Feature Of The Week 4/29/12: Medical University of Vienna Demonstrates Automated Measurement of Choroidal Thickness in the Human Eye by PS-OCT
Researchers at the Medical University of Vienna have a long history of outstanding contributions to the field of Optical Coherence Tomography dating back 25 years. This week work by Dr. Hitzenberger, one of the pioneers of the field, and his colleagues is featured demonstrating automated measurement of choroidal thickness in the human eye by polarization sensitive optical coherence tomography. Below is a summary of their work.
Imaging of retinal layers and quantitative measurements of their thickness are among the most important applications of OCT. E.g., the total retinal thickness is an important quantity for diagnosis of diabetic retinopathy, the thickness of the retinal nerve fiber layer is reduced in glaucoma, providing an important measure for early diagnosis of this disease, and the thickness of the choroid is of interest for disorders like high myopia, diabetic retinopathy, and central serous chorioretinopathy.
While commercially available ophthalmic OCT systems are quite successful in measuring layers down to the retinal pigment epithelium (RPE), they have problems accessing deeper layers like the choroid. The main reason is that imaging of deeper layers is hindered by scattering and absorption of the sampling light in the RPE. Another shortcoming of commercial OCT systems is that they acquire images just based on backscattered intensity which does not provide a tissue specific contrast.
The first problem can be overcome by using light of longer wavelength, around 1050 nm where scattering and absorption of the RPE are reduced and water absorption has a local minimum. The use of this wavelength enabled imaging of the choroid down to the choroid-sclera interface (CSI). Recently, several approaches to measure the thickness of the choroid by OCT operating in the 1050 nm range have been published. Most of them use manual segmentation based on intensity images. This approach is rather subjective and time consuming.
We developed an alternate approach: we use polarization sensitive (PS)-OCT at 1050 nm, a functional extension of OCT that takes advantage of the additional information carried in the light's polarization state. Our method is based entirely on intrinsic, tissue specific polarization contrast mechanisms. In a first step, the anterior boundary of the choroid, the RPE, is segmented based on depolarization. In a second step, the choroid-sclera interface (CSI) is found by using the birefringence of the sclera. The choroidal thickness is then calculated as the distance between RPE and CSI.
We use a state-of-the-art high-speed swept source PS-OCT instrument developed in our lab. The system operates at an A-scan rate of 100 kHz and achieves a nominal maximum sensitivity if 102 dB. 50 B-scans (consisting of 1000 A-scans each) are recorded at the same position. From these, several averaged high-quality B-scan images are derived: reflectivity, retardation, and degree of polarization uniformity (DOPU). From these images, RPE and CSI are segmented, and the choroidal thickness is derived.
Repeated measurements were performed in the eyes of 5 healthy volunteers, and reproducibility was determined. The mean choroidal thickness varied between subjects, ranging from ~ 270 – 380 µm, the average reproducibility was 18.3 µm, in the range of the diurnal variation of the choroidal thickness.
Finally, a comparison to manual segmentation on intensity based images was performed. The overall agreement was good, however, a systematic deviation of ~ 35 µm was observed (PS-OCT results showed slightly larger thickness). Further studies are required to analyze the reasons for this difference.
Despite this unresolved discrepancy, we think choroidal thickness measurement by PS-OCT is an interesting alternative to intensity based methods because it can be fully automated and provides additional, complementary information to intensity based data.
For more information see recent Article. Courtesy Christoph Hitzenberger from the Medical University of Vienna.