Keratoconus is a chronic-degenerative disease which results in progressive corneal thinning and steeping leading to irregular astigmatism and decreased visual acuity that in severe cases may cause debilitating visual impairment. In recent years, Machine Learning methods, especially Convolutional Neural Networks (CNN), have been applied to classify images according to either presence or absence of the disease, based on different corneal maps. This study aims to develop a novel CNN architecture to classify axial curvature maps of the anterior corneal surface in five different grades of disease (i: normal eye; ii: suspect eye; iii: subclinical keratoconus; iv: keratoconus; and v: severe keratoconus). The dataset comprises 3, 832 axial curvature maps represented on relative scale and labeled by ophthalmologists. The images were splitted into three distinct subsets: training (2, 297 images ≈ 60%), validation (771 images ≈ 20%), and test (764 images ≈ 20%) sets. The model achieved an overall accuracy of 78.53%, a macro-average sensitivity of 74.53% (87.50% for normal eyes, 46.56% for suspect eyes, 65.41% for subclinical keratoconus, 93.42% for keratoconus, and 79.25% for severe keratoconus) and a macro-average specificity of 94.42% (92.14% for normal eyes, 95.30% for suspect eyes, 93.82% for subclinical keratoconus, 91.24% for keratoconus, and 99.58% for severe keratoconus). Additionally, the model achieved AUC scores of 0.97, 0.92, 0.90, 0.98, and 0.94 for normal eye, suspect eye, subclinical keratoconus, keratoconus, and severe keratoconus, respectively. The results suggest that the CNN exhibited notable proficiency in distinguishing between normal eyes and various stages of keratoconus, offering potential for enhanced diagnostic accuracy in ocular health assessment.
Keratoconus is a chronic-degenerative disease which results in progressive corneal thinning and steepening leading to irregular astigmatism and decreased visual acuity that in severe cases may cause debilitating visual impairment. In recent years, different Machine Learning methods have been applied to distinguish either normal and keratoconic eyes. These methods utilize both corneal curvature maps and their corresponding numeric indices to perform the classification. The main objective of this study is to evaluate the performance of features extracted with Histograms of Oriented Gradients (HOG) and with Convolutional Neural Networks (CNN) in the classification of normal and keratoconic eyes, using axial map of the anterior corneal surface. Two distinct models were trained using the same Multilayer Perceptron (MLP) architecture: one of them using the HOG features as input, and the other with the CNN features. The Topographic Keratoconus Classification index (TKC) provided by Pentacam™ was used as a label and the KC2-labeled maps were defined as keratoconus. Each model was trained using 3,000 images of normal and 3,000 keratoconic eyes, and then validated and tested on 1,000 images of each label. The model trained with HOG features exhibited a sensitivity of 99.1% and specificity of 98.7%, with an Area Under the Curve (AUC) of 0.999143. The model trained with CNN features showed both sensitivity and specificity of 99.5%, and AUC = 0.999778. The results suggest that the performance of the classifier is similar for both types of features.
Introduction: To date, it has never been demonstrated the propagation sound speed in human corneas and lens in vivo. With the advent of Optical Coherence Tomography (OCT), it became possible to determine the dimensions of the ocular tissues without the interference of sound propagation speed and to use this information to define the real propagation sound speed for each patient and individualized structure. Aim: To determine the sound propagation speed in the cornea and lens from patients that theoretically exhibits differences in tissue elasticity (normal corneas and keratoconus, corneas of young and elderly patients, in addition to clear crystalline lens from young and elderly patients with cataract). Then, relate the determined velocity in each group with the expected tissue elasticity of the cornea and lens. Methods: We studied 100 eyes from 50 patients: 50 with keratoconus and no cataract and 50 with cataract and no corneal changes. All patients measured corneal and lens thickness by ultrasound methods (Ultrasonic Biomicroscopy - UBM and Ultrasonic Pachymetry - USP) and by OCT (RTVue®, Lenstar® and Visante®), then were divided into 2 groups: Group 1 (Cornea) analyzed the central corneal thickness (UBM, USP, RTVue®, Visante®, Lenstar®); Group 2 (Lens) analyzed the axial thickness of the lens (UBM and Lenstar®). Based on standard ultrasonic speed from USP (1640 m/s) and UBM (1548 m/s), we calculated the real sound propagation speed in each tissue. Results: Based on USP, the corneal sound speed on control group (1616 m/s) was faster than on keratoconus group (1547 m/s) (P < 0.0001). Based on UBM, the lens sound speed on cataract group (1664 m/s) was faster that on control group (1605 m/s) (P < 0.0001). Discussion: It is known that sound propagates faster in materials with lower elasticity. We found that the sound speed on keratoconic corneas (high elasticity) was slower and on cataract lens (lower elasticity) was faster than normal corneas and lens in vivo.
Elastography is the mapping of tissues and cells by their respective mechanical properties, such as elasticity and viscosity. Our interest primarily lies in the human eye. Combining Scanning Laser Doppler Vibrometry (SLDV) with geometrically focused mechanical vibratory excitations of the cornea, it is possible to reconstruct these mechanical properties of the cornea. Experiments were conducted on phantom corneas as well as excised donor human corneas to test feasibility and derive a method of modeling. Finite element analysis was used to recreate the phantom studies and corroborate with the experimental data. Results are in close agreement. To further expand the study, lamb eyes were used in MR Elastography studies. 3D wave reconstruction was created and elastography maps were obtained. With MR Elastography, it would be possible to noninvasively measure mechanical properties of anatomical features not visible to SLDV, such as the lens and retina. Future plans include creating a more robust finite element model, improving the SLDV method for in-vivo application, and continuing experiments with MR Elastography.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.