The diagnostic performance of optical coherence tomography and optical coherence tomography angiography for early primary open-angle glaucoma: a cross-sectional study CURRENT STATUS: POSTED

Background: Primary open angle glaucoma (POAG) is often delayed in diagnosis and treatment, as it has no acute attack period, and symptoms usually appear at a relatively late stage. Optical coherence tomography angiography (OCTA) on the basis of Carl Zeiss optical coherence tomography with the support of FOURM platform is recently available in China. It is reportedly more sensitive to POAG. We reported the diagnostic utilities of OCTA and OCT which is already in clinical practice. Methods: 15 patients (23 eyes) with early POAG as observation group and 30 health people (30 eyes) as normal control group were enrolled in this cross-sectional study. OCTA-based superficial macula vessel density, superficial macula perfusion density, superficial optic disc vessel density, superficial optic disc perfusion density and OCT-based macula thickness, ganglion cell complex (GCC) thickness and retinal nerve fiber layer (RNFL) thickness were recorded in the two groups. Independent sample t-test and receiver operating characteristic curve were used for statistical analysis. Area under the receiver operating characteristic curves (AUCs) were used to measure the diagnostic utility. Results: The optimal diagnostic utility parameters were the superficial vessel density and perfusion density in the macula (except the center of macula), and the AUCs were above 0.97. Followed by the superficial vessel density and perfusion density in the optic disc area. The OCT-based diagnostic utility parameters were generally lower than that mentioned above, the top three parameters were the inferior RNFL thickness (AUC=0.919), the superior GCC thickness (AUC=0.919) and the inferior GCC thickness (AUC = 0.908). Conclusions: Changes of thickness in macula, GCC, RNFL, and changes of superficial vessel density and perfusion density in the optic disc or in macula can be detected by OCT or by OCTA in early POAG. The OCTA-based diagnostic utility was generally higher than the OCT-based diagnostic utility. optical micro-angiography (OMAG) used to capture the dynamic motion of the red blood cells and provide a high-resolution 3D visualization of perfused retinal vasculature. The macular area thickness and the GCC thickness were scanned using the Macular Cube 512×128 procedure, the RNFL thickness scanning was performed using the Optic Disc Cube 200×200 procedure, and the optic disc and macular area OCTA scanning were performed using the Angiography mm procedure. The thickness of the RNFL was calculated by the ONH and RNFL

The diagnostic performance of optical coherence tomography and optical coherence tomography angiography for early primary open-angle glaucoma: a cross-sectional study

KEYWORDS
primary open angle glaucoma, optical coherence tomography, optical coherence tomography angiography, diagnostic utility relatively late stage [2,3]. Therefore, early diagnosis and treatment of glaucoma as well as close monitoring are crucial for POAG. The etiology and pathogenesis of POAG are not well understood so far. Studies have shown that patients with POAG had a sharply decreased diastolic blood flow of the ophthalmic artery, a significant decreased systolic and diastolic blood flow velocity of the posterior ciliary artery, and an obviously increased resistance index (RI), all of which indicated that the POAG patients had a short supply of blood flow to the posterior ciliary artery, furtherly suggesting insufficient supply to optic disc and choroid. Meanwhile, The systolic and diastolic blood flow velocities of the central retinal artery (CRA) were also significantly decreased, diastolic blood flow of the part of individuals even zero, similarly indicating that the central retinal artery of patients with POAG had a low flow velocity and high resistance state, and the blood supply to the surface of the optic disc and to the inner layer of the retina was obviously insufficient [4]. Optical coherence tomography angiography (OCTA) is an imaging technology developed on the basis of spectral domain OCT (SD-OCT), which can detect the retinal microvascular noninvasively in a short time without mydriasis. In addition, it can also provide three-dimensional images, which offer researchers a new method to study the morphology of blood vessels more intuitively and make an accurate qualitative and quantitative analysis of them. All in all, OCTA will provide researchers with a possibility to study and early diagnosis for POAG [5]. In current study, SD-OCT was used to detect macular thickness, RNFL thickness and GCC thickness, and OCTA was used to detect superficial macular vessel density and perfusion density, superficial optic disc vessel density and perfusion density. Calculating the diagnostic efficacy of each parameter (presented as AUROC) was to provide a theoretical basis for the application of SD-OCT and OCTA in the diagnosis and evaluation of early POAG. Categorical variables were compared using the chi-square test. Area under the receiver operating characteristic curves (AUROC) were used to describe the utility of each parameter to discriminate glaucomatous eyes from the control eyes. An automatic resampling procedure was also used (n = 1000) as measurements of bilateral eyes nested within subject are more likely to be correlated.

