Home

Media centre

  1. 2016-01-08 Lietuvio vadovaujamos komandos inovacija padės kovoti su vėžiu (www.delfi.lt)
  2. 2015-12-22 Vėžio žymenų tyrimams - inovacija. įkvėpta gamtos (www.ifpa.lt)
  3. 2015-12-22 Vėžio žymenų tyrimams - inovacija, įkvėpta gamtos (www.laisvaslaikrastis.lt)
  4. 2015-12-22 Vėžio žymenų tyrimams - inovacija, įkvėpta gamtos (sc.bns.lt)
  5. 2015-12-22 Vėžio žymenų tyrimams - inovacija, įkvėpta gamtos (www.regionunaujienos.lt)
  6. 2015-12-22 Vėžio žymenų tyrimams - inovacija, įkvėpta gamtos (www.manosveikata.lt)
  7. 2015-12-21 Vėžio žymenų tyrimams pasitelktas "medaus korio" metodas (www.naujienos.vu.lt)
  8. 2015-12-21 Vėžio žymenų tyrimams - inovacija, įkvėpta gamtos (www.vpc.lt)
  9. 2015-12-17 Vėžio žymenų tyrimams - inovacija, įkvėpta gamtos (www.mf.vu.lt)
  10. 2015-12-17 Vėžio žymenų tyrimams - inovacija, įkvėpta gamtos (www.santaroszinios.lt)
  11. 2015-06-02 Kanų ir Vilniaus universitetai pasirašė bendradarbiavimo sutartį skaitmeninės patologijos srityje, Prof. Arvydas Laurinavičius ir Benoit Plancoulaine
  12. 2015-06 Unicaen & l’Universitete de Vilnius ont signe un accord de collaboration en pathologie numerique (2015-06 Kanų universiteto naujienaiškis Phenix)
  13. 2015-06-01 Kanų ir Vilniaus universitetai pasirašė bendradarbiavimo sutartį skaitmeninės patologijos srityje
  14. 2015-05-14 Krūties vėžys “kompiuterinės bites akimis” (Lietuvos sveikata, Nr. 20(116), 2015-05-14)
  15. 2013-06-21 Pradėtas vykdyti tarptautinis projektas
  16. 2013-06-21 Pradėtas vykdyti tarptautinis projektas (SMĮA naujienlaiškis)
  17. 2013-04-16 Gauta parama projektui "Išsamus biožymenų raiškos intranavikinio heterogeniškumo įvertinimas skaitmeninės imunohistochemijos vaizdo analizės būdu"

 

Deliverables

1. Computer software for image analysis
2. Data bases
3. Information system - Virtual collaborative space for digital image analysis
4. Methodologies for tissue heterogeneity image analysis

1. Computer software for image analysis

1.1. Software Tools for Tissue Microarray Image Analysis Data Transformation. This software tool integrates image and statistical analysis systems.Components of software: 1. Aperio Output Converter. This tool mainly handles multiple input files, removes the undesired columns, changes the column headers as specified, appends specimen mapping data and then merges all the files into a single MS Excel (SAS conformant) output file. 2. Image Files Renamer. This tool renames the “markup image file” names into some meaningful filenames; Output of the CONVERTER module becomes an essential reference for reading further details about the input image files. At the completion it adds another column in the Output for renamed files.

1.2. Toolkit to build stereology test grid fore ease the evaluation of histological markers: dHIC Wizard. The tedious work is required for the pathologists to estimate the proportions of the markers on histological sections (Ki67 Li, mitotic count and so on). The arrival of new technologies such as the digital imaging "large field" allows to change the conventional techniques used with the microscope. Thus, the virtual slides (WSI) allow more comfort for the observations of histological sections but increase the surface to be observed for counting the elements (mitosis, nuclei and so on). It is possible to use the sampling of the virtual slide by means of the stereology test grids for the unbiased evaluation in order to keep a reasonable time for the expert. The number of the grid patterns remains high for an easy job keeping a reasonable uncertainty. So, we proposed to start automatically the work of marking the elements of interests in the stereology test grid using the image processing tools.
The toolkit uses a stereology test grid (of frames or of crosses) and various images processing software that generate binary masks. The test grid is automatically built by the grid generator tool and the binary masks are produced from the results of the image processing. The last tool of the toolkit can now make the test grid containing the marked elements.

1.3. The Hex-B process: a proposed toolbox to analyze the tumor heterogeneity. Two steps are to do after scanning histological slides to give quantitative data from the cancerous tumor sections: the image processing and the image analysis. In our project, the image processing is performed by the chaining of two software: Genie and Nuclear of Aperio Company. The results will subsequently subject to statistical analysis.
In the toolbox of the Hex-B, we find several utilities: 1. The stereology tool which adjusts the ROI drawing on the markup image (the result image of the image processing by Genie and Nuclear) and that builds the hexagonal grid drawing (honeycomb) into the ROI drawing. 2. The mask tool which translates the builded mask by Genie and Nuclear towards a composite image. This image contains two binary layers, the first layer in green color shows the negative nuclear profiles and the second layer in red color shows the positive nuclear profiles. 3. The main tool of the Hex-B process which computes the count of the positive nuclear profiles and the count of the negative nuclear profiles in each hexagon. Then, the tool computes the percent of PositiveHex (Ki67 LI) in each hexagonal pattern.

