Dataset Information

Machine learning helps improve diagnostic ability of subclinical keratoconus using Scheimpflug and OCT imaging modalities.

ABSTRACT: Purpose:To develop an automated classification system using a machine learning classifier to distinguish clinically unaffected eyes in patients with keratoconus from a normal control population based on a combination of Scheimpflug camera images and ultra-high-resolution optical coherence tomography (UHR-OCT) imaging data. Methods:A total of 121 eyes from 121 participants were classified by 2 cornea experts into 3 groups: normal (50 eyes), with keratoconus (38 eyes) or with subclinical keratoconus (33 eyes). All eyes were imaged with a Scheimpflug camera and UHR-OCT. Corneal morphological features were extracted from the imaging data. A neural network was used to train a model based on these features to distinguish the eyes with subclinical keratoconus from normal eyes. Fisher's score was used to rank the differentiable power of each feature. The receiver operating characteristic (ROC) curves were calculated to obtain the area under the ROC curves (AUCs). Results:The developed classification model used to combine all features from the Scheimpflug camera and UHR-OCT dramatically improved the differentiable power to discriminate between normal eyes and eyes with subclinical keratoconus (AUC?=?0.93). The variation in the thickness profile within each individual in the corneal epithelium extracted from UHR-OCT imaging ranked the highest in differentiating eyes with subclinical keratoconus from normal eyes. Conclusion:The automated classification system using machine learning based on the combination of Scheimpflug camera data and UHR-OCT imaging data showed excellent performance in discriminating eyes with subclinical keratoconus from normal eyes. The epithelial features extracted from the OCT images were the most valuable in the discrimination process. This classification system has the potential to improve the differentiable power of subclinical keratoconus and the efficiency of keratoconus screening.

SUBMITTER: Shi C

PROVIDER: S-EPMC7507244 | biostudies-literature | 2020

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Machine learning helps improve diagnostic ability of subclinical keratoconus using Scheimpflug and OCT imaging modalities.

Shi Ce C Wang Mengyi M Zhu Tiantian T Zhang Ying Y Ye Yufeng Y Jiang Jun J Chen Sisi S Lu Fan F Shen Meixiao M

Eye and vision (London, England) 20200910

<h4>Purpose</h4>To develop an automated classification system using a machine learning classifier to distinguish clinically unaffected eyes in patients with keratoconus from a normal control population based on a combination of Scheimpflug camera images and ultra-high-resolution optical coherence tomography (UHR-OCT) imaging data.<h4>Methods</h4>A total of 121 eyes from 121 participants were classified by 2 cornea experts into 3 groups: normal (50 eyes), with keratoconus (38 eyes) or with subcli ...[more]

PMID: 32974414

Similar Datasets

Project description:PURPOSE:To determine optimal objective, machine-derived variables and variable combinations from Scheimpflug and spectral-domain (SD) OCT imaging to distinguish the clinically unaffected eye in patients with asymmetric keratoconus (AKC) from a normal control population. DESIGN:Retrospective case-control study. PARTICIPANTS:Thirty clinically unaffected eyes with no physical findings on slit-lamp examination, no definitive abnormalities on corneal imaging, and corrected distance acuity of 20/20 or better from 30 patients with highly AKC eyes and 60 eyes from 60 normal control patients who had undergone uneventful LASIK with at least 2 years of stable follow-up (controls). METHODS:Scheimpflug and SD OCT imaging were obtained in all eyes, and receiver operating characteristic (ROC) curves were generated to determine area under the curve (AUC), sensitivity, and specificity for each machine-derived variable and variable combination. MAIN OUTCOME MEASURES:Distinguishing AKC eyes from controls as determined by AUC, sensitivity, and specificity. RESULTS:No individual machine-derived metric from Scheimpflug or SD OCT technology yielded an AUC higher than 0.75. Combining 5 Scheimpflug metrics (index height decentration [IHD], index vertical asymmetry [IVA], pachymetry apex, inferior-superior value, and Ambrosio's Relational Thickness Maximum [ARTmax]) yielded the best Scheimpflug results (AUC 0.86, sensitivity 83%, specificity 83%). Combining 11 SD OCT thickness metrics (minimum-median, temporal outer, superior nasal outer, minimum, epithelium minimum-maximum, epithelial standard deviation, superior inner, superior outer, superior temporal outer, superior nasal inner, central) yielded the best SD OCT results (AUC 0.96, sensitivity 89%, specificity 89%). Combining 13 total Scheimpflug/SD OCT metrics yielded the best results overall (AUC 1.0, sensitivity 100%, specificity 100%). The most impactful variables in combined models included epithelial thickness variability and total focal corneal thickness variability from SD OCT and anterior curvature and topometric indices from Scheimpflug technology. No posterior corneal metrics were impactful in modeling. CONCLUSIONS:Individual machine-derived metrics from Scheimpflug and SD OCT imaging poorly distinguished normal eyes from minimally affected eyes from patients with highly AKC. Combined SD OCT metrics performed better than combined Scheimpflug metrics. Combining anterior curvature and asymmetry indices from Scheimpflug with regional total thickness and epithelial thickness variability metrics from SD OCT clearly distinguished the 2 populations. Posterior corneal indices were not useful in distinguishing populations.

