Distinct profiles for the two cone diseases (BCM and ACH) were calculated with the method of longitudinal reflectivity profile analysis of OCT images. Although these diseases are clinically difficult to distinguish, our analysis algorithm allows reliable and simple differentiation by using the LRPs that yield a one-to-one configuration.
Misinterpretation of profiles is highly unlikely because of the narrow 99% confidence interval. Because of the marked differences and the high reproducibility of LRPs, it is possible to differentiate the two diseases based on the LRP only.
Achromatopsia is a rare congenital retinal disease characterized by an almost absent cone function
20 and is usually inherited in an autosomal recessive manner. Responsible mutations such as CNGA3, CNGB3, and GNAT2
21 22 23 24 25 have been found in several genes. In some patients, an exact chromosomal localization of the mutated gene has not been established yet (e.g., ACHM1).
26 A dysfunction of cones due to these mutations has been proposed as the reason for visual impairment.
21 Patients with ACH usually present with severely impaired visual acuity, a relative central scotoma, congenital nystagmus, and absence of color discrimination. Biomicroscopy shows normal anterior and posterior segment morphology
(Fig. 1b)and sometimes a reduced foveal reflex or fine pigment mottling. Visual function in ACH is stationary, and all clinical findings are already present at birth (congenital functional defect of cone system). Falls et al.
12 estimated the number of foveal cones to be normal, whereas extrafoveal cones seemed to be reduced in number. Retinal and foveal thickness were reported to be normal.
12 19 In contrast, the quantity of rods seemed to be normal throughout the entire retina. Foveal cones showed an abnormal configuration of the inner segment as well as dislocated nuclei. Ganzfeld electroretinogram (ERG) findings are described as characteristic for ACH.
27
On OCT images of patients with ACH, P2 (photoreceptor reflectivity) disappeared completely, whereas histologic examinations have shown that photoreceptors are present.
12 Peak 2 presumably represents a highly reflective structure between the inner and outer photoreceptor segments. This structure may represent the ovoid region of the photoreceptors. The ovoid lies in the outermost part of the inner segment of the photoreceptor and contains densely packed mitochondria. Published reports on histology of eyes with ACH date back to 1921 and 1965, a time before the availability of electron microscopy that can show the presence or absence of mitochondria in the diseased cones. It is tempting to speculate that the missing P2 in OCT scans of patients with ACH reflects alterations in this mitochondria-rich region of the photoreceptors.
BCM is usually a stationary, X-linked, recessively inherited disease that occurs even less frequently than ACH. However, recent reports indicate that there is evidence of progressive loss of cone function in older individuals.
28 29 Mutations in the red and green opsin genes have been identified (OPN1LW,
30 OPN1MW
31 ). Patients with BCM usually have better visual acuity than do patients affected by ACH, and nystagmus often regresses eventually. Despite the reduced cone function, affected persons have a residual ability to distinguish colors—especially shades of blue—used in specific color vision tests.
14 Much as in ACH, biomicroscopy of patients with BCM reveals a normal retinal aspect
(Fig. 1C) . The ganzfeld ERG shows similar findings as it does for ACH.
32 33 Until now, no histopathological data on BCM have been reported. Clinical distinction between ACH and BCM is difficult, especially in children
(Fig. 1) . Although BCM is clinically well defined, BCVA, Ganzfeld ERG, and color testing are needed to establish a diagnosis. The method of LRP analysis of OCT images complements the diagnostic procedure, revealing high specificity and reproducibility. The finding that P2 (photoreceptors) is shifted toward P1 (RPE) and reflectivity is significantly reduced indicates a reduction in size of the photoreceptors and a structural modification either of the cell itself or of intracellular organelles (mitochondria). The lack of P3 (external limiting membrane) supports the thesis that a structural change in the photoreceptors and/or surrounding tissue is present. P2 cannot be misinterpreted as P3 because the external limiting membrane never reaches the reflectivity of P2 and the advancement of P2 toward P1 can clearly be seen in all patients
(Fig. 1F) . Unfortunately, these findings cannot be correlated to morphologic observations, because histologic data of patients with BCM are lacking. Foveal thickness was reduced in patients with BCM compared with those with ACH and controls subjects (
P < 0.001).
With reference to the progressive forms of BCM with increasing loss of cone function,
28 29 a potential misinterpretation with other forms of progressive maculopathy such as Stargardt’s disease could be hypothesized. Besides the phenotypical differences, LRPs in Stargardt’s diseases cannot be misinterpreted for BCM. LRPs of patients with Stargardt’s disease show only P1 and P4. Peaks 2 and 3 are missing, and foveal thickness is dramatically reduced (∼20 μm; Barthelmes D, Fleischhauer JC, unpublished observation, 2005).
We speculate that mutations resulting in the phenotype of BCM lead to a reduction in the size of the photoreceptors (reduced foveal thickness, distance from P1 to P4) and to an alteration of the area between inner and outer photoreceptor segments (shifting of P2 toward P1, absence of P3). These results indicate an alteration of the outer segment of the photoreceptors (shifting of P2 toward P1). The distance from P2 to P4 was less in patients with BCM (121 ± 10 μm) than that in normal subjects (148 ± 8 μm). This difference would indicate an alteration in the area of the somata of the photoreceptors. Again, unfortunately no histologic reports are available on BCM to support this hypothesis.
Reports in the literature on foveal thickness measured with OCT devices range from 146 to 153 μm.
3 34 35 These values were acquired by using the built-in algorithm of the OCT software. Because we propose P2 to represent the photoreceptors and P1 the RPE, the standard OCT software calculates foveal thickness from P2 to P4. Already published values are in accordance with ours if the additional ∼50 μm from P1 to P2 is added.
3 34 35
The method of using LRPs to analyze OCT images has primarily been described by Huang et al.
7 In a recent publication, Ishikawa et al.
36 used LRPs to reevaluate retinal thickness in normal and glaucomatous eyes in the macular area.
36 In both studies, the original data sets recorded by the instrument were used to evaluate the reflectivity. In the study by Huang et al.,
36 OCT-1 was used (Carl Zeiss Meditec, AG), which has a much lower resolution (∼25 μm) than the newer versions and indeed, LRP revealed much more information on the layerlike structure in the tissue. Ishikawa et al.
36 used the original scan data exported from the Stratus OCT-3 to compare changes in the different retinal layers between control subjects and patients who had glaucoma. For glaucomatous eyes, significant changes in the retinal ganglion cell layer were detected. We exported 8-bit grayscale images for analysis in our approach. Patients with cone disorders often exhibit nystagmus that can be more or less pronounced. Nystagmus makes image acquisition very difficult, especially if there is a need for the fovea to be well discriminated on the images. Off-line analysis of distorted images due to eye movements is much easier if performed on the 8-bit grayscale image, where the region of interest (ROI) can be chosen arbitrarily. Thus, the original A-scans do not have to be selected within the large raw data file itself.
In conclusion, we demonstrate a method to analyze OCT images quantitatively. Two cone diseases were investigated that showed clinically high similarity. They exhibited characteristic features of tissue reflectivity and thus were easily distinguished by longitudinal reflectivity profiles of OCT image analysis.