Purpose: This study aimed to employ the incremental digital image correlation (DIC) method to obtain displacement and strain field data of the cornea from Corvis ST (CVS) sequences and access the performance of embedding these biomechanical data with machine learning models to distinguish forme fruste keratoconus (FFKC) from normal corneas. Methods: 100 subjects were categorized into normal (N = 50) and FFKC (N = 50) groups. Image sequences depicting the horizontal cross-section of the human cornea under air puff were captured using the Corvis ST tonometer. The high-speed evolution of full-field corneal displacement, strain, velocity, and strain rate was reconstructed utilizing the incremental DIC approach. Maximum (max-) and average (ave-) values of full-field displacement V, shear strain γxy, velocity VR, and shear strain rate γxyR were determined over time, generating eight evolution curves denoting max-V, max-γxy, max-VR, max-γxyR, ave-V, ave-γxy, ave-VR, and ave-γxyR, respectively. These evolution data were inputted into two machine learning (ML) models, specifically Naïve Bayes (NB) and Random Forest (RF) models, which were subsequently employed to construct a voting classifier. The performance of the models in diagnosing FFKC from normal corneas was compared to existing CVS parameters. Results: The Normal group and the FFKC group each included 50 eyes. The FFKC group did not differ from healthy controls for age (p = 0.26) and gender (p = 0.36) at baseline, but they had significantly lower bIOP (p < 0.001) and thinner central cornea thickness (CCT) (p < 0.001). The results demonstrated that the proposed voting ensemble model yielded the highest performance with an AUC of 1.00, followed by the RF model with an AUC of 0.99. Radius and A2 Time emerged as the best-performing CVS parameters with AUC values of 0.948 and 0.938, respectively. Nonetheless, no existing Corvis ST parameters outperformed the ML models. A progressive enhancement in performance of the ML models was observed with incremental time points during the corneal deformation. Conclusion: This study represents the first instance where displacement and strain data following incremental DIC analysis of Corvis ST images were integrated with machine learning models to effectively differentiate FFKC corneas from normal ones, achieving superior accuracy compared to existing CVS parameters. Considering biomechanical responses of the inner cornea and their temporal pattern changes may significantly improve the early detection of keratoconus.

Keratoconus (KC), a leading cause of vision impairment, has an unclear aetiology. This study used Mendelian randomization (MR) to explore the causal links between various factors (smoking, asthma, Down syndrome, inflammatory bowel disease, atopic dermatitis, and serum 25-hydroxyvitamin D levels) and KC. A two-sample MR design, grounded in genome-wide association study (GWAS) summary statistics, was adopted using data from FinnGen, UK Biobank, and other GWAS-related articles. The inverse-variance weighted (IVW) method was employed, complemented by the Wald ratio method for factors with only one single-nucleotide polymorphism (SNP). Sensitivity and stability were assessed through Cochrane’s Q test, the MR-Egger intercept test, MR-PRESSO outlier test, and the leave-one-out analysis. The IVW results for the ORA (Ocular Response Analyzer) biomechanical parameters indicated significant associations between tobacco smoking (CH: p < 0.001; CRF: p = 0.009) and inflammatory bowel disease (CH: p = 0.032; CRF: p = 0.001) and corneal biomechanics. The Wald ratio method showed tobacco smoking was associated with a lower risk of KC (p = 0.024). Conversely, asthma (p = 0.009), atopic dermatitis (p = 0.012), inflammatory bowel disease (p = 0.017), and serum 25-hydroxyvitamin D levels (p = 0.039) were associated with a higher risk of KC by IVW, and the same applied to Down syndrome (p = 0.004) using the Wald ratio. These results underscore the role of corneal biomechanics as potential mediators in KC risk, warranting further investigation using Corvis ST and Brillouin microscopy. The findings emphasise the importance of timely screening for specific populations in KC prevention and management.

Purpose: The aim of this study is to develop an index for distinguishing between very asymmetric ectasia with normal topography (VAE-NT) eyes and normal eyes, with good performance in validity, reliability, and predictive values.Methods: In the training dataset, this single-center retrospective study involved 102 healthy eyes and 97 VAE-NT eyes. After propensity score matching (PSM), data from 53 healthy eyes and 53 VAE-NT eyes, including demographic and Corvis ST examination results, were collected. The area under the receiver operating characteristic curve (AUC), sensitivity, specificity, intraclass correlation coefficient (ICC), and positive and negative likelihood ratios were calculated for the dynamic corneal response (DCR) parameters of Corvis ST. The Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) model was used to objectively and comprehensively evaluate the Corvis ST DCRs, and logistic regression was used to determine the optimal combination of parameters that can accurately separate VAE-NT from normal corneas. In the validation dataset, 44 VAE-NT eyes and 49 normal eyes were involved. The validity, reliability, and predictive value of the index were further assessed using the validation dataset. The VAE-NT index was compared with the tomographic and biomechanical index (TBI) in both the training and validation datasets.Results: In the training dataset, the optimal parameter combination forming the VAE-NT index included the following DCRs: SP A1, SP HC, A1 Time, DA Ratio Max (2 mm), DA Ratio Max (1 mm), Integrated Radius, and stress–strain index version 2 (SSI2). The receiver operating characteristic (ROC) curve analysis showed an AUC value of 0.971, with a cut-off value of 0.425, an accuracy of 95.283%, a specificity of 94.340%, and a sensitivity of 96.230%. In the validation dataset, the AUC value of the VAE-NT index was 0.980. The sensitivity and specificity of the VAE-NT index were 93.180% and 95.920%, respectively. The positive and negative likelihood ratios of the VAE-NT index were 22.830 and 0.071, respectively. The ICC of the VAE-NT index was 0.835, and the accuracy was 94.624%. The VAE-NT index outperformed TBI in both the training and validation datasets.Conclusion: The VAE-NT index was developed, exhibiting high sensitivity, specificity, and AUC, along with favorable likelihood ratios and repeatability, suggesting that the VAE-NT index is a robust and reliable tool for distinguishing VAE-NT eyes from normal eyes. Further validation in broader populations and over longer follow-up periods is needed to support clinical translation.