June 2017
Volume 58, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2017
Image analysis of conjunctival staining with Lissamine green in dry eye syndrome
Author Affiliations & Notes
  • Emilie COURRIER
    Corneal Graft Biology, Engineering and Imaging Laboratory, Faculty of Medicine - Jean Monnet University, Saint-Etienne, France
  • Didier Renault
    Thea Laboratories, Clermont-Ferrand, France
    Corneal Graft Biology, Engineering and Imaging Laboratory, Faculty of Medicine - Jean Monnet University, Saint-Etienne, France
  • Caroline Urrea
    Department of Ophthalmology, University Hospital of Saint-Etienne, Saint-Etienne, France
  • Mathilde Kaspi
    Department of Ophthalmology, University Hospital of Saint-Etienne, Saint-Etienne, France
  • Elian Dib
    Corneal Graft Biology, Engineering and Imaging Laboratory, Faculty of Medicine - Jean Monnet University, Saint-Etienne, France
    Hubert Curien Laboratory (UMR 5516 CNRS), Jean Monnet University, Saint-Etienne, France
  • Corinne Fournier
    Hubert Curien Laboratory (UMR 5516 CNRS), Jean Monnet University, Saint-Etienne, France
  • Thierry Lépine
    Corneal Graft Biology, Engineering and Imaging Laboratory, Faculty of Medicine - Jean Monnet University, Saint-Etienne, France
    Hubert Curien Laboratory (UMR 5516 CNRS), Jean Monnet University, Saint-Etienne, France
  • Frederic Chiambaretta
    Department of Ophthalmology, University Hospital of Clermont-Ferrand, Clermont-Ferrand, France
  • Guillaume Hor
    Corneal Graft Biology, Engineering and Imaging Laboratory, Faculty of Medicine - Jean Monnet University, Saint-Etienne, France
  • Zhiguo HE
    Corneal Graft Biology, Engineering and Imaging Laboratory, Faculty of Medicine - Jean Monnet University, Saint-Etienne, France
  • Gilles Thuret
    Corneal Graft Biology, Engineering and Imaging Laboratory, Faculty of Medicine - Jean Monnet University, Saint-Etienne, France
    Department of Ophthalmology, University Hospital of Saint-Etienne, Saint-Etienne, France
  • Philippe GAIN
    Corneal Graft Biology, Engineering and Imaging Laboratory, Faculty of Medicine - Jean Monnet University, Saint-Etienne, France
    Department of Ophthalmology, University Hospital of Saint-Etienne, Saint-Etienne, France
  • Footnotes
    Commercial Relationships   Emilie COURRIER, None; Didier Renault, None; Caroline Urrea, None; Mathilde Kaspi, None; Elian Dib, None; Corinne Fournier, None; Thierry Lépine, None; Frederic Chiambaretta, None; Guillaume Hor, None; Zhiguo HE, None; Gilles Thuret, None; Philippe GAIN, None
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science June 2017, Vol.58, 2703. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Emilie COURRIER, Didier Renault, Caroline Urrea, Mathilde Kaspi, Elian Dib, Corinne Fournier, Thierry Lépine, Frederic Chiambaretta, Guillaume Hor, Zhiguo HE, Gilles Thuret, Philippe GAIN; Image analysis of conjunctival staining with Lissamine green in dry eye syndrome. Invest. Ophthalmol. Vis. Sci. 2017;58(8):2703.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose : Lissamine Green (LG) is often used in addition to fluorescein to assess the severity of conjunctival lesions during dry eye syndrome but its quantification remains manual. Aim: to describe an algorithm destined to the image analysis of LG stained conjunctival lesions.

Methods : Ten pictures of 6 patients suffering from dry eye and presenting visible LG positive conjunctival staining were selected among pre-existing data. Pictures had been taken during routine consultation, after instillation of LG, using 2 different slit-lamps with a white light source and a red filter transmitting over the wavelengths absorbed by LG. Conjunctival staining consequently appeared black against a red background. 16x magnification was used to image the entire conjunctival surface. With the ImageJ software, the red channel was extracted from the original image. Edge detection was performed using a Laplacian of Gaussian (LoG) filter. A manual threshold was determined on a subset of images in order to select only the stained areas. The same threshold remained constant thereafter. Finally, the detected areas were overlaid onto the original RGB image. The lesions were also manually outlined by two experts for comparison.

Results : In addition to enhancing the global contrast on the conjunctival surface, the red filter combined with a LoG filter allowed highlighting and localizing LG staining while avoiding the erroneous detection of blood vessels. The delineation obtained by the algorithm closely matched the actual contours of small lesions, and was therefore in good agreement with the experts. The large areas of confluent LG stained lesions were separated by the algorithm in numerous small dots, resulting in a large overestimation of the total number of dots. Nevertheless, these cases were already classified as Oxford grade V (>316 dots) and the grading was therefore not modified by the overestimation. Despite slight changes in picture quality, results were similar for both slit-lamps indicating that the algorithm was robust.

Conclusions : This new image analysis algorithm could increase the reliability (fidelity, accuracy, reproducibility) of the quantification of LG conjunctival staining, especially in clinical trials.

This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.

 

Delineation of LG conjunctival staining obtained through our approach thanks to a filtered image (A) and overlaid onto the RGB image (B). Scale bar: 1.25mm.

Delineation of LG conjunctival staining obtained through our approach thanks to a filtered image (A) and overlaid onto the RGB image (B). Scale bar: 1.25mm.

×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×