Methods
Statistical analyses were performed using statistical software (Graphpad Prism7, GraphPad Software, CA, USA). A P-value of ≤0.05 was considered statistically significant.

Results
A total of 45 subjects (53 eyes), consisting of 15 early POAG subjects (23 eyes), 30 normal health subjects (30 eyes) were included in this study. Demographic and ophthalmic characteristics of the study subjects are summarized in Table 1. There was no statistically significant difference in terms of age and gender between the two groups (both P values >0.05). The IOP was significant higher in the early POAG subjects as compared with the control subjects (P<0.001). The two groups also differed by VFs (P<0.001). Table 2 summarizes the measurements of OCT parameters of the two groups. Significantly thicker macula was observed in healthy eyes compared with the early POAG eyes. Parameters of macular area thickness in the inner ring (temporal, nasal, superior, inferior) and the outer ring (nasal, superior, inferior) were found had statistically significant differences (all P values<0.05). No statistically significant difference, However, was found at the center of macular area thickness (P = 0.990) or at the temporal part of outer ring (P = 0.312) between the two groups. The GCC thickness in the normal control group was significantly higher than that in the early POAG group in all six sectors (temporal, superotemporal, inferotemporal, nasal, superonasal, inferonasal) (all P values<0. 05). Besides, the RNFL thickness in the normal health eyes was significantly higher than that in the early POAG eyes, and there were statistically significant differences both in the superior hemifield and inferior hemifield (both P values<0.05). Table 3 summarizes the measurements of OCTA parameters in the macular area for the two groups.
The calculated results showed that no statistically significant difference was found in central macula vessel density (P = 0.087) and in central macula perfusion density (P = 0.126) between the early POAG eyes and healthy eyes. Vascular density (in the inner ring, outer ring, and whole en face) and perfusion density (in the inner ring, outer ring, and whole en face) were all found significantly higher in the normal control subjects than that in the early POAG subjects (all P values<0. 05). (Figure 1) Table 4 summarizes the measurements of OCTA parameters in the optic disc area for the two groups.
Vascular density (in the center, inner ring, outer ring, and whole en face) and perfusion density (in the center, inner ring, and whole en face) were all found statistically significant differences between the normal control subjects and the early POAG subjects (all P values<0. 05). However, the final results revealed that there was no statistically significant difference in outer ring perfusion density (P = 0.185) between the early POAG eyes and healthy eyes.
The AUROCs of all the parameters to differentiate early POAG from healthy control eyes are shown in Table 5. Among these AUROCs, 6 parameters' AUROC ( thickness of the central macular area, thickness of the inferior sector of the inner ring, thickness of the temporal sector of the outer ring, the central macula vessel density, the central macula perfusion density, and the perfusion density of the outer ring of the optic disc area) showed no statistically significant difference (all P values>0.05).
Other parameters' AUROC were found had statistically significant differences to discriminate early POAG from healthy eyes (all P<0.05). The parameters with the highest diagnostic utilities were the vascular density in the macular area (except the macular center), and the AUROC reached 0.98. The diagnostic utilities of perfusion density in the macular area (except the macular center) were comparable to that in the macular area (except the macular center), and the AUROC were above 0.97.
The diagnostic abilities of vascular density in the optic disc area were put at second place, and the optimal parameter was the inner ring vascular density, and its AUROC reached 0.97. The diagnostic abilities of perfusion density in the optic disc area were slightly worse than vessel density's in the optic disc area. The optimal parameter was the central optic disc area perfusion density, and its AUC was 0.95. The diagnostic abilities of OCT-based parameters were generally worse than that of OCTAbased parameters, the optimal three parameters were the inferior hemifield RNFL thickness (AUC = 0.919), superior GCC thickness (AUC = 0.919), and the inferior GCC thickness (AUC = 0.908). ( Figure   2) Discussion POAG is a chronic optic neuropathy characterized by progressive apoptosis of the retinal ganglion cells and elevated intraocular pressure. The two main theories of optic nerve damage mechanism are IOP-induced mechanical injury theory and vascular ischemia theory [3]. Although the IOP-induced mechanical injury theory is generally accepted, in some patients with normal intraocular pressure, the observable excavated narrow-rim ONH appearance and VF defects are also presented. Meanwhile, some glaucoma patients persist in suffering optic disc and VF damages even after using drugs or surgery to control intraocular pressure, suggesting that there are other factors affecting the occurrence of POAG, and the focus on vessel-perfusion changes is one of its research