Automation of dry eye disease quantitative assessment: A review

Role of artificial intelligence, machine learning and deep learning models in corneal disorders - A narrative review

In the last decade, artificial intelligence (AI) has significantly impacted ophthalmology, particularly in managing corneal diseases, a major reversible cause of blindness. This review explores AI's transformative role in the corneal subspecialty, which has adopted advanced technology for superior clinical judgment, early diagnosis, and personalized therapy. While AI's role in anterior segment diseases is less documented compared to glaucoma and retinal pathologies, this review highlights its integration into corneal diagnostics through imaging techniques like slit-lamp biomicroscopy, anterior segment optical coherence tomography (AS-OCT), and in vivo confocal biomicroscopy. AI has been pivotal in refining decision-making and prognosis for conditions such as keratoconus, infectious keratitis, and dystrophies. Multi-disease deep learning neural networks (MDDNs) have shown diagnostic ability in classifying corneal diseases using AS-OCT images, achieving notable metrics like an AUC of 0.910. AI's progress over two decades has significantly improved the accuracy of diagnosing conditions like keratoconus and microbial keratitis. For instance, AI has achieved a 90.7% accuracy rate in classifying bacterial and fungal keratitis and an AUC of 0.910 in differentiating various corneal diseases. Convolutional neural networks (CNNs) have enhanced the analysis of color-coded corneal maps, yielding up to 99.3% diagnostic accuracy for keratoconus. Deep learning algorithms have also shown robust performance in detecting fungal hyphae on in vivo confocal microscopy, with precise quantification of hyphal density. AI models combining tomography scans and visual acuity have demonstrated up to 97% accuracy in keratoconus staging according to the Amsler-Krumeich classification. However, the review acknowledges the limitations of current AI models, including their reliance on binary classification, which may not capture the complexity of real-world clinical presentations with multiple coexisting disorders. Challenges also include dependency on data quality, diverse imaging protocols, and integrating multimodal images for a generalized AI diagnosis. The need for interpretability in AI models is emphasized to foster trust and applicability in clinical settings. Looking ahead, AI has the potential to unravel the intricate mechanisms behind corneal pathologies, reduce healthcare's carbon footprint, and revolutionize diagnostic and management paradigms. Ethical and regulatory considerations will accompany AI's clinical adoption, marking an era where AI not only assists but augments ophthalmic care.

Automated tear film break-up time measurement for dry eye diagnosis using deep learning

In the realm of ophthalmology, precise measurement of tear film break-up time (TBUT) plays a crucial role in diagnosing dry eye disease (DED). This study aims to introduce an automated approach utilizing artificial intelligence (AI) to mitigate subjectivity and enhance the reliability of TBUT measurement. We employed a dataset of 47 slit lamp videos for development, while a test dataset of 20 slit lamp videos was used for evaluating the proposed approach. The multistep approach for TBUT estimation involves the utilization of a Dual-Task Siamese Network for classifying video frames into tear film breakup or non-breakup categories. Subsequently, a postprocessing step incorporates a Gaussian filter to smooth the instant breakup/non-breakup predictions effectively. Applying a threshold to the smoothed predictions identifies the initiation of tear film breakup. Our proposed method demonstrates on the evaluation dataset a precise breakup/non-breakup classification of video frames, achieving an Area Under the Curve of 0.870. At the video level, we observed a strong Pearson correlation coefficient (r) of 0.81 between TBUT assessments conducted using our approach and the ground truth. These findings underscore the potential of AI-based approaches in quantifying TBUT, presenting a promising avenue for advancing diagnostic methodologies in ophthalmology.

Comparing automated and manual assessments of tear break-up time using different non-invasive devices and a fluorescein procedure

To evaluate the agreement and repeatability of an automated topography-based method for non-invasive break-up time (NIBUT) analyses in comparison with two other NIBUT procedures, the fluorescein procedure (fBUT), and with the manual assessment with the same device. In the first experiment, a semi-randomised crossover study was performed on forty-three participants (23.1 ± 2.1 years). NIBUT measurements were collected in a randomised order, in both eyes of participants with EasyTear View + (Easytear, Rovereto), Polaris, and Sirius + (CSO, Firenze). Then a fBUT was collected. The overall measurement procedure was repeated in a further session (retest) on the same day. In a second experiment, a retrospective randomised crossover study was performed on eighty-five NIBUT videos previously recorded by the Sirius+. Two observers assessed manually the videos and the NIBUTs were compared with the automatic ones. In the first experiment, ANOVA showed a significant difference between the four measures in both eyes (p < 0.001). Significant differences were found in the paired comparisons between each NIBUT procedure and fBUT (Wicoxon; p < 0.05). Sirius+ resulted in agreement only with Polaris in the left eye. Correlations between all NIBUT procedures resulted in statistical significance in both eyes. All procedures showed very good test-rest reliability. In the second experiment, a significant correlation between automated and manual NIBUT was found, but also a significant statistical difference between the two measurements, although clinically negligible (0.3 s). The investigated NIBUT devices perform differently from each other (and from fBUT), so they cannot be considered interchangeable. The automated measure of NIBUT with Sirius+ has a negligible clinical difference compared to manual assessment on the same device.

