Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Comparative Study
. 2013 Jun 6;8:92.
doi: 10.1186/1746-1596-8-92.

Analyzing Huge Pathology Images With Open Source Software

Affiliations
Free PMC article
Comparative Study

Analyzing Huge Pathology Images With Open Source Software

Christophe Deroulers et al. Diagn Pathol. .
Free PMC article

Abstract

Background: Digital pathology images are increasingly used both for diagnosis and research, because slide scanners are nowadays broadly available and because the quantitative study of these images yields new insights in systems biology. However, such virtual slides build up a technical challenge since the images occupy often several gigabytes and cannot be fully opened in a computer's memory. Moreover, there is no standard format. Therefore, most common open source tools such as ImageJ fail at treating them, and the others require expensive hardware while still being prohibitively slow.

Results: We have developed several cross-platform open source software tools to overcome these limitations. The NDPITools provide a way to transform microscopy images initially in the loosely supported NDPI format into one or several standard TIFF files, and to create mosaics (division of huge images into small ones, with or without overlap) in various TIFF and JPEG formats. They can be driven through ImageJ plugins. The LargeTIFFTools achieve similar functionality for huge TIFF images which do not fit into RAM. We test the performance of these tools on several digital slides and compare them, when applicable, to standard software. A statistical study of the cells in a tissue sample from an oligodendroglioma was performed on an average laptop computer to demonstrate the efficiency of the tools.

Conclusions: Our open source software enables dealing with huge images with standard software on average computers. They are cross-platform, independent of proprietary libraries and very modular, allowing them to be used in other open source projects. They have excellent performance in terms of execution speed and RAM requirements. They open promising perspectives both to the clinician who wants to study a single slide and to the research team or data centre who do image analysis of many slides on a computer cluster.

Virtual slides: The virtual slide(s) for this article can be found here:http://www.diagnosticpathology.diagnomx.eu/vs/5955513929846272.

Figures

Figure 1
Figure 1
A sample slide. (a): macroscopic view of the whole slide (the black rectangle on the left is 1x2 cm). (b,c): Influence of the magnification on the quality of results. (b): a portion of the slide scanned at magnification level 10x. The white contours show the result of an automatic detection of the dark cell nuclei with the ImageJ software. A significant fraction of the cell nuclei is missed and the contours are rather pixelated. (c): the same portion of the slide scanned at magnification 40x. The white contours show the result of the same automatic detection. Almost all cell nuclei are detected and the shapes of the contours are much more precise. Scale bar: 4 μm.
Figure 2
Figure 2
A typical session using ImageJ and the NDPITools plugins. A NDPI file has been opened with the NDPITools plugins and it is displayed as a preview image (image at largest resolution which still fits into the computer’s screen) — top window. A rectangular region has been selected and extracted as a TIFF image, then opened — bottom window.
Figure 3
Figure 3
Preview image of a NDPI file with several focalization levels in ImageJ. The NDPI file 08.ndpi contains images at 5 different focalization levels. Therefore, its preview image is displayed as a stack of 5 images.
Figure 4
Figure 4
Dialog box for customized extraction in ImageJ from an NDPI file with production of a mosaic. The dialog box shows some options which can be customized while producing a mosaic from a rectangular selection of a NDPI file preview image (here, using the file previewed in Figure 3).
Figure 5
Figure 5
The positions of the 154240 identified nuclei were obtained from the analysis with ImageJ of the digital slide on a laptop computer. Since the slide was too large to fit into the computer’s memory, it was turned into a mosaic of 16 pieces with overlap of 60 pixels, and each piece underwent automated analysis independently. Then the results were aggregated. The graph shows the probability density function of the distance of a cell nucleus to its nearest neighbor in the whole sample.

Similar articles

See all similar articles

Cited by 29 articles

See all "Cited by" articles

References

    1. Diamond J, McCleary D. In: Advanced Techniques in Diagnostic Cellular Pathology. Hannon-Fletcher M, Maxwell P, editor. Chichester UK: John Wiley & Sons, Ltd;; 2009. Virtual microscopy.
    1. Ameisen D, Yunès JB, Deroulers C, Perrier V, Bouhidel F, Battistella M, Legrès L, Janin A, Bertheau P. Stack or Trash? Fast quality assessment of virtual slides. Diagn Pathol. 2013. in press.
    1. García Rojo M, Castro AM, Gonçalves L. COST action “EuroTelepath”: digital pathology integration in electronic health record, including primary care centres. Diagn Pathol. 2011;6(Suppl 1):S6. doi: 10.1186/1746-1596-6-S1-S6. - DOI - PMC - PubMed
    1. Ameisen D. Intégration des lames virtuelles dans le dossier patient électronique. PhD thesis. 2013. Univ Paris Diderot-Paris 7.
    1. Collan Y, Torkkeli T, Personen E, Jantunen E, Kosma VM. Application of morphometry in tumor pathology. Anal Quant Cytol Histol. 1987;9(2):79–88. - PubMed

Publication types

LinkOut - more resources

Feedback