A study recently published in Ophthalmology and Therapy analyzed diffuse chorioretinal atrophy (DCA) in pathologic myopia to develop a standardized classification system using artificial intelligence (AI).
Give me some background first.
Pathologic myopia has become one of the major causes of low vision in the working-aged and elderly populations.
Most notably: DCA is recognized as the initial stage of pathologic myopia and plays a crucial role in its progression, according to the International Photographic Classification and Grading System for Myopic Maculopathy (META-PM).
Keep going…
The appearance of DCA is characterized by a yellow-white lesion with an indistinct boundary located at the posterior pole of the eye, along with blurred choroidal vessels within the atrophic area.
Advances in AI-based imaging processing technology have allowed for the quantitative measurement of DCA area and density through the analysis of fundus photography.
Now talk about the study.
In this single-center, retrospective observational study, investigators performed a comprehensive analysis of 202 patients (338 eyes) diagnosed with high myopia to develop a new grading system for pathologic myopia.
High myopia was defined as spherical equivalent (SE) <-6 D and/or axial length (AL) >26 mm.
How did they develop the grading system?
To establish the grading system, investigators:
- Preprocessed fundus images
- Labeled samples
- Used deep learning segmentation models to analyze the images
- Measured and calculated the area and density of DCA lesions
Hierarchical clustering was used to categorize diffuse atrophy on fundus into three groups based on the area and density of diffuse atrophy (G1, G2, G3), and G0 was reserved for patients with high myopia without diffuse atrophy on fundus.
Findings?
Based on the area and density of DCA in pathologic myopia, the condition was categorized into four grades:
- G0: No diffuse atrophy
- G1
- DCA area: 0 < area < 10.170786 mm2
- DCA density: 0 < density ≤ 0.093
- G2
- DCA area: 10.170786 < area < 27.034940 mm2
- DCA density: 0.093 < density < 0.245
- G3:
- DCA area: 27.034940 < area < 77.894083 mm2
- DCA density: 0.245 < density < 0.712
Talk about the clinical characteristics of DCA.
Fundus photographs showed a progressive enlargement of atrophic lesions, evolving from punctate-shaped to patchy with indistinct boundaries with increasing pathologic myopia severity.
DCA atrophy lesions exhibited a gradual shift in color from brown-yellow to yellow-white, originating from the temporal side of the optic disc and extending towards the macula; severe cases showed widespread distribution throughout the posterior pole.
Who tended to have DCA?
In comparison to patients without DCA (G0), patients with DCA were (P < 0.001):
- Significantly older
- DCA: 34 years (range 27-48)
- Without DCA: 29 years (range 26-34 years)
- Had a longer AL
- DCA: 28.85 ± 1.57 mm
- Without DCA: 27.11 ± 1.01 mm
- Exhibited a more myopic SE
- DCA: -13 D (range: -16 to -10.5 D)
- Without DCA: -9.09 ± 2.41 D
Anything else?
In eyes with DCA, a trend emerged as grades increased from G1 to G3, showing associations with (P<0.001):
- Older age
- Longer axial length
- Deeper myopic SE
- Larger area of parapapillary atrophy
- Increased fundus tessellated density (FTD)
Expert opinion?
Per the study authors, “In G1 and G2, the lesions were predominantly localized in the temporal side of the optic disc, with fewer observed in the nasal side of the optic disc and the temporal region of the macula."
“Upon reaching G3, atrophic lesions were distributed throughout the entire posterior pole of the eyeball,” they added.
Take home.
These findings suggest that this new grading system for pathologic myopia leveraging DCA area and density measurements is a reliable measure for evaluating the severity of DCA—making it suitable for broad application in the screening of pathologic myopia.
Next steps?
Due to the prospective nature of this study, assessing the potential clinical applications and validating this classification system requires multicenter clinical cohort studies in the future to ensure broader generalizability and thorough testing in real-world clinical settings.