Advancements in Nanogels for Enhanced Ocular Drug Delivery: Cutting-Edge Strategies to Overcome Eye Barriers

Nanomedicine in gel or particle formation holds considerable potential for enhancing passive and active targeting within ocular drug delivery systems. The complex barriers of the eye, exemplified by the intricate network of closely connected tissue structures, pose significant challenges for drug administration. Leveraging the capability of engineered nanomedicine offers a promising approach to enhance drug penetration, particularly through active targeting agents such as protein peptides and aptamers, which facilitate targeted release and heightened bioavailability. Simultaneously, DNA carriers have emerged as a cutting-edge class of active-targeting structures, connecting active targeting agents and illustrating their potential in ocular drug delivery applications. This review aims to consolidate recent findings regarding the optimization of various nanoparticles, i.e., hydrogel-based systems, incorporating both passive and active targeting agents for ocular drug delivery, thereby identifying novel mechanisms and strategies. Furthermore, the review delves into the potential application of DNA nanostructures, exploring their role in the development of targeted drug delivery approaches within the field of ocular therapy.

Dynamic Aspects of Pre-Soft Contact Lens Tear Film and Their Relation to Dry Eye: Basic Science and Clinical Relevance

Soft contact lens (SCL) perturbs the intimate connection between the pre-lens tear film (PLTF) and the ocular surface in various ways, i.e., (i) decrease in tear meniscus radius and aqueous tear thickness, (ii) attenuation of tear film lipid layer spread, (iii) limited wettability of SCL surface, (iv) increased friction with eyelid wiper, etc. This often results in SCL-related dry eye (SCLRDE) manifested as PLTF instability and contact lens discomfort (CLD). In this review, the individual contributions of factors (i-iv) to PLTF breakup patterns (BUP) and CLD are considered via the tear film-oriented diagnosis framework adopted by the Asia Dry Eye Society from a clinical and basic science perspective. It is shown that SCLRDE (due to aqueous deficiency, increased evaporation, or decreased wettability) and BUP of PLTF classify within the same types as the ones observed for the precorneal tear film. The analysis of PLTF dynamics reveals that the inclusion of SCL enhances the manifestation of BUP associated with (i) decreased thickness of PLTF aqueous layer and (ii) limited SCL wettability as shown by the rapid expansion of BUP area. PLTF thinness and instability result in increased blink-related friction and lid wiper epitheliopathy as major contributor to CLD.

The efficacy of ocular surface assessment approaches in evaluating dry eye treatment with artificial tears

This study evaluates the effectiveness of objective techniques in assessing dry eye disease (DED) treatment compared with the subjective assessment commonly used in clinical practice. Thirty subjects were recruited for two visits separated by 28(± 3) days of treatment with artificial tears. A buttery of common subjective assessment methods were accompanied by a set of objective techniques including measurement of noninvasive tear film break-up time (NIBUT), lipid layer thickness (LLT), and quantitative evaluation of tear film surface quality and dynamics (TFD). Additionally, meibography was performed. Two commercially available videokeratoscopes and a prototype of a lateral shearing interferometer were used for the measurements. Both subjective and objective techniques showed a positive effect of artificial tears in DED treatment. Statistically significant improvements were observed in subjective symptoms (from P < 0.001 for Ocular Surface Disease Index, OSDI to p = 0.019 for tearing), conjunctival redness (P = 0.022), ocular staining (P = 0.012), fluorescein tear film break-up time (P = 0.015), NIBUT (P = 0.037), LLT (P < 0.001), and TFD (P = 0.048). In general, weak or statistically insignificant correlations were observed between subjective and objective assessment methods. The apparent lack of correlation between these methods might indicate the complementary character of objective techniques that likely assess other characteristics of ocular surface health than those assessed subjectively.