To evaluate the diagnostic performance of swept-source anterior segment optical coherence tomography (SS-OCT) in differentiating eyes with primary angle closure disease (PACD) from eyes of control subjects, as well as eyes with PAC and PAC glaucoma (PACG) from eyes with PAC suspect (PACS) disease.
Multicenter cross-sectional study.
Chinese patients were classified into control, PACS, and PAC/PACG groups. The area under the receiving operating characteristic curve (AUC) from logistic regression models was used to evaluate discriminating ability. Sensitivity and specificity were calculated, and performance of the models was validated using an independent dataset.
A total of 2928 SS-OCT images from 366 eyes of 260 patients were recruited to develop diagnostic models. The validation dataset included 1176 SS-OCT images from 147 eyes of 143 patients. For distinguishing PACD from control eyes, average anterior chamber depth had the highest AUC (0.94). With a cutoff of 2.2 mm for average anterior chamber depth, the sensitivity and specificity were 90.2% and 85.2% in the training set. For distinguishing PAC/PACG from PACS, a multivariate model had an AUC of 0.83, with sensitivity and specificity of 82.0% and 62.8% in the training set. The validation set confirmed the findings.
SS-OCT of the anterior segment showed excellent diagnostic performance distinguishing PACD from normal eyes and moderate diagnostic ability distinguishing eyes with PAC/PACG from eyes with PACS. ACD alone may provide a simple and effective way to diagnose PACD from control subjects. As ACD can be obtained using other more available modalities, this has implications for the early diagnosis of PACD.
P rimary angle closure disease (PACD) is clinically classified into subgroups, including primary angle closure suspect (PACS), prinary angle clsoure (PAC), and primary angle clsoure glaucoma (PACG). In general, PACS is the benign form of PACD, but patients with PACS may progress to PAC and PACG , or develop a vision-threatening condition known as acute PAC (APAC). PACD is most common in people of Asian descent, though it can occur in other ethnic groups and is commonly underdiagnosed. , A delayed or missed diagnosis of PACD, especially of PAC or PACG, can lead to irreversible vision loss, emphasizing the importance of early identification and treatment of these patients.
Anterior segment anatomy, particularly the angle configuration, is the most important component in diagnosing PACD, with gonioscopy accepted as the criterion standard. Gonioscopy requires advanced clinical training, is subjective, and may vary between examiners. In contrast, swept-source optical coherence tomography (SS-OCT) of the anterior segment is a noncontact cross-sectional imaging modality that provides high-resolution quantitative images of anterior segment landmarks and parameters. Compared with a single scan obtained using traditional anterior segment OCT (AS-OCT), SS-OCT can image 360° of the anterior segment, obtaining 128 circumferential scans within seconds.
Traditional AS-OCT has been studied as a tool to distinguish PACD from control subjects. One study evaluated the diagnostic ability of SS-OCT to identify PACD and concluded that anterior chamber volume had the highest diagnostic value; however, anterior chamber volume is not easy to obtain and may not be practical in routine clinical practice. Moreover, the diagnostic ability of SS-OCT to classify PACD has not been explored. The aim of this study was to further evaluate the diagnostic performance of SS-OCT in differentiating PACD from control eyes and then classifying PACD into PAC/PACG and PACS. The latter is important because patients with PAC/PACG are at high risk for PACD, and these patients require timely treatment. SS-OCT classification could serve as an objective and noninvasive method to help identify patients with PACD who are at risk of vision loss, particularly in areas lacking eye professionals.
STUDY POPULATION AND RECRUITMENT
This was a multicenter cross-sectional study approved by the Institutional Review Boards of Zhongshan Ophthalmic Center, Beijing Tongren Hospital, and the University of California, San Francisco. The study adhered to the tenets of the Declaration of Helsinki and informed consent was obtained from all patients. The study design included establishing a “training” dataset then validating the findings on a separate “validation” dataset.
Consecutive adult patients evaluated by the Glaucoma Service at Zhongshan Ophthalmic Center between October 2018 and June 2019 were enrolled to develop an SS-OCT diagnostic model. The model was then tested on a validation set of patients seen by the Glaucoma Service at Beijing Tongren Hospital in December 2019. Patients were excluded if they had any of the following: previous iris laser treatment or intraocular surgery, history of secondary angle closure, previous episode of APAC, history of trauma, congenital abnormalities, or other ocular disease apart from cataract. Both eyes from each participant were enrolled if they fulfilled both met inclusion and exclusion criteria.
