23 Optical Coherence Tomography for Imaging Anterior-Chamber Inflammatory Reaction in Uveitis
Anterior chamber (AC) cellular reaction is often clinically observed by slit-lamp biomicroscopy. 1 However, it cannot be performed in eyes with postoperative corneal edema or haze, and the grading ability is also dependent on the experience of the examiner. The existing objective method for AC reaction grading is the flare meter, which is based on the principle of laser photometry. Counting cells by laser flare meter is effective and linear in controlled laboratory situations, but it appears to be less accurate than flare measurements in vivo. 2 Moreover, its reliability in edematous cornea has not been proved so far. Laser flare photometry has been known to be affected by the aqueous humor protein concentration, the mydriatic agent used, and the presence of red blood cells. 2
Optical coherence tomography (OCT) has evolved as a widely used imaging modality in ophthalmology in the last decade. Earlier, we reported a case series on the objective method of diagnosing the AC cells in anterior uveitis with time-domain (TD) anterior-segment OCT. 3 Recently, OCT machines with sophisticated optics with spectral (SD) or Fourier domain (FD), which enabled examination of both the anterior and posterior segment in a single setup, have been developed. 4 , 5 , 6 , 7 , 8 , 9 This development has enabled high-definition scanning of the AC. In this chapter, we discuss the use of OCT, both TD and the FD, in grading AC reaction and the methods of evaluation in uveitis.
23.1 Time-Domain OCT in Anterior-Chamber Cell Detection
Various causes of uveitis, including acute and chronic idiopathic anterior uveitis, postoperative uveitis, panuveitis, herpetic keratouveitis, interstitial keratitis, corneal ulcer, posterior corneal abscess, and endophthalmitis were included in the study. Sixty-two eyes of 45 patients referred to the uvea clinic were studied. Patients not willing to undergo follow-up were excluded. A detailed slit-lamp examination was performed in all patients, and the AC reaction was graded clinically from 0 to 4 using the Standardization of the Uveitis Nomenclature (SUN). 10 Cross-sectional imaging of AC with anterior-segment OCT (Carl Zeiss Meditec, Dublin, CA) was used. Corneal high-resolution single-scan mode (Fig. 23.1) was used. Four images were taken in one axis. The single capture image of the AC included the central cornea to the anterior surface of the lens.
23.1.1 Automated Cell Analysis
The hyperreflective spots detected in the AC were counted in nonenhanced image manually and using an automated computer algorithm. The mean numbers of cells were determined by counting the hyperreflective spots from the OCT image (Fig. 23.1). In the automated method, hyperreflective spots were segmented (Fig. 23.2) from the OCT image by a pixel-based candidate object extraction and counted with a connected component-labeling technique using custom MATLAB software version 7.1 (MathWorks, Natick, MA). 3
23.2 Comparison between Manual and Automated Cell-Counting Methods
23.2.1 Anterior Chamber Cells
In the manual counting method, mean hyperreflective spots were 3 ± 1.8 in grade 1, 12 ± 3.5 in grade 2, 33.8 ± 10.2 in grade 3, and 61.4 ± 9.6 in grade 4. The automated method showed mean 3 ± 1.9 hyperreflective spots in grade 1, 12.4 ± 3.6 in grade 2, 33.2 ± 9.6 in grade 3, and 74.8 ± 17 in grade 4. We observed good correlation (Pearson coefficient for grade 1: 0.995, grade 2: 0.948, grade 3: 0.985, and grade 4: 0.893) between automated and manual methods. Except for grade 4 (P = 0.009), there were no significant differences in mean values between the manual and automated method in lower grades. The automated method was more sensitive in grade 4 uveitis and detected a greater number of cells (Fig. 23.2).
23.2.2 Aqueous Flare
Aqueous flare up to grade 3 was not detected in anterior-segment OCT. However, grade 4 aqueous flare, which was characterized by intense fibrinous reaction, was seen by OCT in seven (11.2%) eyes.
23.2.3 Keratic Precipitates
Keratic precipitates were seen as discrete hyperreflective spots attached to the endothelium of the cornea (Fig. 23.3). They were counted similar to AC cells. Keratic precipitates were noted in 12 eyes in the study.
Fibrinous membrane was seen in three eyes in the pupillary area and in one eye on the endothelium of the cornea (Fig. 23.4).
23.2.5 Corneal Edema and OCT Cell Counting
Of the 12 eyes with corneal edema, anterior-segment OCT was also able to detect AC cells (Fig. 23.5) in 11 (91.6%) eyes in which slit-lamp grading was not possible because of corneal edema. The central corneal thickness ranged from 702 to 1020 µm (mean, 843± 109 µm). The eyes with postoperative uveitis, endophthalmitis, and corneal infection had corneal edema. The mean number of hyperreflective spots detected in eyes with corneal edema using the manual method was 12.27 ± 12.1 and 12.9 ± 13 using the automated method.