Automated video image morphometry of the corneal endothelium

Abstract

The central corneal endothelium of 13 eyes in 13 subjects was visualized with a non-contact specular microscope. This report describes the computer-assisted morphometric analysis of enhanced digitized images, using a direct input by means of a frame grabber. The output consisted of mean cell area, cell density, frequency distribution of the individual cell area, and cell polygonality. Results showed that the mean coefficient of variation of images analyzed three times consecutively was 0.95 percent. The cell analysis of three different images of the same subject was accurate, with a mean coefficient of variation of 4.2 percent.

Influence of Cataract Surgery on Corneal Endothelial Cell Density Estimation

“…12,17,23,24 Previous studies demonstrated a low mean coefficient of variation of 1% to 2.2%, 12,17 and several authors have pointed out that the analysis of 1 image per eye is sufficient in most clinical circumstances. 11,23,24 However, most sources of imprecision will increase as cell density decreases. 25 Several evaluations have involved cell counts within a defined area or some form of fixed frame, and evidence has been presented that cell counting techniques can differ in the estimation of ECD.…”

“…[6][7][8] Recent technological advances enabled to examine cell densities and other morphometric parameters directly from the original photograph or video image. [9][10][11][12][13] Thus, a rapid morphometric analysis was possible, although some manual involvement may still be required. The most popularly used counting methods are the fixed frame, the variable frame, and the center cell analysis.…”

Automatic estimation of endothelium cell density in donor corneas by means of Fourier analysis

“…Some of them are based on classic image processing techniques (LESTER et al, 1981;HIRST et al, 1984;HARTMANN and KODITZ, 1984;NISHI and HANASAKI, 1988;SIERTSEMA et al, 1993;CORKIDI et al, 1993); others employ more advanced techniques such as mathematical morphology (HASEGAWA et al, 1992), artificial neural networks (HASEGAWA et al, 1996), fuzzy logic (REINHARD et al, 1999), or some combination of them (FoRACCHIA and RUGGERI, 2000). Some algorithms were also implemented in commercial systems, provided either by microscope makers, such as Konan or Topcon, or by software providers, such as Bio-Optics (BENETZ et al, 1999), Rhine-Tec (REINHARD et al, 1999) or Nidek Technologies (FoRACCHIA and RUGGERI, 2000).…”

Utilização do programa Cells Analyser® no estudo comparativo entre imagens do endotélio de córneas obtidas por microscopia especular

“…A ampliação da imagem captada faz-se por um sistema óptico- RESUMO eletrônico para então ser direcionada a um software que realizará a análise quantitativa e qualitativa do endotélio corneal (4)(5) . A microscopia especular é um recurso importante para o estudo da morfologia endotelial afirmativa da Academia Americana de Oftalmologia (6) , corroborada por outros autores (1,(7)(8)(9)(10)(11) . O software Cells Analyser ® complementa a tecnologia dos microscópios especulares existentes, pela incorporação dos resultados destes equipamentos, obtendo o erro amostral e indicando o número de células que devem ser consideradas para tornar o resultado do exame endotelial representativo da realidade do mosaico endotelial como um todo, inferindolhes validade estatística e conseqüentemente médica, para melhor definirem diagnósticos e condutas clínicas e/ou cirúr-gicas, de modo a elevar a qualidade deste procedimento a um novo patamar de confiabilidade médica (12) .…”