Aim: To investigate the application of anterior and posterior corneal higher order aberrations (HOAs) in detecting keratoconus (KC) and suspect keratoconus (SKC). Method: This is a retrospective, case-control study which evaluated non-ectatic (normal) eyes, SKC eyes, and KC eyes. The Sirius Scheimpfug (CSO, Italy) analyzer was used to measure HOAs of the anterior and posterior corneal surfaces. Sensitivity, specificity and area under receiver operating characteristic curve (AUC) were calculated. Results: Two-hundred and twenty eyes were included in the analysis (normal n = 108, SKC n= 42, KC n= 70). Receiver operating characteristic (ROC) curve analysis revealed a high predictive ability for anterior corneal HOAs parameters: Root mean square (RMS) total corneal HOAs, RMS trefoil and RMS coma to detect keratoconus (AUC > 0.9 for all). RMS Coma (3, ±1) derived from the anterior corneal surface was the parameter with the highest ability to discriminate between suspect keratoconus and normal eyes (AUC = 0.922; cutoff > 0.2). All posterior corneal HOAs parameters were insufficient in discriminating between SKC and normal eyes (AUC < 0.8 for all). In contrast, their ability to detect KC was excellent with AUC of > 0.9 for all except RMS spherical aberrations (AUC = 0.846). Conclusion: Anterior and posterior corneal higher order aberrations can differentiate between keratoconus and normal eyes, with a high level of certainty. In suspect keratoconus disease however, only anterior corneal HOAs, and in particular coma-like aberrations, are of value. Corneal aberrometry may be of value in screening for keratoconus in populations with a high prevalence of the disease.

Aim: This study measures and evaluates the socioeconomic burden of people living with keratoconus in Saudi Arabia. Methods: The study employed a cross-sectional design, a Keratoconus Economic Burden Questionnaire, and a convenient sample of 89 keratoconus patients (58.4% male) drawn from multiple regions in Saudi Arabia. It was conducted using online surveys and the data was analysed using appropriate quantitative techniques. Results: The mean age and annual income of participants were 33.24 years and Saudi Riyal (SAR) 33,505.6180 (SD=62,215.29), respectively, with only 37% being employed for wages. Up to 94.4% needed glasses or contact lenses at least once a week and 73.0% received care from optometrists. The condition forced 45.9% of the respondents to change careers or leisure activities, with a further 51.3% having to take time off work. The mean annual out-of-pocket expenses for buying and maintaining glasses or contact lenses as well as traveling and accommodation for keratoconus-related treatment were SAR 8,673.19 (SD=11,307.73), with 48.32 incurring upwards of SAR 12,000 over the period. The treatment costs increased with disease duration, r(89) = .216, p < .05. Regression results show that the existence of comorbid eye disease, changing glasses at least once a year, and wearing either glasses or contact lenses at least once a week individually have statistically significant, negative effects on the total annual keratoconus treatment costs, while disease duration, utilisation of optometrists, and taking time off had a statistically significant increase on the total cost (p<.05). Conclusion: With a prevalence rate of 1 in 375, progressive debilitation, and the lifetime nature of the disease, keratoconus is a critical public health concern in Saudi Arabia. The resulting visual impairment and discomfort as well as both direct and indirect economic burdens have considerable impacts on the patient's quality of life.

Nowadays smartphone utilization for disease diagnosis and remote health care applications has become promising due to their ubiquity. Here, a novel convolutional neural network method for detecting keratoconus that is wholly implemented on a smartphone is proposed. The proposed method provides accurate detection of over 72.9% for all stages of keratoconus. Preliminary results indicate 90%, 83%, 64% and 52% detection rate for severe, advanced, moderate and mild stages of disease, respectively.

Machine learning (ML) has a large capacity to learn and analyze a large volume of data. This study aimed to train different algorithms to discriminate between healthy and pathologic corneal images by evaluating digitally processed spectral-domain optical coherence tomography (SD-OCT) corneal images. A set of 22 SD-OCT images belonging to a random set of corneal pathologies was compared to 71 healthy corneas (control group). A binary classification method was applied; three approaches of ML were used. Once all images were analyzed, representative areas from every digital image were also processed and analyzed for a statistical feature comparison between healthy and pathologic corneas. The best performance was obtained from transfer learning - support vector machine (TL-SVM) (AUROC = 0.94, SPE 88%, SEN 100%) and transfer learning – random forest (TL- RF) method (AUROC = 0.92, SPE 84%, SEN 100%), followed by convolutional neural network (CNN) (AUROC = 0.84, SPE 77%, SEN 91%) and random forest (AUROC = 0.77, SPE 60%, SEN 95%). The highest diagnostic accuracy in classifying corneal images was achieved with the TL-SVM and the TL-RF models. In image classification, CNN was a strong predictor. This pilot experimental study developed a systematic mechanized system to discern pathologic from healthy corneas.