Ocular imaging at the cutting-edge

Real-time OCT feedback-controlled RPE photodisruption in ex vivo porcine eyes using 8 microsecond laser pulses

Selective retinal pigment epithelium (RPE) photodisruption requires reliable real-time feedback dosimetry (RFD) to prevent unwanted overexposure. In this study, optical coherence tomography (OCT) based RFD was investigated in ex vivo porcine eyes exposed to laser pulses of 8 µs duration (wavelength: 532 nm, exposure area: 90 × 90 µm2, radiant exposure: 247 to 1975 mJ/µm2). For RFD, fringe washouts in time-resolved OCT M-scans (central wavelength: 870 nm, scan rate: 85 kHz) were compared to an RPE cell viability assay. Statistical analysis revealed a moderate correlation between RPE lesion size and applied treatment energy, suggesting RFD adaptation to inter- and intraindividual RPE pigmentation and ocular transmission.

Bessel Beams in Ophthalmology: A Review

The achievable resolution of a conventional imaging system is inevitably limited due to diffraction. Dealing with precise imaging in scattering media, such as in the case of biomedical imaging, is even more difficult owing to the weak signal-to-noise ratios. Recent developments in non-diffractive beams such as Bessel beams, Airy beams, vortex beams, and Mathieu beams have paved the way to tackle some of these challenges. This review specifically focuses on non-diffractive Bessel beams for ophthalmological applications. The theoretical foundation of the non-diffractive Bessel beam is discussed first followed by a review of various ophthalmological applications utilizing Bessel beams. The advantages and disadvantages of these techniques in comparison to those of existing state-of-the-art ophthalmological systems are discussed. The review concludes with an overview of the current developments and the future perspectives of non-diffractive beams in ophthalmology.

Spotlight on Targeted Chemotherapy in Retinoblastoma: Safety, Efficacy, and Patient Outcomes

As the most common primary intraocular malignancy of childhood, retinoblastoma (RB) has had a complex journey in its management, following a course from enucleation as the first life-saving treatment to numerous globe-salvaging therapies during the last century. Currently, this potentially lethal disease has achieved high survival rates owing to multidisciplinary management and the introduction of neoadjuvant and multimodal chemotherapy. Therefore, the goal of treatment is shifting toward conserving the globe and vision as much as possible. Up until recently, many advanced cases of RB were enucleated primarily; however, targeted chemotherapy via the ophthalmic artery and management of intraocular seeding by local administration of chemotherapeutic agents have revolutionized the globe-conserving therapies. The added benefit of avoiding systemic complications of cytotoxic drugs resulted in these methods gaining popularity, and they are becoming a main part of care in many referral centers. Initially, there were some safety concerns regarding these approaches; however, increasing experience has shown that these modalities are relatively safe procedures and many complications can be averted by changing the choice of the drug and using some prophylactic measures. It is hoped that, in the near future, with advances in early diagnosis and patient-targeted molecular therapies, as well as gene-editing techniques, the patient's vision can be saved even in advanced RB.

Predicting Systemic Health Features from Retinal Fundus Images Using Transfer-Learning-Based Artificial Intelligence Models

While color fundus photos are used in routine clinical practice to diagnose ophthalmic conditions, evidence suggests that ocular imaging contains valuable information regarding the systemic health features of patients. These features can be identified through computer vision techniques including deep learning (DL) artificial intelligence (AI) models. We aim to construct a DL model that can predict systemic features from fundus images and to determine the optimal method of model construction for this task. Data were collected from a cohort of patients undergoing diabetic retinopathy screening between March 2020 and March 2021. Two models were created for each of 12 systemic health features based on the DenseNet201 architecture: one utilizing transfer learning with images from ImageNet and another from 35,126 fundus images. Here, 1277 fundus images were used to train the AI models. Area under the receiver operating characteristics curve (AUROC) scores were used to compare the model performance. Models utilizing the ImageNet transfer learning data were superior to those using retinal images for transfer learning (mean AUROC 0.78 vs. 0.65, p-value < 0.001). Models using ImageNet pretraining were able to predict systemic features including ethnicity (AUROC 0.93), age > 70 (AUROC 0.90), gender (AUROC 0.85), ACE inhibitor (AUROC 0.82), and ARB medication use (AUROC 0.78). We conclude that fundus images contain valuable information about the systemic characteristics of a patient. To optimize DL model performance, we recommend that even domain specific models consider using transfer learning from more generalized image sets to improve accuracy.

Fractal dimension of retinal vasculature as an image quality metric for automated fundus image analysis systems

Automated fundus screening is becoming a significant programme of telemedicine in ophthalmology. Instant quality evaluation of uploaded retinal images could decrease unreliable diagnosis. In this work, we propose fractal dimension of retinal vasculature as an easy, effective and explainable indicator of retinal image quality. The pipeline of our approach is as follows: utilize image pre-processing technique to standardize input retinal images from possibly different sources to a uniform style; then, an improved deep learning empowered vessel segmentation model is employed to extract retinal vessels from the pre-processed images; finally, a box counting module is used to measure the fractal dimension of segmented vessel images. A small fractal threshold (could be a value between 1.45 and 1.50) indicates insufficient image quality. Our approach has been validated on 30,644 images from four public database.