Retinal Shape-Based Classification of Retinal Detachment and Posterior Vitreous Detachment Eyes

Artificial intelligence-assisted management of retinal detachment from ultra-widefield fundus images based on weakly-supervised approach

Background

Retinal detachment (RD) is a common sight-threatening condition in the emergency department. Early postural intervention based on detachment regions can improve visual prognosis.

Methods

We developed a weakly supervised model with 24,208 ultra-widefield fundus images to localize and coarsely outline the anatomical RD regions. The customized preoperative postural guidance was generated for patients accordingly. The localization performance was then compared with the baseline model and an ophthalmologist according to the reference standard established by the retina experts.

Results

In the 48-partition lesion detection, our proposed model reached an 86.42% (95% confidence interval (CI): 85.81-87.01%) precision and an 83.27% (95%CI: 82.62-83.90%) recall with an average precision (PA) of 0.9132. In contrast, the baseline model achieved a 92.67% (95%CI: 92.11-93.19%) precision and limited recall of 68.07% (95%CI: 67.25-68.88%). Our holistic lesion localization performance was comparable to the ophthalmologist's 89.16% (95%CI: 88.75-89.55%) precision and 83.38% (95%CI: 82.91-83.84%) recall. As to the performance of four-zone anatomical localization, compared with the ground truth, the un-weighted Cohen's κ coefficients were 0.710(95%CI: 0.659-0.761) and 0.753(95%CI: 0.702-0.804) for the weakly-supervised model and the general ophthalmologist, respectively.

Conclusion

The proposed weakly-supervised deep learning model showed outstanding performance comparable to that of the general ophthalmologist in localizing and outlining the RD regions. Hopefully, it would greatly facilitate managing RD patients, especially for medical referral and patient education.

Swept-Source OCT Mid-Peripheral Retinal Irregularity in Retinal Detachment and Posterior Vitreous Detachment Eyes

Irregularities in retinal shape have been shown to correlate with axial length, a major risk factor for retinal detachment. To further investigate this association, a comparison was performed of the swept-source optical coherence tomography (SS OCT) peripheral retinal shape of eyes that had either a posterior vitreous detachment (PVD) or vitrectomy for retinal detachment. The objective was to identify a biomarker that can be tested as a predictor for retinal detachment. Eyes with a PVD (N = 88), treated retinal detachment (N = 67), or retinal tear (N = 53) were recruited between July 2020 and January 2022 from hospital retinal clinics in South Australia. The mid-peripheral retina was imaged in four quadrants with SS OCT. The features explored were patient age, eye axial length, and retinal shape irregularity quantified in the frequency domain. A discriminant analysis classifier to identify retinal detachment eyes was trained with two-thirds and tested with one-third of the sample. Retinal detachment eyes had greater irregularity than PVD eyes. A classifier trained using shape features from the superior and temporal retina had a specificity of 84% and a sensitivity of 48%. Models incorporating axial length were less successful, suggesting peripheral retinal irregularity is a better biomarker for retinal detachment than axial length. Mid-peripheral retinal irregularity can identify eyes that have experienced a retinal detachment.