A Comparison of Optical Coherence Tomography Angiography Metrics and Artifacts on Scans of Different Sizes in Diabetic Macular Ischemia: The effect of different size scans on DMI evaluation.
Am J Ophthalmol
; 2024 Sep 13.
Article
en En
| MEDLINE
| ID: mdl-39278388
ABSTRACT
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 3â¯×â¯3 versus 6â¯×â¯6 mm optical coherence tomography angiography (OCTA) macular scans in patients with DMI.DESIGN:
Prospective, comparative image quality analysis with one-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 3â¯×â¯3 and 6â¯×â¯6 mm scans at baseline were included in the image quality analysis. The 6â¯×â¯6 mm scan presented with more severe motion artifact (Pâ¯=â¯.02). Conversely, the 3â¯×â¯3 mm scans were more susceptible to mild decentration (Pâ¯=â¯.009). After removing all the poor-quality images, 55 eyes with both size scans entered the longitudinal analysis. The 3â¯×â¯3 mm FAZ was significantly larger than the 6â¯×â¯6 mm FAZ using either aFAZ or cFAZ (both P < .05). In contrast, the 6â¯×â¯6 mm wiSVD and wiDVD were remarkably higher than those on the 3â¯×â¯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 3â¯×â¯3 mm aFAZ decreased numerically at 52 weeks (Pâ¯=â¯.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 3â¯×â¯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.
Texto completo:
1
Colección:
01-internacional
Base de datos:
MEDLINE
Idioma:
En
Revista:
Am J Ophthalmol
Año:
2024
Tipo del documento:
Article
Pais de publicación:
Estados Unidos