PATIENT ASSESSMENT FOR PACD
After obtaining medical and ocular histories, all patients underwent a standardized ophthalmic examination including measuring best-corrected visual acuity, intraocular pressure (IOP) with Goldmann applanation, slit-lamp biomicroscopy, gonioscopy, and optic disc assessment using a 90-diopter lens at the slit lamp. Optic nerve spectral-domain OCT and a Humphrey visual field test were performed in addition to anterior segment SS-OCT (CASIA SS-1000).
Gonioscopy was performed by an experienced glaucoma specialist (HY, YY, SL) masked to SS-OCT findings using a 4-mirror Zeiss gonioprism (Volk Optical Inc, Mentor, Ohio, USA). The examination was performed in a dark environment using a 1-mm light beam not passing through the pupil to avoid inducing pupillary constriction. If the trabecular meshwork was not visible, dynamic gonioscopy was performed to assess for peripheral anterior synechiae. Based on this clinical examination, patients were divided into the control, PACS, or PAC/PACG groups.
PACD included patients with PACS and PAC/PACG, all of whom had nonvisualization of the posterior trabecular meshwork for ≥180° on gonioscopy. PACS was defined if there was lack of IOP elevation, peripheral anterior synechiae, or glaucomatous damage (vertical cup:disc ratio ≤0.6, interocular vertical cup:disc ratio difference <0.2, no optic disc hemorrhages, no optic nerve rim loss or notch, no retinal nerve fiber layer defects on OCT, and no Humphrey visual field defects). PAC/PACG eyes had presence of IOP >21 mmHg, peripheral anterior synechiae, or glaucomatous optic nerve damage. Control subjects were those with normal gonioscopy without other intraocular diseases apart from cataract.
SS-OCT IMAGING AND IMAGE ANALYSIS
SS-OCT anterior segment imaging was performed by a masked experienced examiner in a dark environment (approximately <1 lux) before the patients received any pupil dilation or constriction medications. The upper and lower eyelids were gently retracted, taking care to avoid inadvertent pressure on the globe. Each eye was scanned using the 3-dimensional angle analysis model, taking 128 consecutive meridional scans, each consisting of 512 A-scans across the anterior chamber.
Four cross-sectional images, every 45° starting from 0° (0° and 180°, 45° and 225°, 90° and 270°, and 135° and 315°), were analyzed with the Tomey Software (version 7M.1) as previously described. , For each image, the scleral spur and angle recess (AR) were automatically marked by the SS-OCT system and then manually corrected by masked experienced observers (M.P., J.P.). The observers were trained by a glaucoma specialist (YH.) and demonstrated ≥90% agreement with the specialist in the labeling of prespecified scans. Anterior chamber parameters collected from SS-OCT images are shown in Figure 1 with detailed description 12 : 1) volume parameters: cornea volume (CorVol), anterior chamber volume (ACVol), and iris volume (IrisVol); 2) dimension parameters: anterior chamber depth (ACD), anterior chamber width (ACW), and lens vault (LV); and 3) angle parameters: angle opening distance (AOD), AR area (ARA), trabecular-iris space area (TISA), and trabecular-iris angle (TIA). AOD, ARA, TISA, and TIA were all assessed 250, 500, and 750 μm from the scleral spur.
Demographic data across the control, PACS, and PAC/PACG groups were compared using analysis of variance for means and the χ 2 test for proportions. To compare IOP and SS-OCT parameters across 3 groups, we used generalized linear models, and generalized estimating equations were used to account for intereye correlation among patients with 2 eligible eyes. To account for increasing disease severity from the control, PACS, and PAC/PACG groups, linear trend P value was calculated for comparison across these 3 groups by assigning an equally spaced increasing scale to them.
We performed univariable and multivariable logistic regression modeling to evaluate the association between clinical and SS-OCT findings in PACD. From univariable logistic regression models, we calculated the area under the receiving operating characteristic curve (AUC) to summarize the diagnostic power of each clinical and SS-OCT parameter in distinguishing between groups. The sensitivity and specificity for discriminating PACD using various SS-OCT parameters were calculated to determine the optimal cutoff point that provides sufficiently high sensitivity and specificity. For the multivariable logistic regression modeling, only variables with < .05 were kept in in the final multivariable model for calculating the AUC, sensitivity, and specificity. To account for intereye correlation, the clustered bootstrap method was used to derive the 95% confidence intervals (CIs) for AUC, sensitivity, and specificity. All statistical analyses were performed using SAS software (SAS Institute Inc), and 2-sided P < .05 without correcting for multiple comparisons was considered statistically significant.
