Ophthalmologists use OPT data to diagnose and characterize tissues and abnormalities in transverse and axial locations within the eye. For example, an ophthalmologist might request an OPT of the macula, the optic nerve or the cornea in either or both eyes for a given patient. Serial reports can be compared to monitor disease progression and response to treatment. OPT devices produce two categories of clinical data: B-scan images and tissue measurements.
Prior to interpreting an OPT B-scan (or set of B-scans), users must first determine if the study is of adequate quality to answer the diagnostic question. Examples of inadequate studies include:
The pathology that needs to be visualized does not appear within the field of the scan
The image quality is not sufficient to see the tissue layers of interest (i.e. media opacity, blink, etc)
The scans are not in the expected anatomic order (i.e. due to eye movements)
In some cases, inadequate images can be corrected by capturing another scan in the same area. However, in other cases, the patient’s eye disease interferes with visualization of the tissues of interest making adequate image quality impossible. Ideally, when choosing between multiple scans of the same tissue area, physicians would have access to information about the above questions so they can select only the best scan(s).
The physician may then choose to view and assess each B-scan in the dataset individually. When assessing OPT B-scans, ophthalmologists often identify normal or expected tissue boundaries first, then proceed to identify abnormal interfaces or structures next. The identification of pathology is both qualitative (i.e. does a structure exist) and quantitative (i.e. how thick is it). If previous scans are present for this patient, the physician may choose to compare the most recent scan data with prior visits. Due to workflow constraints, it may be difficult for B-scan interpretations to happen on the same machine that captures the images. Therefore, remote image assessment, such as image viewing in the examining room with the patient, is optimal.
In addition to viewing B-scan image data, clinicians also use quantitative measurements of tissue thicknesses or volumes extracted automatically from the OPT images. As with image quality, the accuracy of automated segmentation must be assessed prior to use of the numerical measurements based on these boundaries. This is typically accomplished by visual inspection of boundary lines placed on the OPT images but also can be inferred from analysis confidence measurements provided by the device software. In addition to segmentation accuracy, it is also important to determine if the region of interest has been aligned appropriately with the intended sampling area of the OPT.
The analysis software application segments OPT images using the raw data of the instrument to quantify tissue optical reflectivity and location in longitudinal scan or B-scan images. Many boundaries can be identified automatically with software algorithms, see Figure x.3-1.
Figure UU.3-1 OPT B-scan with Layers and Boundaries Identified
The innermost (anterior) layer of the retina, the internal limiting membrane (ILM) is often intensely hyperreflective and defines the innermost border of the nerve fiber layer. The nerve fiber layer (NFL) is bounded posteriorly by the ganglion cell layer and is not visible within the central foveal area. In high quality OPT scans, the sublamina of the inner plexiform layer may be identifiable. The external limiting membrane is the subtle interface between the outer nuclear layer and the photoreceptors. The junction between the photoreceptor inner segments and outer segments (IS/OS junction) is often intensely hyperreflective and in time domain OPT systems, was thought to represent the outermost boundary of the retina. Current thought, however, suggests that the photoreceptors extend up to the next bright interface, often referred to as the retinal pigment epithelium (RPE) interdigitation. This interface may be more than 35 micrometers beyond the IS/OS junction. When three high intensity lines are not present under the retina, however, this interdigitation area may not be visible. The next bright region typically represents the RPE cell bodies which consist of a single layer of cuboidal cells with reflective melanosomes oriented at the innermost portion of the cells. Below the RPE cells is a structure called Bruch's membrane which is contiguous with the outer RPE cell membrane.
The axial thickness and volume of tissue layers can be measured using the boundaries defined above. For example, the nerve fiber layer is typically measured from the innermost ILM interface to the interface of the NFL with the retina. Time domain OPT systems measure retinal thickness as the axial distance between the innermost ILM interface and the IS/OS junction. However, high resolution OPT systems now offer the potential to measure true retinal thickness (ILM to outermost photoreceptor interface) in addition to variants that include tissue and fluid that may intervene between the retina and the RPE. The RPE layer is measured from the innermost portion of the RPE cells, which is the hyper reflective melanin-containing layer to the outermost highly reflective interface. Pathologic structures that may intervene between normal tissue layers may obscure their appearance but often can be measured using the same methods as normal anatomic layers.
