS: A nation wide questionnaire-based survey was conducted on all registered, practicing ophthalmologists in Singapore by electronic mail and post. An ophthalmologist was defined as a doctor who had completed at least 3 years of specialist training in ophthalmology and had passed the Master's Degree in Ophthalmology or equivalent examinations. The survey asked the ophthalmologists about the methods and criteria used for the diagnosis of angle closure. It also inquired about the management of PACS and opinions on the effectiveness of prophylactic laser peripheral iridotomy (LPI) in preventing both acute and chronic angle closure. RESULTS: A total of 126 out of 158 ophthalmologists responded (79.7% response rate). Of the respondents, 84.9% would advise prophylactic LPI for asymptomatic PACS; 84.9% believed that prophylactic LPI would prevent acute angle closure but only 44% thought that it could prevent the development of glaucomatous optic nerve damage. The preferred method of prophylactic LPI was sequential argon-YAG LPI. In assessing patients for angle closure, 85.4% used gonioscopy, 92.9% assessed anterior chamber depth, and 17.5% performed provocative tests. CONCLUSIONS: Ophthalmologists in Singapore vary in the method of assessment of patients with angle closure. Most believe that prophylactic LPI should be performed for asymptomatic PACS.

Comparison of the scanning peripheral anterior chamber depth analyzer and the modified van Herick grading system in the assessment of angle closure.

PURPOSE: To compare the performance of the scanning peripheral anterior chamber depth analyzer (SPAC) and the modified van Herick grading system in the assessment of angle closure, using gonioscopy as the reference standard. DESIGN: Prospective comparative study. PARTICIPANTS: One hundred twenty phakic subjects (53 with primary angle closure and 67 with open angles) with no history of laser or intraocular surgery. METHODS: The modified van Herick grading system was used to grade the peripheral anterior chamber depth (ACD) clinically. The SPAC was used to assess ACD, and this was graded categorically as S (suspect angle closure), P (potential angle closure), or N (normal) and numerically from 1 to 12. One randomly selected eye of each subject underwent examination with the modified van Herick grading, SPAC, Goldmann applanation tonometry, optic disc examination, and gonioscopy. The SPAC measurements, modified van Herick grades (0%, 5%, 15%, 25%, 40%, 75%, and > or =100% of corneal thickness), and gonioscopy (graded as narrow or open) were correlated using the Spearman correlation coefficient. The receiver operating characteristics of the 2 methods to detect angle closure also were compared. MAIN OUTCOME MEASURES: Grading of ACD by the SPAC, the modified van Herick system, and gonioscopic grading of the angle. RESULTS: The SPAC results correlated well with the modified van Herick grading system (categorical grade, r = 0.527; numerical grade, r = 0.542; P<0.0001). For eyes graded as having narrow angles by gonioscopy, the area under the curve (AUC) for SPAC categorical grade S or P was 0.790; sensitivity and specificity were 84.9% and 73.1%, respectively. For the modified van Herick grading system, using a cutoff of peripheral ACD < or = 25% corneal thickness, the AUC was 0.872 and sensitivity and specificity were 84.9% and 89.6%. Compared with gonioscopy that found 53 of 120 cases of narrow angles, the SPAC graded more eyes as having narrow angles (63/120) than the modified van Herick system (52/120). CONCLUSIONS: The SPAC correlated well with the modified van Herick system in grading peripheral ACD. However, the SPAC appeared to overestimate the proportion of eyes with narrow angles relative to gonioscopy and the modified van Herick grading system.

Anterior chamber angle assessment using gonioscopy and ultrasound biomicroscopy.

PURPOSE: Comparison of anterior chamber angle measurements using ultrasound biomicroscopy (UBM) and gonioscopy. METHODS: Five hundred subjects were evaluated for grading of angle width by the Shaffer method. UBM was done in the same group to document angle width, angle opening distance (AOD 500), and anterior chamber depth. Biometric parameters were documented in all subjects. UBM and gonioscopic findings were compared. RESULTS: A study was conducted in 282 men and 218 women with a mean age of 57.32 +/- 12.48 years. Gonioscopic grading was used to segregate occludable (slit-like, grades 1 and 2) from nonoccludable (grades 3 and 4) angles. Subjective assessment by gonioscopy resulted in an overestimation of angle width within the occludable group when compared with values obtained by UBM. This did not affect the segregation of occludable versus nonoccludable angles by gonioscopy. Biometric parameters in eyes with occludable angles were significantly lower in comparison with eyes with nonoccludable angles, except for lens thickness. AOD 500 correlated well with angle width. CONCLUSIONS: We concluded that clinical segregation into occludable and nonoccludable angles by an experienced observer using gonioscopy is fairly accurate. However, UBM is required for objective quantification of angles, and AOD 500 can be a reliable and standard parameter to grade angle width.