Doppler Optical Coherence Tomography of Retinal Circulation

The overall goal of this procedure is to measure total retinal blood flow using Doppler Optical Coherence Tomography or OCT and a semi-automated grading software. This is accomplished by first scanning the patient with Doppler OCT according to a dual angle protocol. The second step is to automatically evaluate the quality of OCT images and detect blood vessels in the images.

Next, the graders judge and revise the vessel location, diameter, and type according to OCT images and a color photograph of the optic disc. The final step is to calculate the Doppler signal and doppler angle of the detected retinal veins and estimate the blood flow and velocity in each vein. Ultimately, total retinal blood flow is obtained by summing the blood flow in all veins around the optic disc.

The main advantage of this technique over existing methods, such as laser doppler techniques and ultrasound color doppler imaging, is that it provides total retinal blood flow measurements using scans taken in just two seconds. Video demonstration of this master is critical as the steps are difficult to learn, and because the accuracy of result depends on the image quality and the subjective greeting of vessel position. Demonstrating the scanning procedure will be ophthalmic technician Janis Van Norman.

For this protocol, patients should be scanned using an RT view for your domain optical coherence tomography or OCT system, utilizing a circum papillary double circular scan pattern. A 3D optic disc OCT scan is also needed. The double circular scan pattern consists of two concentric circles around the optic nerve head with an inner ring diameter of 3.4 millimeters and an ounce of ring diameter of 3.75 millimeters.

This pattern transects all branch retinal arteries and veins emanating from the optic nerve head. The pattern is performed six times consecutively over two seconds, so that blood flow measurement can be averaged over approximately two cardiac cycles. The Doppler scans are performed according to a dual angle protocol where three scans are obtained with the OCT beam passing through the SUP nasal portion of the pupil and three scans are obtained through the inferior nasal portion.

This provides flow measurements with two incidents angles on each vessel on the OCT machine. After each scan is obtained, the image quality indices are automatically calculated, keep only scans of acceptable quality. The real-time quality check improves the likelihood of valid flow measurements.

Next, perform the 3D optic disc scan. This is a raster scan covering a six by six millimeter region around the optic disc. This provides a detailed on fas image of this area.

Also import a color photograph of the optic disc to help distinguish arteries and veins. To begin analysis, first, transfer the Doppler OCT images, 3D disc, OCT images and color disc photography to the blood flow grading center where grad use doctor C software to measure retinal blood flow velocity and vessel area in the reading center. The greater first classifies the data of each eye as either good or poor in quality.

According to the average eye movement and signal strength of the repeated scans, only the scans with good quality are graded for acceptable scans. Apply an automated segmentation algorithm to register the frames on the same circle and to segment each OCT image for vessel detection. Then apply an automated algorithm that matches vessels to locate the same vessel in each frame.

For an overview project the detected vessel as a line segment onto the on FFAs fundus image calculated from the 3D disc scan. Have the graders review each vessel on a small portion of the averaged inner and outer ring frames, overlaid with a circle that represents the automated segmentation result. The grader should judge whether the location, vessel, diameter, and vessel type correctly match the vessel on the two rings.

According to the OCT image on fas OCT image and disc photograph of the same eye, allow the grader to change any of the above values if deemed necessary. A confidence score of zero to five is automatically given to each vessel based on the Doppler signal strength in the vessel area. The grader then manually corrects this based on the strength of the Doppler signal regularity of vessel boundary agreement between inner and outer rings of vessel size and sign agreement of the Doppler shift between the inner and outer rings.

The confidence score will be used for weighted averaging flow among scans. After all the vessels are verified and corrected, apply an automated algorithm to calculate the blood flow S area and velocity of each vein. Grading qualities and total retinal blood flow venous and artery area velocity are calculated.

The algorithms detail are illustrated in this flow chart. First, integrate the doppler signal over the vessel area and then average it among all frames. Then calculate the doppler flow as the summed doppler signal divided by the Doppler angle.

The vessel area is also averaged among frames for each vessel. The blood flow is calculated based on a weighted average from the valid scans. The validity is decided by the previously calculated subjective confidence score and an objective evaluation of the reliability of the Doppler angle.

The weight of the valid scans is decided by the strength of the summed doppler signal and variation of the doppler angle. Calculate the average flow speed by summing the flows in the valid veins and dividing by the summed area of those veins. The vessels with no valid scans.

Estimate the flow using the vessel area and the average flow speed from valid veins. Add the calculated flow for all veins to determine the total retinal blood flow. Evaluate the reliability of total retinal blood flow measurement based on the valid venous area percentage eye movement and signal strength.

Next, calculate the superior and inferior hemisphere blood flow. Also obtain venous area and total arterial area by adding the vessel areas of all veins and all arteries. Assuming the total retinal blood flow is the same in arteries and veins, calculate the arterial and venous velocities by dividing total blood flow with arterial area and venous area.

The results were inputted into the Doctor C central database. Researchers can review the result online. The results include two parts, the grading qualities and the measurements.

The results can also be downloaded as a spreadsheet. Here we see the total retinal blood flow as determined by two graders using Doctor Z software. The total retinal blood flow measurements from each grader are similar and similar to flow rates determined by the other grader using the manual software and literature.

The yield rate using the single angle protocol is about 65%but can be improved to 80%using the dual angle protocol and realtime image quality control for three graders. A good correlation exists between total blood flow and the pattern standard deviation from visual field tests the glaucoma to eyes for all graders. The integrator reproducibility as measured by the coefficient of variation, is similar for both glaucomatous and normal eyes.

Likewise, the reproducibility measurements for the two methods, Dr.C and the manual software are similar. Once mastered, the dual angle protocol can be performed in several minutes and the greeting can be done in approximately 20 minutes. While performing doper OCT scan, it is important to remember to follow the dual angle protocol and check image quality for each scan.

This technique has wide applications for optic nerve and retinal diseases such as glaucoma and diabetic retinopathy.