The training dataset included 2928 SS-OCT images from 366 eyes of 260 patients: 152 control eyes, 86 eyes with PACS, and 128 eyes with PAC/PACG ( Table 1 ), among which data collected from 40 eyes with PACD were also used in another study (ie, an additional 174 eyes with PACD were newly added to the current study to build the statistical model). All participants were Chinese. The mean (± SD) age was 59.8 ± 13.1 years with no significant age difference across groups (trend P = .81), though there were more females in the PACS and PAC/PACG groups compared with the control subjects (77.9% vs 71.9% vs 53.3%; trend P = .009). Mean IOP (± SD) was 14.2 ± 3.3 mmHg in the control group, 14.2 ± 3.2 mmHg in the PACS group, and 20.5 ± 10.1 mmHg in the PAC/PACG group, with a significant difference across groups (trend P < .001) as well as between the PACS and PAC/PACG groups ( P < .001).
|Control||PACS||PAC/PACG||Linear Trend P Value for Control, PACS, and PAC/PACG||P Value for PACS vs PAC/PACG|
|Training set (eyes), N||152||86||128|
|Mean (SD)||59.4 (16.0)||61.9 (8.3)||58.8 (11.7)|
|Sex, n (%)||.009||.30|
|Female||81 (53.3)||67 (77.9)||92 (71.9)|
|Male||71 (46.7)||19 (22.1)||36 (28.1)|
|Mean (SD)||14.2 (3.3)||14.2 (3.2)||20.5 (10.1)|
|Validation set (eyes), N||56||64||27|
|Mean (SD)||60.5 (7.0)||64.2 (6.5)||66.3 (6.8)|
|Sex, n (%)||.94||.54|
|Female||40 (71.4)||43 (67.2)||20 (74.1)|
|Male||16 (28.6)||21 (32.8)||7 (25.9)|
|Mean (SD)||16.4 (1.9)||16.8 (1.9)||18.7 (3.9)|
The validation dataset included 1176 SS-OCT images from an independent sample of 143 Chinese patients: 56 control eyes, 64 eyes with PACS, and 27 eyes with PAC/PACG ( Table 1 ). As expected, mean IOP was significantly different across the 3 groups (trend P = .001).
COMPARISON OF SS-OCT PARAMETERS IN CONTROL, PACS, AND PAC/PACG GROUPS
Volume measurements including ACVol, CorVol, and IrisVol were all significantly different across 3 groups ( Table 2 ; trend P < .001). CorVol and IrisVol were lower in the PAC/PACG group compared with the PACS group ( P = .002 and P = .01), but there was no difference in ACVol between the 2 groups ( P = .92).
|SS-OCT Parameters||Control, Mean (SD), N = 152||PACS, Mean (SD), N = 86||PAC/PACG, Mean (SD),N = 92||Linear Trend P Value for Control, PACS, and PAC/PACG||P Value forPACS vs PAC/PACG|
|ACVol (mm 2 )||132.98 (40.39)||72.58 (18.25)||72.84 (19.28)||<.001||.92|
|CorVol (mm 2 )||117.98 (16.43)||115.37 (12.62)||108.84 (13.43)||<.001||.002|
|IriVol (mm 2 )||34.13 (5.49)||31.06 (4.89)||29.14 (5.51)||<.001||.01|
|ACD_mean (mm)||2.65 (0.43)||1.89 (0.26)||1.88 (0.28)||<.001||.93|
|ACW_mean (mm)||11.59 (0.43)||11.18 (0.48)||11.03 (0.53)||<.001||.04|
|LV_mean (mm)||0.38 (0.35)||0.90 (0.22)||0.80 (0.22)||<.001||.003|
|AOD250_mean (mm)||0.22 (0.12)||0.09 (0.05)||0.06 (0.05)||<.001||<.001|
|TIA250_mean (°)||25.00 (12.50)||11.11 (7.90)||6.45 (6.14)||<.001||<.001|
|ARA250_mean (mm 2 )||0.07 (0.04)||0.04 (0.02)||0.02 (0.02)||<.001||<.001|
|TISA250_mean (mm 2 )||0.05 (0.02)||0.02 (0.01)||0.01 (0.01)||<.001||<.001|