Duke Eye Center, Duke University Medical Center, Durham, North Carolina, USA.
Ophthalmology. 2010 Jun;117(6):1177-1189.e4. doi: 10.1016/j.ophtha.2009.10.029. Epub 2010 Feb 19.
To characterize the types and frequencies of image artifacts associated with macular scanning using 2 common spectral-domain optical coherence tomography (SD OCT) instruments and to evaluate the impact of artifacts on foveal thickness measurements.
Retrospective, observational chart review.
For the Cirrus HD-OCT (Carl Zeiss Meditec, Dublin, CA), scans of 98 eyes from 58 patients were included in the study. For the Spectralis HRA+OCT (Heidelberg Engineering, Heidelberg, Germany), scans of 88 eyes from 54 patients were included.
Macular volume scans of healthy and diseased eyes were evaluated systematically for image artifacts within each scan overall and within the center 1-mm area. The frequency of each artifact type was compared for scans stratified by diagnosis category. Artifacts in the center 1-mm area were graded for severity and were corrected manually using each instrument's software. Artifacts that resulted in errors of more than 50 microm or more than 10% of retinal thickness or that caused a misdiagnosis of macular edema or retinal thinning were defined as clinically significant and were analyzed further.
Overall frequency of image artifacts by artifact type, relative frequency of artifacts in scans stratified by posterior segment disease diagnosis, and retinal thickness measurements of the center 1-mm subfield before and after artifact corrections.
For Cirrus, 84.7% of scans had artifacts and 32.7% had at least 1 artifact in the center 1-mm area of the scan. For Spectralis, 90.9% of scans had at least 1 artifact, and 37.5% had at least 1 artifact in the center 1-mm area. Certain artifact types were observed more frequently with specific disease states. Clinically significant artifacts involving the center 1-mm area were seen in 5.1% of Cirrus and 8.0% of Spectralis scans.
Image artifacts in SD OCT volume scanning are common and frequently involve segmentation errors. Artifacts are relatively less common in the center 1-mm area of scans, but may affect retinal thickness measurements in a clinically significant manner. Careful review of scans for artifacts is important when using SD OCT images and retinal thickness measurements in patient care or clinical trials.
描述使用两种常见的谱域光学相干断层扫描（SD OCT）仪器进行黄斑扫描时的图像伪影类型和频率，并评估伪影对中心凹厚度测量的影响。
回顾性、观察性图表审查。
对于 Cirrus HD-OCT（Carl Zeiss Meditec，都柏林，CA），纳入了 58 例 98 只眼的扫描结果；对于 Spectralis HRA+OCT（Heidelberg Engineering，海德堡，德国），纳入了 54 例 88 只眼的扫描结果。
对健康和患病眼的黄斑容积扫描进行系统评估，以确定整个扫描以及中心 1mm 区域内的图像伪影。比较按诊断类别分层的扫描中每种伪影类型的频率。对中心 1mm 区域内的伪影进行严重程度分级，并使用每种仪器的软件手动校正。将导致误差超过 50μm 或视网膜厚度超过 10%，或导致黄斑水肿或视网膜变薄误诊的伪影定义为临床显著，并进一步分析。
按伪影类型分类的总体图像伪影频率、按后部节段疾病诊断分层的扫描中伪影的相对频率、以及中心 1mm 子域的视网膜厚度测量值在进行伪影校正前后的变化。
Cirrus 中，84.7%的扫描存在伪影，32.7%的扫描在中心 1mm 区域存在至少 1 种伪影。Spectralis 中，90.9%的扫描存在至少 1 种伪影，37.5%的扫描在中心 1mm 区域存在至少 1 种伪影。某些伪影类型与特定疾病状态下观察到的更为频繁。Cirrus 和 Spectralis 扫描中分别有 5.1%和 8.0%存在涉及中心 1mm 区域的临床显著伪影。
SD OCT 容积扫描中的图像伪影很常见，并且经常涉及分割错误。在扫描的中心 1mm 区域中，伪影相对较少，但可能以临床显著的方式影响视网膜厚度测量值。在使用 SD OCT 图像和视网膜厚度测量值进行患者护理或临床试验时，仔细检查扫描中的伪影非常重要。
