Recent Research in Ocular Cystinosis: Drug Delivery Systems, Cysteamine Detection Methods and Future Perspectives.

Cystinosis is a rare genetic disorder characterized by the accumulation of cystine crystals in different tissues and organs. Although renal damage prevails during initial stages, the deposition of cystine crystals in the cornea causes severe ocular manifestations. At present, cysteamine is the only topical effective treatment for ocular cystinosis. The lack of investment by the pharmaceutical industry, together with the limited stability of cysteamine, make it available only as two marketed presentations (Cystaran® and Cystadrops®) and as compounding formulations prepared in pharmacy departments. Even so, new drug delivery systems (DDSs) need to be developed, allowing more comfortable dosage schedules that favor patient adherence. In the last decades, different research groups have focused on the development of hydrogels, nanowafers and contact lenses, allowing a sustained cysteamine release. In parallel, different determination methods and strategies to increase the stability of the formulations have also been developed. This comprehensive review aims to compile all the challenges and advances related to new cysteamine DDSs, analytical determination methods, and possible future therapeutic alternatives for treating cystinosis.

On the development of conjunctival hyperemia computer-assisted diagnosis tools: Influence of feature selection and class imbalance in automatic gradings.

OBJECTIVE: The sudden increase of blood flow in the bulbar conjunctiva, known as hyperemia, is associated to a red hue of variable intensity. Experts measure hyperemia using levels in a grading scale, a procedure that is subjective, non-repeatable and time consuming, thus creating a need for its automatisation. However, the task is far from straightforward due to data issues such as class imbalance or correlated features. In this paper, we study the specific features of hyperemia and propose various approaches to address these problems in the context of an automatic framework for hyperemia grading. METHODOLOGY: Oversampling, undersampling and SMOTE approaches were applied in order to tackle the problem of class imbalance. 25 features were computed for each image and regression methods were then used to transform them into a value on the grading scale. The values and relationships among features and experts' values were analysed, and five feature selection techniques were subsequently studied. RESULTS: The lowest mean square error (MSE) for the regression systems trained with individual features is below 0.1 for both scales. Multi-layer perceptron (MLP) obtains the best values, but is less consistent than the random forest (RF) method. When all features are combined, the best results for both scales are achieved with MLP. Correlation based feature selection (CFS) and M5 provide the best results, MSE=0.108 and MSE=0.061 respectively. Finally, the class imbalance problem is minimised with the SMOTE approach for both scales (MSE<0.006). CONCLUSIONS: Machine learning methods are able to perform an objective assessment of hyperemia grading, removing both intra- and inter-expert subjectivity while providing a gain in computation time. SMOTE and oversampling approaches minimise the class imbalance problem, while feature selection reduces the number of features from 25 to 3-5 without worsening the MSE. As the differences between the system and a human expert are similar to the differences between experts, we can therefore conclude that the system behaves like an expert.

New software application for clarifying tear film lipid layer patterns.

PURPOSE: The lipid layer of the tear film limits evaporation during the interblink interval and also affects tear stability. This study was designed to validate a new software application designed to characterize the tear film lipid layer through texture and color pattern recognition. METHODS: Using the Tearscope-plus (slit-lamp magnification ×200), the lipid layer was examined in 105 healthy young adults, and interference photographs were acquired with a Topcon DV-3 digital camera. The photographs were classified by a clinician experienced in examining lipid layer patterns (LLPs), and these classifications were used as the reference standard (reference examiner). Next, LLPs were graded using the new software and further by 2 observers (observer 1 and observer 2) with experience in examining the ocular surface. RESULTS: Strong correlation was detected between the categories determined by the new application and reference examiner (Cramer V 0.85-0.93, P < 0.001). The classifications made using the new application and by observer 1 and observer 2 were also consistent although correlation was weaker (Cramer V 0.56-0.87, P < 0.001). For thinner LLPs, greatest correspondence with the reference was observed using the new software (96.6%), whereas the 2 observers showed better agreement when grading thicker patterns. Notwithstanding, agreement between the 2 observers and the reference examiner was good with at least 81% matched classifications. Best agreement (96.2%) was noted between the new method and observers 1 and 2 for recognizing meshwork patterns, whereas observers 1 and 2 showed greatest correspondence when classifying color fringe patterns. CONCLUSIONS: The new application can objectively categorize LLPs using the Tearscope-plus.

Clinical evaluation of the Canon TX-10 noncontact tonometer in healthy eyes.

PURPOSE: To assess the accuracy and repeatability of intraocular pressure (IOP) measurements obtained with the Canon TX-10 noncontact tonometer (NCT), using the Goldmann applanation tonometer (GAT) as the gold standard. METHODS: Seventy-three young healthy subjects were enrolled in the study. Central corneal thickness (CCT) was obtained with Orbscan Topography System and then repeated IOP measurements were obtained with both tonometers. Canon TX-10 NCT was performed first, followed by GAT. Another study sample of 32 young healthy subjects was used to test the repeatability with both tonometers. Two sets of data were acquired separated by 1 week. Plotting the difference between the methods against the mean was done to compare the tonometers and the intersession variability. The hypothesis of zero bias was examined by a paired t test. The 95% limits of agreement (LoA) were also calculated. RESULTS: No statistically significant differences were found when comparing the Canon TX-10 NCT with GAT, displaying close level of agreement with GAT as seen by the 95% LoA (-4.78 mmHg to +4.00 mmHg). CCT ranged from 419 microm to 585 microm and no relationship was found between CCT and IOP measurements. The coefficients of repeatability were 3.70 mmHg and 3.41 mmHg for GAT and TX-10 tonometers, respectively. CONCLUSIONS: There was a good agreement between Canon TX-10 and GAT in healthy subjects. From the results obtained here, CCT appears to not affect IOP measurements when CCT is thinner than 570 microm.

A comparison of two pachymetric systems: slit-scanning and ultrasonic.

PURPOSE: This article compares measurements of corneal thickness obtained with the Orbscan Topography System with those obtained with the ultrasonic pachymeter, when used to measure central-comeal thickness in normal subjects. METHODS: Slit-scan topography (Orbscan II system, version 3.0, Bausch & Lomb, Rochester, NY) and ultrasonic pachymetry (Ophthasonic A-Scan/ Pachometer III, Accutome, Malvern, PA) were used to measure central-corneal thickness in 92 right corneas of 92 healthy adult subjects. A correlation analysis was used to evaluate the relationship between both instruments. Measurements were also compared by plotting the difference between the methods against the average. The hypothesis of zero bias was examined by a paired t-test. The 95% limits of agreement also were calculated. RESULTS: The mean corneal thickness was 559.9 microm +/-3.73 SEM with the Orbscan system and 551.2 microm +/- 4.57 SEM with the ultrasonic pachymeter, values that were significantly different (paired t-test, P = 0.000). The coefficient of determination was 0.860 (y = -65.78 + 1.135 x; P = 0.000). The mean difference between the measurements from both devices was found to be significantly different from zero (mean: 8.74 microm; paired t-test; P = 0.000), with Orbscan being slightly higher than the Ophthasonic A-Scan. CONCLUSIONS: The Orbscan system measurements of central corneal thickness were greater than ultrasonic pachymeter measurements, a difference that was statistically significantly.