Project description:BackgroundKeratoconus is a disorder characterized by progressive thinning and distortion of the cornea. If detected at an early stage, corneal collagen cross-linking can prevent disease progression and further visual loss. Although advanced forms are easily detected, reliable identification of subclinical disease can be problematic. Several different machine learning algorithms have been used to improve the detection of subclinical keratoconus based on the analysis of multiple types of clinical measures, such as corneal imaging, aberrometry, or biomechanical measurements.ObjectiveThe aim of this study is to survey and critically evaluate the literature on the algorithmic detection of subclinical keratoconus and equivalent definitions.MethodsFor this systematic review, we performed a structured search of the following databases: MEDLINE, Embase, and Web of Science and Cochrane Library from January 1, 2010, to October 31, 2020. We included all full-text studies that have used algorithms for the detection of subclinical keratoconus and excluded studies that did not perform validation. This systematic review followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) recommendations.ResultsWe compared the measured parameters and the design of the machine learning algorithms reported in 26 papers that met the inclusion criteria. All salient information required for detailed comparison, including diagnostic criteria, demographic data, sample size, acquisition system, validation details, parameter inputs, machine learning algorithm, and key results are reported in this study.ConclusionsMachine learning has the potential to improve the detection of subclinical keratoconus or early keratoconus in routine ophthalmic practice. Currently, there is no consensus regarding the corneal parameters that should be included for assessment and the optimal design for the machine learning algorithm. We have identified avenues for further research to improve early detection and stratification of patients for early treatment to prevent disease progression.

Project description:PURPOSE:To compare the diagnostic ability of wide-field swept-source optical coherence tomography (SS-OCT) retinal nerve fiber layer (RNFL) maps with spectral-domain OCT (SD-OCT) maps for detection of preperimetric (PPG) and early glaucoma (EG). PATIENTS AND METHODS:One hundred and forty-six eyes, including 37 healthy eyes, 38 eyes with PPG, and 71 eyes with EG, were analyzed. The patients underwent both SD-OCT (Cirrus HD-OCT; Carl Zeiss Meditec, Dublin, CA, USA) and wide-field SS-OCT scanning (DRI-OCT-1 Atlantis; Topcon, Tokyo, Japan). By SD-OCT, circumpapillary RNFL and macular ganglion cell analyses were performed. SS-OCT provides a wide-field RNFL thickness map and a SuperPixel map, which are composed of an RNFL deviation map of the peripapillary area and a deviation map of the composition of the ganglion cell layer with the inner plexiform layer and RNFL [GC-IPL+RNFL] in the macular area. The ability to discriminate PPG and EG from healthy eyes was assessed according to sensitivity, specificity and area under the receiver operating characteristic curve for parameters and criteria provided by SD-OCT and wide-field SS-OCT scanning. RESULTS:The wide-field RNFL thickness map obtained by SS-OCT showed the highest sensitivity to PPG and EG (92.1 and 97.2%, respectively) as compared with the other, SD-OCT criteria. The wide-field RNFL thickness map showed PPG-diagnostic performance comparable to the SD-OCT RNFL thickness and GC-IPL deviation maps (p?=?0.453 and 0.180), and PPG-diagnostic performance superior to the SD-OCT RNFL deviation and GC-IPL thickness maps (p?=?0.003 and 0.039). In EG, the wide-field RNFL thickness and SuperPixel maps showed diagnostic performance comparable to the SD-OCT thickness and deviation maps (p?=?0.065 to 0.100), except for the GC-IPL thickness map (p?=?0.004). CONCLUSIONS:The wide-field SS-OCT RNFL thickness maps showed a diagnostic ability for distinguishing PPG and EG from healthy eyes that was similar to that of SD-OCT. In the clinical setting, these maps can be effective for detection of early-glaucomatous changes.

