糖尿病黄斑缺血患者不同尺寸光学相干断层扫描血管成像扫描的指标与伪影比较
该前瞻性影像质量分析纳入57例(81眼)有糖尿病黄斑缺血证据且完成3×3与6×6 mm光学相干断层扫描血管成像配对扫描的患者,比较两种尺寸扫描的伪影频率与中心凹无血管区指标。6×6 mm扫描运动伪影更严重(P = .02),3×3 mm扫描更易出现轻度偏心(P = .009);3×3 mm中心凹无血管区更大而6×6 mm血管密度更高。作者认为糖尿病黄斑缺血中中心凹无血管区指标最好在3×3 mm扫描上评估,但需手动校正中心凹无血管区边界。
要点
- 6×6 mm扫描运动伪影更严重(P = .02),3×3 mm扫描更易出现轻度偏心(P = .009)。
- 3×3 mm扫描的自动与校正中心凹无血管区均显著大于6×6 mm扫描(均P < .05)。
- 6×6 mm扫描的全图像浅层与深层血管密度均显著高于3×3 mm扫描(均P < .001)。
核心结论
- 在糖尿病黄斑缺血中,中心凹无血管区指标最好在3×3 mm扫描上评估,因为分辨率更好,但需要手动校正中心凹无血管区边界。
- 伪影频率和自动中心凹无血管区勾画错误提示需要进一步技术改进。
研究设计与样本
- 前瞻性、比较性影像质量分析,随访1年。
- 招募有糖尿病视网膜病变且光学相干断层扫描血管成像显示糖尿病黄斑缺血证据的患者,定义为自动中心凹无血管区≥0.5 mm²或旁中心凹毛细血管无灌注≥1个象限(当自动中心凹无血管区<0.5 mm²时)。
- 仅纳入同时具有两种尺寸扫描的患者进行分析;图像质量分析纳入57例(81眼)基线配对3×3与6×6 mm扫描。
- 排除所有质量差的图像后,55眼同时具有两种尺寸扫描进入纵向分析。
主要发现
- 6×6 mm扫描运动伪影更严重(P = .02),而3×3 mm扫描更易出现轻度偏心(P = .009)。
- 使用自动或校正中心凹无血管区,3×3 mm的中心凹无血管区均显著大于6×6 mm(均P < .05)。
- 6×6 mm的全图像浅层血管密度和全图像深层血管密度均显著高于3×3 mm扫描(均P < .001)。
- 两种尺寸扫描的校正中心凹无血管区在1年内均稳步增加(均P < .01),但3×3 mm的自动中心凹无血管区在52周时数值下降(P = .02)。
- 回顾所有扫描后,旁中心凹毛细血管无灌注识别不佳是自动中心凹无血管区勾画错误最常见的原因。
关键数值与推荐
- 糖尿病黄斑缺血定义为自动中心凹无血管区≥0.5 mm²或旁中心凹毛细血管无灌注≥1个象限(当自动中心凹无血管区<0.5 mm²时)。
- 3×3 mm扫描的自动中心凹无血管区与校正中心凹无血管区均显著大于6×6 mm扫描(均P < .05)。
- 6×6 mm扫描的全图像浅层血管密度与全图像深层血管密度均显著高于3×3 mm扫描(均P < .001)。
- 6×6 mm扫描运动伪影更严重(P = .02),3×3 mm扫描更易出现轻度偏心(P = .009)。
临床怎么用
- 作者认为在糖尿病黄斑缺血中中心凹无血管区指标最好在3×3 mm扫描上评估,因为分辨率更好,但需要手动校正中心凹无血管区边界。
- 作者指出伪影频率和自动中心凹无血管区勾画错误提示需要进一步技术改进。
局限性
- 原文未报告局限性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
PURPOSE: Changes in the foveal avascular zone (FAZ) metrics over time are key outcome measures for clinical trials in diabetic macular ischemia (DMI). However, artifacts and automatically delineated FAZ measurements may influence the results. We aimed to compare the artifact frequency and FAZ metrics on 3 × 3 versus 6 × 6 mm optical coherence tomography angiography (OCTA) macular scans in patients with DMI. DESIGN: Prospective, comparative image quality analysis with 1-year follow-up. METHODS: Patients with diabetic retinopathy (DR) were recruited if they presented with OCTA evidence of DMI, defined as an automated FAZ (aFAZ) ≥0.5 mm2 or parafoveal capillary nonperfusion (CNP) ≥1 quadrant if the aFAZ <0.5 mm2. Only those who had both size scans were included in the analysis. The types of artifacts and FAZ delineation errors were graded before manual correction. After excluding scans with poor quality, the aFAZ, corrected FAZ (cFAZ), whole image superficial vessel density (wiSVD), and whole image deep vessel density (wiDVD) were compared on both size scans. RESULTS: Fifty-seven patients (81 eyes) with paired OCTA 3 × 3 and 6 × 6 mm scans at baseline were included in the image quality analysis. The 6 × 6 mm scan presented with more severe motion artifact (P = .02). Conversely, the 3 × 3 mm scans were more susceptible to mild decentration (P = .009). After removing all the poor-quality images, 55 eyes with both size scans entered the longitudinal analysis. The 3 × 3 mm FAZ was significantly larger than the 6 × 6 mm FAZ using either aFAZ or cFAZ (both P < .05). In contrast, the 6 × 6 mm wiSVD and wiDVD were remarkably higher than those on the 3 × 3 mm scans (both P < .001). There was a steady increase in cFAZ over one year on both size scans (both P < .01). However, the 3 × 3 mm aFAZ decreased numerically at 52 weeks (P = .02). After reviewing all the scans, poor identification of parafoveal CNP was the most common reason for erroneous aFAZ delineation. CONCLUSIONS: In DMI, the FAZ metrics are best evaluated on the 3 × 3 scan due to better resolution. However, manual correction of the FAZ margin is needed. The frequency of artifacts and aFAZ delineation errors suggest that further technical refinement is required.
引用
- Tsai WS, Thottarath S, Gurudas S, et al. A Comparison of Optical Coherence Tomography Angiography Metrics and Artifacts on Scans of Different Sizes in Diabetic Macular Ischemia. Am J Ophthalmol. 2024. PMID: 39278388. doi:10.1016/j.ajo.2024.09.012
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。