Automated vessel diameter quantification and vessel tracing for OCT angiography

Innovative Macula Capillaries Plexuses Visualization with OCTA B-Scan Graph Representation: Transforming OCTA B-Scan into OCTA Graph Representation

Purpose

The aim of this study is to transform optical coherence tomography angiography (OCTA) scans into innovative OCTA graphs, serving as novel biomarkers representing the macular vasculature.

Patients and Methods

The study included 90 healthy subjects and 39 subjects with various abnormalities (29 with diabetic retinopathy, 5 with age-related macular degeneration, and 5 with choroid neovascularization). OCTA 5µm macular coronal views (MCVs) were generated for each subject, followed by blood vessel segmentation and skeleton processing. Subsequently, the blood vessel density index, blood vessel skeleton index, and blood vessel tortuosity index were computed. The graphs of each metric were plotted against the axial axes of the OCTA B-scan, representing the integrity of vasculature at successive 5µm macular depths.

Results

The results revealed two significant findings. First, the B-scans from OCTA can be transformed into OCTA graphs, yielding three specific OCTA graphs in this study. These graphs provide new biomarkers for assessing the integrity of deep vascular complex (DVC) and superficial vascular complex (SVC) within the macula. Second, a statistically significant difference was observed between normal (n=90) and abnormal (n=39) subjects, with a t-test p-value significantly lower than 0.001. The Mann-Whitney u-test also yielded significant difference but only between the 90 normal and 29 DR subjects.

Conclusion

The novel OCTA graphs offer a unique representation of the macula's SVC and DVC, suggesting their potential in aiding physicians in the diagnosis of eye health within OCTA clinics. Further research is warranted to finalize the shape of these newly derived OCTA graphs and establish their clinical relevance and utility.

Early Retinal Microvasculopathy in Prediabetic Patients and Correlated Factors

INTRODUCTION

We aimed to detect early retinal microcirculation changes in prediabetic patients and investigate their correlation with clinical examinations.

METHODS

Forty-seven prediabetic individuals, 29 controls, and 81 type 2 diabetic mellitus (T2DM) patients were enrolled in this study. A review of clinical data and spectral-domain optical coherence tomography angiography (SD-OCTA) parameters of macular vessel diameter (VD), foveal avascular zone (FAZ), and macular vessel area density (VAD) was performed.

RESULTS

Levels of low-density lipoprotein cholesterol and triglycerides in prediabetes and T2DM groups were significantly higher than those in the control group. The urine microalbumin-to-creatinine ratio (ACR) was mildly and moderately increased in the prediabetes and T2DM groups, respectively. The estimated glomerular filtration rate of the three groups was within the normal range. SD-OCTA showed that VAD in the superficial macular area was decreased in the prediabetes group compared to the control group (p = 0.01). The FAZ size, particularly in the deep layer, was expanded in the prediabetes group. In the deep retinal layer of the macular area, VD and FAZ size in the prediabetes group were larger than those in the control group. In the prediabetes group, the axial length was significantly correlated with macular VD and FAZ size (p < 0.05), and ACR was correlated with FAZ size (p < 0.05). Age had a negative correlation with VAD (p < 0.01). ACR had a positive correlation with FAZ size (p < 0.05).

CONCLUSIONS

Enlargement and irregularity of the FAZ area, deep capillary dilation, and a decrease in VAD occur in the retina of prediabetic patients with mild kidney function impairment.

Multi-step deep neural network for identifying subfascial vessels in a dorsal skinfold window chamber model

Automatic segmentation of blood vessels in the dorsal skinfold window chamber (DWSC) model is a prerequisite for the evaluation of vascular-targeted photodynamic therapy (V-PDT) biological response. Recently, deep learning methods have been widely applied in blood vessel segmentation, but they have difficulty precisely identifying the subfascial vessels. This study proposed a multi-step deep neural network, named the global attention-Xnet (GA-Xnet) model, to precisely segment subfascial vessels in the DSWC model. We first used Hough transform combined with a U-Net model to extract circular regions of interest for image processing. GA step was then employed to obtain global feature learning followed by coarse segmentation for the entire blood vessel image. Secondly, the coarse segmentation of blood vessel images from the GA step and the same number of retinal images from the DRIVE datasets were combined as the mixing sample, inputted into the Xnet step to learn the multiscale feature predicting fine segmentation maps of blood vessels. The data show that the accuracy, sensitivity, and specificity for the segmentation of multiscale blood vessels in the DSWC model are 96.00%, 86.27%, 96.47%, respectively. As a result, the subfascial vessels could be accurately identified, and the connectedness of the vessel skeleton is well preserved. These findings suggest that the proposed multi-step deep neural network helps evaluate the short-term vascular responses in V-PDT.

Model-based optical coherence tomography angiography enables motion-insensitive vascular imaging

We present a significant step toward ultrahigh-resolution, motion-insensitive characterization of vascular dynamics. Optical coherence tomography angiography (OCTA) is an invaluable diagnostic technology for non-invasive, label-free vascular imaging in vivo. However, since it relies on detecting moving cells from consecutive scans, high-resolution OCTA is susceptible to tissue motion, which imposes challenges in resolving and quantifying small vessels. We developed a novel OCTA technique named ultrahigh-resolution factor angiography (URFA) by modeling repeated scans as generative latent variables, with a common variance representing shared features and a unique variance representing motion. By iteratively maximizing the combined log-likelihood probability of these variances, the unique variance is largely separated. Meanwhile, features in the common variance are decoupled, in which vessels with dynamic flow are extracted from tissue structure by integrating high-order factors. Combined with Gabor-domain optical coherence microscopy, URFA successfully extracted high-resolution cutaneous vasculature despite severe involuntary tissue motion and scanner oscillation, significantly improving the visualization and characterization of micro-capillaries in vivo. Compared with the conventional approach, URFA reduces motion artifacts by nearly 50% on average, evaluated on local differences.