Doheny Image Reading Center, Doheny Eye Institute, Los Angeles, CA, USA.
Department of Ophthalmology, David Geffen School of Medicine at the University of California, Los Angeles, Los Angeles, CA, USA.
Acta Ophthalmol. 2018 Nov;96(7):e820-e827. doi: 10.1111/aos.13740. Epub 2018 May 31.
To investigate the effect of multiple enface image averaging on image quality of the optical coherence tomography angiography (OCTA).
Twenty-one normal volunteers were enrolled in this study. For each subject, one eye was imaged with 3 × 3 mm scan protocol, and another eye was imaged with the 6 × 6 mm scan protocol centred on the fovea using the ZEISS Angioplex™ spectral-domain OCTA device. Eyes were repeatedly imaged to obtain nine OCTA cube scan sets, and nine superficial capillary plexus (SCP) and deep capillary plexus (DCP) were individually averaged after registration.
Eighteen eyes with a 3 × 3 mm scan field and 14 eyes with a 6 × 6 mm scan field were studied. Averaged images showed more continuous vessels and less background noise in both the SCP and the DCP as the number of frames used for averaging increased, with both 3 × 3 and 6 × 6 mm scan protocols. The intensity histogram of the vessels dramatically changed after averaging. Contrast-to-noise ratio (CNR) and subjectively assessed image quality scores also increased as the number of frames used for averaging increased in all image types. However, the additional benefit in quality diminished when averaging more than five frames. Averaging only three frames achieved significant improvement in CNR and the score assigned by certified grades.
Use of multiple image averaging in OCTA enface images was found to be both objectively and subjectively effective for enhancing image quality. These findings may of value for developing optimal OCTA imaging protocols for future studies.
研究多次 Enface 图像平均对光相干断层扫描血管造影（OCTA）图像质量的影响。
本研究纳入 21 名正常志愿者。对每位受试者，一只眼采用 3×3mm 扫描方案进行成像，另一只眼采用以黄斑为中心的 6×6mm 扫描方案进行成像，使用 ZEISS Angioplex™ 光谱域 OCTA 设备。对眼睛进行重复成像，以获得 9 个 OCTA 立方扫描集，并在注册后分别对 9 个浅层毛细血管丛（SCP）和深层毛细血管丛（DCP）进行平均。
对 18 只具有 3×3mm 扫描视野的眼睛和 14 只具有 6×6mm 扫描视野的眼睛进行了研究。随着平均所用帧数的增加，在 SCP 和 DCP 中，平均图像显示出更连续的血管和更少的背景噪声，两种扫描方案均如此。血管的强度直方图在平均后发生显著变化。在所有图像类型中，随着平均所用帧数的增加，对比噪声比（CNR）和主观评估的图像质量评分也随之增加。然而，当平均超过 5 帧时，质量的额外增益会减少。仅平均 3 帧即可显著提高 CNR 和经认证等级分配的评分。
在 OCTA Enface 图像中使用多次图像平均被发现对增强图像质量具有客观和主观的有效性。这些发现可能对未来研究中开发最佳 OCTA 成像方案具有价值。