Project description:PurposeTo compare the predicted ablation depth (AD) with the postoperatively measured corneal ablation depth (postop-AD) at central, paracentral, and midperipheral locations using two rotating Scheimpflug analyzers and a Fourier-domain optical coherence tomographer in eyes that underwent femtosecond laser-assisted LASIK (FS-LASIK).MethodsThe values of corneal thickness were measured preoperatively and postoperatively at one and three months. The difference between preoperative and postoperative was defined as postop-AD. Measurements were performed at the corneal vertex and mid-peripheral area. The mid-peripheral corneal thickness was measured at the superior, inferior, nasal, and temporal locations at a distance of 1.0 or 2.5 mm from the corneal vertex. The predicted AD was calculated by ORK-CAM software (Schwind eye tech-solutions GmbH, Kleinostheim, Germany), and the difference between the predicted AD and postop-AD was defined as Δ-AD. Paired t-test analysis was employed to evaluate the differences, agreement was assessed by the Bland-Altman method.ResultsForty-two eyes of 42 patients were investigated. At one month, the predicted AD in the central and paracentral areas was underestimated by the Pentacam HR (Oculus, Wetzlar, Germany), Sirius (Costruzione Strumenti Oftalmici, Florence, Italy) and RTVue OCT (Optovue Inc., Freemont, CA, United States), whereas Δ-AD was negative as established by all devices and predominantly statistically significant. The Δ-AD values approximated zero at three months. The mean difference of Δ-AD at three months at the corneal vertex was 0.67 ± 9.39 mm, -7.92 ± 9.05 mm and -1.36 ± 8.31 mm, respectively. The mid-peripheral measurements had positive values at one month and even more highly positive at three months (with statistically significant differences in most of the cases). The agreement between the predicted and postop-AD was moderate with all devices, but slightly better with RTVue.ConclusionThe predicted AD seems to be underestimated in the central and paracentral corneal area and overestimated in the mid-periphery.Translational relevanceThe study could help to partly explain and prevent the refractive errors after FS-LASIK.

Project description:Custom high-resolution high-speed anterior segment spectral domain Optical Coherence Tomography (OCT) was used to characterize three-dimensionally (3-D) corneal topography in keratoconus before and after implantation of intracorneal ring segments (ICRS). Previously described acquisition protocols were followed to minimize the impact of the motions of the eye. The collected set of images was corrected from distortions: fan (scanning) and optical (refraction). Custom algorithms were developed for automatic detection and classification of volumes in the anterior segment of the eye, in particular for the detection and classification of the implanted ICRS. Surfaces were automatically detected for quantitative analysis of the corneal elevation maps (fitted by biconicoids and Zernike polynomials) and pachymetry. Automatic tools were developed for the estimation of the 3-D positioning of the ICRS. The pupil center reference was estimated from the segmented iris volume. The developed algorithms are illustrated in a keratoconic eye (grade III) pre- and 30 days post-operatively after implantation of two triangular-section, 0.3-mm thick Ferrara ring segments. Quantitative corneal topographies reveal that the ICRS produced a flattening of the anterior surface, a steepening of the posterior surface, meridional differences in the changes in curvature and asphericity, and increased symmetry of the anterior topography. Optical distortion correction through the ICRS (of a different refractive index from the cornea) allowed accurate pachymetric estimates, which showed increased thickness in the ectatic area as well as in peripheral corneal areas. Automatic tools allowed estimation of the depth of the implanted ICRS ring, as well as its rotation with respect to the pupil plane. Anterior segment sOCT provided with fan and optical distortion correction and analysis tools is an excellent instrument for evaluating and monitoring keratoconic eyes and for the quantification of the changes produced by ICRS treatment.

Dataset Information

Machine learning helps improve diagnostic ability of subclinical keratoconus using Scheimpflug and OCT imaging modalities.

Publications

Machine learning helps improve diagnostic ability of subclinical keratoconus using Scheimpflug and OCT imaging modalities.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets