Penetrating keratoplasty (PK) ranks among the oldest and most common kinds of human tissue transplantation. Based on the hypothesis that reported indications for PK significantly vary between global regions and over time, the present systematic review aimed to provide a thorough overview of global PK indications as reported in peer-reviewed manuscripts. A literature search of PubMed and MEDLINE was conducted to retrieve articles published from January 1980 to May 2014. Indications for PK within 7 global regions were compared using a modified classification system for PK indications and analyzed via multivariate regression. A total of 141 publications from 37 countries were included, recording 180 865 PK cases. Postcataract surgery edema was the predominant indication in North America (28.0%) and ranked second in Europe (20.6%), Australia (21.1%), the Middle East (13.6%), Asia (15.5%), and South America (18.6%). Keratoconus was the leading indication in Europe (24.2%), Australia (33.2%), the Middle East (32.8%), Africa (32.4%), and South America (22.8%). It ranked third in North America (14.2%). Keratitis was the primary indication in Asia (32.3%). Fuchs endothelial corneal dystrophy was the fourth most common indication in North America (12.9%) and Europe (10.2%) and fifth in South America (3.8%). Multivariate analysis supported these results and revealed individual regional changes over time. Systematic analysis reveals characteristic chronological and regional differences in reported global PK indications. Leading reported indications for PK between 1980 and 2014 were keratoconus (Europe, Australia, the Middle East, Africa, and South America), pseudophakic bullous keratopathy/aphakic bullous keratopathy (North America), and keratitis (Asia).

Combining multiple classifiers, known as ensemble methods, can give substantial improvement in prediction performance of learning algorithms especially in the presence of non-informative features in the data sets. We propose an ensemble of subset of kNN classifiers, ESkNN, for classification task in two steps. Firstly, we choose classifiers based upon their individual performance using the out-of-sample accuracy. The selected classifiers are then combined sequentially starting from the best model and assessed for collective performance on a validation data set. We use bench mark data sets with their original and some added non-informative features for the evaluation of our method. The results are compared with usual kNN, bagged kNN, random kNN, multiple feature subset method, random forest and support vector machines. Our experimental comparisons on benchmark classification problems and simulated data sets reveal that the proposed ensemble gives better classification performance than the usual kNN and its ensembles, and performs comparable to random forest and support vector machines.

The feasibility of shifted-excitation Raman difference spectroscopy (SERDS) as a label-free and non-invasive technique for an objective diagnosis of oral cancer (OSCC) was investigated by analyzing 12 ex vivo OSCC samples. 72 mean SERDS spectra from each three physiological tissue points and pathological lesions were correlated with the histo-pathological diagnosis. Principal component analysis (PCA) and linear discriminant analysis (LDA) showed excellent results with an area under the curve of 94.5% and a classification error of 9.7% (sensitivity: 86.1%; specificity: 94.4%). The SERDS Raman spectra of malignant and benignant tissues were discriminable with respect to the spectral features of proteins, lipids and nucleic acids. The presented method is capable of a highly accurate identification of OSCC. These findings suggest a high validity and reproducibility of SERDS combined with PCA and LDA analysis regarding oral cancer tissue.

Purpose: The purpose of this study was to assess the diagnostic power of the novel two-dimensional parameter Bruch's membrane opening minimal rim area (BMO-MRA) in spectral-domain optical coherence tomography (SD-OCT) for detection of glaucoma compared to minimal rim width (BMO-MRW) and retinal nerve fiber layer (RNFL) thickness in large and small optic discs. Methods: In this case-control, cross-sectional study, 207 eyes of 207 participants, including 89 controls and 97 patients with glaucoma and 21 with ocular hypertension (OHT), with a disc size 2.43 mm² underwent SD-OCT, confocal laser scanning tomography (CSLT), visual field testing, and clinical examination. Bruch's membrane opening-MRA BMO-MRW, RNFL thickness of SD-OCT and disc margin rim area (DM-RA) of CSLT were evaluated and analyzed for diagnostic power to detect glaucoma. Results: In healthy eyes with macrodiscs, mean BMO-MRW of 243.14 ± 43.12 μm was significantly smaller than BMO-MRW in microdiscs (338.97 ± 69.39; P < 0.001). Bruch's membrane opening-MRA was comparable between disc size groups with 1.22 ± 0.25 mm² for macrodiscs and 1.26 ± 0.27 mm² for microdiscs (P = 0.51), as was RNFL thickness (82.69 ± 15.76 μm versus 78.53 ± 11.01 μm, respectively; P = 0.28). Perimetric mean deviation was −8.7 ± 6.3 dB in glaucoma and −0.6 ± 1.60 dB in OHT patients. Correlation of BMO-MRA and visual field function was rho (ρ) = 0.70 (P < 0.001). Diagnostic power to differentiate glaucoma patients was highest for BMO-MRA. Partial area under the curve (pAUC) for BMO-MRA was 0.14 for specificity between 0.8 and 1.0, exceeding pAUCs of BMO-MRW (P < 0.001), RNFL thickness (P = 0.03), and DM-RA (P = 0.01). Conclusions: Bruch's membrane opening-based minimum rim area measurements offer advantages compared to one-dimensional parameters assessing neuroretinal rim by SD-OCT. In nonglaucomatous eyes, BMO-MRA values seem comparable for the full range of disc sizes. Bruch's membrane opening-MRA surpasses other parameters in diagnostic power for glaucoma.