The macular grid is based upon the grid employed by the Early Treatment of Diabetic Retinopathy Study (ETDRS) to measure area and proximity of macular edema to the anatomic center of the macula, also called the fovea. This grid was developed as an overlay for use with 32mm film color transparencies and fluorescein angiograms in the seminal trials of laser photocoagulation for the treatment of diabetic retinopathy. Subsequently, this grid has been in common use at reading centers since the 1970s, has been incorporated into ophthalmic camera digital software, and has been employed in grading other macular disease in addition to diabetic retinopathy. This grid was slightly modified for use in Time Domain OPT models developed in the 1990s and early 2000s in that the dimensions of the grid were sized to accommodate a 6 mm diameter sampling area of the macula.
The grid for macular OPT is bounded by circular area with a diameter of 6 mm. The center point of the grid is the center of the circle. The grid is divided into 9 standard subfields. The center subfield is a circle with a diameter of 1 mm. The grid is divided into 4 inner and 4 outer subfields by a circle concentric to the center with a diameter of 3 mm. The inner and outer subfields are each divided by 4 radial lines extending from the center circle to the outermost circle, at 45, 135, 225, and 315 degrees, transecting the 3 mm circle in four places. Each of the 4 inner and 4 outer subfields is labeled by its orientation with regard to position relative to the center of the macula – superior, nasal, inferior, and temporal. For instance, the superior inner subfield is the region bounded by the center circle and the 3 mm circle the 315 degree radial line, and the 45 degree radial line. The nasal subfields are those oriented toward the midline of the patient's face, nearest to the optic nerve head. The grids for the left and right eyes are reversed with respect to the positions of the nasal and temporal subfields – in viewing the grid for the left eye along the antero-posterior (Z) axis, the nasal subfields are on the left side and in the right eye the nasal subfields are on the right side (nasal as determined by the location of the subfield closest to the nose).
The OPT macula thickness report consists of the thickness at the center point of the grid, and the mean retinal thickness calculated for each of the 9 subfields of the grid. In the context of the macular disease considered for the diagnosis, and qualitative interpretation of morphology from examination and OPT and/or other modalities, the clinician uses the macula thickness report to determine if the center and the grid subfield averages fall outside the normative range. Monitoring of macular disease by serial grid measurements allows assessment of disease progression and response to intervention. Serial measurements are assessed by comparing OPT thickness or volume reports, provided that the grids are appropriately centered upon the same location in the macula for each visit.
Figure UU.5-1 Macular Grid Thickness Report Display Example
The center point of the grid should be aligned with the anatomic center of the macula, the fovea. This can be approximated by having the patient fixate upon a target coincident with the center of the grid. However, erroneous retinal thickness measurements are obtained when the center of the grid is not aligned with the center of the macula. This may occur in patients with low vision that cannot fixate upon the target, or in patients that blink or move fixation during the study. To determine the expected accuracy of intervisit comparisons, clinicians would benefit from knowing the alignment accuracy of the OPT data from the two visits. Ophthalmologists may also want to customize locations on the fundus to be monitored at each visit.
The following figure illustrates how the content items of the Macular Grid Thickness and Volume Report are related to the ETDRS Grid. Figure shown is not drawn to scale.
Figure UU.5.2 – ETDRS GRID Layout
The process of evaluation of diabetic macular edema will help illustrate the role of the OPT macula thickness report. In diabetic macular edema there is a breakdown in the blood retina barrier which can lead to focal and/or diffuse edema (or thickening) of the macula. The report of the thickness of each subfield area of the macula grid will help direct treatment. For instance, laser treatment to a specific thickened quadrant would be expected to reduce the thickness of retina in the treated zone. Serial comparisons of OPT thicknesses should demonstrate a reduction in thickness in the successfully treated zone. A zone that subsequently became thicker on follow-up scans may warrant further treatment. In addition to an expected local response to specific zonal treatment such as laser, there are treatments with drugs and biologics which are less localized. For instance, the injection of intravitreal drugs in a successfully treated eye would be expected to have a global reduction of thickness in all zones with DME. Patients with severe retinal disease may lose the ability to fixate making the acquisition of OPT images to represent a specific zone less reliable.