hotspots. There are many technologies to document attenuation in ocular blood flow and dropout of retinal microvascular in glaucoma. However, they have played a limited role in elucidating the mechanism of microvasculature damage. Color doppler ultrasound Imaging (CDI) was once considered the most ideal non-invasive imaging technology, but CDI can only measure blood velocity not the actual blood flow volume. Mansouri et al found that the orbital blood flow in POAG patients was impaired, but the results of these kind of studies were not completely consistent, some even oppositely. The main reason is that CDI relies on the operator, hence, resulting in low repeatability of measurements [6]. In this study, we used the Angiography 6×6mm program, which has a wider scan range than Angiography 3×3mm, which offer more indicators for the diagnosis of early glaucoma. In addition, it has been reported that 6×6mm macula scans showed higher diagnostic accuracy compared with 3×3mm scans for differentiating between healthy and glaucoma eyes because the most vulnerable macula areas to glaucoma lie mostly outside the central 3×3mm [9,10]. Compared with Angiography 8×8 mm program, Angiography 6×6mm program provides higher resolution scan pictures. All in all, Angiography 6×6mm program ensured that we get as many parameters as possible, at the same time, obtain reliable scan quality. In previous studies, reports on the vessel density in the macular area or in the optic disc area often referred as whole image vessel density, and there was no differentiation of the center, the inner ring, and the outer ring. In this study, we found that the vessel density and perfusion density in the central part of the macular area of early glaucoma patients had no statistically significant difference compared to that of normal control eyes. Excluding the center of the macular area from the whole en face scan would be helpful to reduce the error of evaluating the vessel density and perfusion density. To the best of our knowledge, there are relatively few reports focusing on perfusion density to date which is obtained by depicting the width of the blood vessel and calculating the density of blood vessel coverage in the scan region, which can better reflect the perfusion volume of blood flow in the blood vessels. In this study, we also found that the perfusion density in both the macular area and the optic disc area was statistically different between the early POAG group and the normal control group and the diagnostic efficiency was high (except for the outer ring of optic disc area).
Some of the results of this study are consistent with previous findings [11][12][13]. The thickness of the macular area (except the center of the macular area, superior sector of the inner ring, temporal sector of the outer ring), GCC thickness, and RNFL thickness the vessel density and perfusion density in the optic disc area and the vascular density and perfusion density in the macular area in the early POAG were significantly reduced, compared with the normal control eyes. However, the difference is that this study showed that the diagnostic abilities in the macular area and in optic disc area (including vessel density and perfusion density) were higher than that of Rao, Hou and Zhong [14][15][16].
We consider the following points to be relevant. and other parameters of younger POAG patients will be more obvious, due to the higher baseline. This can partly explain why our results showed better diagnostic performance than that of previous reports. Hence, we speculate that OCTA has higher diagnostic value for young patients with early POAG.
There are also some limitations to the current study. Although 30 healthy eyes were included in the healthy control group in order to obtain a more reliable reference, the sample size was still small.
Moreover, patients who were over 60 years old were not included in this study. Inclusion of patients in the older age group would be more conducive to explaining the differences in diagnostic performance of macular and optic disc areas in different age groups. In addition, we didn't analyze deep retinal layer blood vessels, for the device has not been able to obtain deep retinal vascular data. After the software update in the future, more available parameters will help us to further understand the pathogenesis of glaucoma. Finally, since this was a cross-sectional study, we are not be able to evaluate the diagnostic value in terms of the progression of POAG. A longitudinal study will be helpful to document the changes of microvascular in progression of glaucoma.

Conclusion
Changes in macular thickness, GCC thickness, RNFL thickness, superficial vessel density and perfusion density in the optic disc or in macular area can be detected by OCT or OCTA in early POAG.
OCTA-based superficial macular vessel density and perfusion density had the highest diagnostic utility. The OCT-based diagnostic utility parameters were generally lower than the OCTA-based diagnostic utility parameters. OCTA has an important clinical application value in diagnosis and evaluation for early POAG.   Receiver operating characteristic curves of superior inner ring macula thickness, whole en face perfusion density of optic disc area, macular inner ring vessel density, superior GCC thickness and superior RNFL thickness