Woo Jyh Haur MD
The type of imaging device used to measure corneal thickness in keratoconus patients after accelerated corneal crosslinking (CXL) matters, according to Woo Jyh Haur MBBS, MMed(Ophth), FRCOphth, FRCSE. Readings are not interchangeable between devices using Scheimpflug cameras and spectral-domain optical coherence tomography (OCT) – and the CXL procedure itself may affect the accuracy of some measurements, he said.
With colleagues at the Singapore National Eye Centre, Dr Woo obtained central and thinnest corneal measurements using the Pentacam® (OCULUS) Scheimpflug camera and Optovue spectral-domain OCT devices from 47 eyes in 47 patients before and after undergoing an accelerated CXL procedure.
The procedure involved removal of the corneal epithelium followed by application of a 0.1% riboflavin solution, and UVA illumination at 30mW/cm2 for four minutes for a total energy of 7.2 J/cm2. Postoperative topical antibiotics, 0.5% moxifloxacin, and corticosteroids, 0.12% prednisolone, were tapered at one month after complete epithelial healing.
Patients were at least 18 years old and displayed axial topography consistent with keratoconus as measured with the Pentacam. All patients’ steepest K values were 47.0 dioptres or more, and were progressing as determined by topography, visual acuity or refractive changes.
Patients with thinnest pachymetry less than 400 microns, those with gaze disorders or other ocular conditions, or pregnancy, were excluded. The study group included 37 males and 10 females with a mean age of 28 ±7 years, ranging from 19 to 52. Baseline mean K was 49.78 ±5.00 dioptres, maximum K 59.45 ±8.49 dioptres, and mean spherical equivalent -4.30 ±3.00 dioptres.
In general, measurements from the two devices were highly correlated, Dr Woo reported. Mean Pentacam estimates of central corneal thickness ran about seven microns to 22 microns thicker than Optovue readings at baseline, and three, six and 12 months after the procedure.
However, at one month post-CXL, the Pentacam estimates dipped sharply and were nearly identical with the Optovue measurements. Mean Pentacam central corneal measurements dropped from 492.95 microns at baseline to 474.03 at one month, recovering to 490.61 at 12 months. Overall, mean Optovue measurements varied much less, ranging from 471.16 microns at baseline to 475.38 microns at one month and 476.66 microns at 12 months’ follow-up.
The consistent differences in corneal thickness readings between the Scheimpflug and OCT devices likely come down to differences in how the technologies capture images and their processing algorithms, Dr Woo said. The Scheimpflug device uses a rotating camera which captures 50 slit images in two seconds, directly measuring elevation at more than 5,000 data points. The OCT device calculates elevation from 26,000 radial A-scans per second.
Using OCT, the anterior corneal boundary may be delineated slightly below the anterior corneal surface, which may account for its usually thinner readings, Dr Woo suggested. Faster image acquisition and better edge detection with OCT may also influence OCT readings, possibly accounting for their greater consistency.
These technical differences may also explain why the Scheimpflug device appeared to underestimate corneal measurements one month after surgery, Dr Woo said. Postoperative stromal haze along with changes in corneal microstructure and reflectivity may have interfered with measurements. An inverse relationship between central corneal thickness measured by dual Scheimpflug tomography and density of stromal haze has previously been reported, he noted (Antonios R et al. Am J Ophthalmol. 2016 Jul; 167:38-47).
Limitations of Dr Woo’s study included the small sample size, and the fact that reproducibility of measurements and possible intra-observer bias were not controlled for, he noted.
Systematic differences exist between these modalities and measurements are not interchangeable
CONSIDER THE SOURCE
Noting that accurate corneal thickness measurements are essential in the assessment and follow-up of patients with keratoconus, Dr Woo cautioned against directly comparing readings from Scheimpflug and OCT devices. “Systematic differences exist between these modalities and measurements are not interchangeable,” he said.
Post-crosslinking stromal changes may further increase discrepancies in corneal thickness measurements, with the Pentacam possibly underestimating measurements one month after surgery, Dr Woo said. With this in mind, clinicians should interpret corneal thickness measurements in the context of the imaging modality used.
Woo Jyh Haur: email@example.com