Abstract
Purpose: :
In clinical and scientific practice central retinal thickness of central millimeter (CMT) and centerpoint (CPT) is frequently used as a quantative parameter obtained from spectral-domain optical coherence tomography (SD-OCT) examinations. However this value might be dependent on correct plotting of the centerpoint (CP). An incorrect output of CMT (and CPT) values could lead to wrong clinical decisions (for example in PRN retreatment criteria). The reliability of this parameter is therefore of major importance. In this study the influence of CP plotting on CMT and CPT was investigated.
Methods: :
29 eyes of 19 patients (19 eyes with macular edema due to diabetic retinopathy/central or branch vein occlusion/age-related macular disease, 6 with degenerative disease including two patients with macular holes and 4 healthy eyes) were scanned on four commonly used SD-OCTs (Heidelberg Spectralis® (SPEC), Zeiss CirrusTM HD-OCT (CIRR), Topcon 3D-OCT 2000 (TOP2), Nidek RS-3000 (NID3)) on the same day in random order. Data were exported from the devices and graded in grey-scale using Vienna Reading Center custom software. In a blinded fashion two experienced SD-OCT grading supervisors manually marked the true CP in all image stacks. All segmentation errors that could potentially confound thickness measurements were corrected manually. The distance between the manually set CP and the device’s CP as well as the resulting differences in CMT and CPT were automatically calculated by the software and compared between the devices.
Results: :
The agreement of CP plotting between the two graders was excellent. The mean distance between the two plottings was 71,63 µm (SPEC: 58,83 µm, CIRR: 90,92 µm, TOP2: 78,9 µm, NID3: 56,88 µm). The mean distance between the true CP and the device’s CP was 290,85 µm (SPEC: 190,5 µm, CIRR: 248,28 µm, TOP2: 530,9 µm, NID3: 193,71 µm). This inaccuracy of CP plotting resulted in mean thickness errors of 19,76 µm (CMT) and 37,44 µm (CPT). Thickness deviation was significantly different between the four devices (CMT/CPT; SPEC: 17,38/41,55 µm, CIRR: 16,93/31,93 µm, TOP2: 34,1/55,76 µm, NID3: 10,66/20,52 µm). The maximum deviations were (CMT/CPT) for SPEC 177/479 µm, CIRR 137/173 µm, TOP2 141/197 µm, NID3 48/96 µm.
Conclusions: :
In all tested devices inaccuracy of CP plotting was frequent and resulted in clinically relevant CMT and CPT deviations. However there is substantial inter-device variability in plotting performance. Manual replotting of the device’s CP is necessary to control for thickness deviations that might confound clinical decisions.
Keywords: imaging/image analysis: clinical • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • clinical research methodology