1.4. Analysis using the Ki67 LI computed into hexagonal area. A main component of the researcher’s workflow is the data analysis. The data analysis process comes after the image digitization, the image processing and the image analysis. The obtained results can help the pathologist to confirm the prognosis and/or the diagnosis. This data analysis method is presented round the stereology grid of hexagons. This tool uses the computed data from the contiguous grid based on the analysis of the hexagon neighbors. Itl determines the occurrence matrix to compute the Haralick parameters and this tool is able give the neighbors of each hexagon always in a same order.

1.5. WSI Hexagonizer. The HexB process was developed to extract information for statistical analysis of histological (KI67) whole slide images (WSIs) processed in chain by Genie and Nuclear image analysis tools developed by Aperio. It was initially decided to extend the HexB process with statistical analysis of the image analysis tool Halo by Indica Labs, but as the study of tumor heterogeneity is supported by many commercial tools for automated image analysis of histological slides, it is important to facilitate subsequent statistical analysis of the results produced by any of these. Thus, while extending the initial prototype with support for statistical analysis of results produced by the nuclear classifier available in the Halo software package, the development of next generation analysis toolbox for tumor heterogeneity was extended to include a generic interface for both supporting multiple medical image analysis vendors and for development of methods for extraction of data for statistical analysis independent of vendor.

1.6. Colfit tool. Tissue staining always remains a very delicate operation in histology. The stain variations can be due to the tissue nature and the tissue thickness but also to the stain concentration, although the performances of the automated slide stainers are well improved. This technical problem traps often the image processing algorithms based on the color. The images from histological slide series must have just little color variations in order to use automated processes with a minimum of the manual settings. In this way, the tissue detection or the cell recognition may be improved a slide to an another. The software prototype "colfit" allows to stain the images considering normalized values from which are performed the correction of the histograms of the three colors, red, green and blue. Thus, the variation of the colors can be reduced for slide series.

2. Data bases

2.1. Reference image data base with established ground truth.
2.2. Image data base for image analysis training, validation, and research.

3. Information system - Virtual collaborative space for digital image analysis

4. Methodologies for tissue heterogeneity image analysis.

Methodology 1: Automated detection and evaluation of "hot spots" - Pareto hot spots Concept.
Methodology 2: Evaluation of dimensional differences indicators of texture and other marker expression.

Digital pathology opens new opportunities for artificial intelligence applications in disease diagnosis and treatment. Major benefits come with extraction and quantification of novel features of pathology, often not visible by usual microscopy examination. The field is rapidly developing in many areas of medicine.

We started digital pathology research in 2010 with image analysis-based quantification1, 2 and exploring the benefits of multivariate analytics of image analysis data to generate combined image-based biomarkers3, 4. Further, we performed accuracy and calibration experiments for image analysis-based quantification5, 6, 7, 8.

In 2015, we proposed a methodology based on hexagonal tiling of image analysis data to quantify intra-tumor heterogeneity of biomarker expression9; this allowed to generate rich data set to compute spatial indicators. We conceptualized this approach as “comprehensive immunohistochemistry”10. The heterogeneity indicators were demonstrated in later studies as independent prognostic factors, often exceeding the informative value of the average level of the biomarker expression11, 12.

In 2020, we published another hexagonal grid-based methodology to automatically detect tumor-host interface zone and compute immune cell density profile across the zone13; this actually measures the “willingness” of immune cells to enter the tumor (immunogradient) and provides independent prognostic value11, 13, 14. Compared to other methods proposed to measure immune response in the tumor microenvironment, the interface zone immunogradient provides quantitative directional assessment in the very frontline of tumor-host interaction. 

We demonstrated that intratumoral heterogeneity- and immunogradient-based computational biomarkers could be combined to predict patient survival without requirement for any conventional clinical or pathology data12. Also, computational augmentation of a single CD8 immunohistochemistry image data generated three independent prognostic features15. Tumor-associated fibrillary collagen architecture, retrieved from 1 mm breast cancer tissue microarray core by deep learning network and assessed by multiple morphometry metrics, could predict survival of the patients16. The methods were recently reviewed in the context of artificial intelligence applications for tumor pathology17, 18

 

Prof. Arvydas Laurinavičius with a team of researchers 

Log In or Register

Cookies make it easier for us to provide you with our services. With the usage of our services you permit us to use cookies. More information