Computational pathology


TOWARDS MULTIMODAL COMPUTATIONAL PATHOLOGY MODELS

Our research focuses on developing mechanistic computational models by integrating hexagonal-grid analytics and deep learning-based feature extraction from microscopy images. In parallel, we are investigating label-free imaging techniques to generate novel image-based biomarkers for predictive pathology models. Specifically, we are currently conducting a study on polychromatic polarization microscopy funded by the Lithuanian Research Council. Our overarching vision is to create multimodal computational pathology models based on the synergy between microscopy imaging and machine learning. These models will aim to generate parametrized disease phenotypes for medical learning systems.




OUR STORY

Our computational pathology research began in 2010, initially focusing on image analysis-based quantification1, 2. We then explored the utility of multivariate analytics applied to image analysis data to develop combined image-based biomarkers3, 4. Furthermore, we conducted experiments to assess the accuracy and calibration of image analysis-based quantification methods5-8.

 In 2015, we introduced a method using hexagonal tiling of image analysis data to quantify intra-tumoral heterogeneity of biomarker expression9. This approach generated rich datasets, enabling the computation of spatial indicators. We conceptualized this as “comprehensive immunohistochemistry10. Subsequent studies demonstrated that these heterogeneity indicators were independent prognostic factors, often providing more informative value than the average biomarker expression level11-13.

Building on this, in 2020, we published another hexagonal grid-based method for the automated detection of the tumor-host interface zone. This allowed us to compute immune cell density profiles across this critical zone 14. This method effectively measures the "willingness" of immune cells to infiltrate the tumor (immunogradient) and has shown independent prognostic value12, 14-16. Compared to other approaches for assessing immune response within the tumor microenvironment, the Interface Zone Immunogradient (patent No. US 12,131,481 B2) offers a quantitative and directional evaluation at the immediate frontline of tumor-host interaction.

We have demonstrated that intratumoral heterogeneity- and immunogradient-based computational biomarkers synergistically predict patient survival, even without the need for conventional clinical or pathology data12, 17. Additionally, computational augmentation of spatially resolved data from a single CD8 immunohistochemistry image yielded three independent prognostic features18. We also showed that the architecture of tumor-associated fibrillary collagen, extracted from histology images using a deep learning network and assessed through multiple morphometry metrics, could predict patient survival19, 20. Our methods have been reviewed within the broader context of artificial intelligence applications in tumor pathology21, 22.

Watch the animation Immunogradient - Computational Biomarkers of Anti-Tumour Responses

Guided tour @NPC „What we do in computational pathology“ 











Prof. Arvydas Laurinavičius with a team of researchers digital@vpc.lt


    [1] Brazdziute E, Laurinavicius A: Digital pathology evaluation of complement C4d component deposition in the kidney allograft biopsies is a useful tool to improve reproducibility of the scoring. Diagnostic Pathology 2011.

     http://www.diagnosticpathology.org/content/pdf/1746-1596-6-S1-S5.pdf

     

    [2] Laurinaviciene A, Dasevicius D, Ostapenko V, Jarmalaite S, Lazutka J, Laurinavicius A: Membrane connectivity estimated by digital image analysis of HER2 immunohistochemistry is concordant with visual scoring and fluorescence in situ hybridization results: algorithm evaluation on breast cancer tissue microarrays. Diagnostic Pathology 2011.

     http://www.diagnosticpathology.org/content/pdf/1746-1596-6-87.pdf

     

    [3] Laurinavicius A, Laurinaviciene A, Ostapenko V, Dasevicius D, Jarmalaite S, Lazutka J: Immunohistochemistry profiles of breast ductal carcinoma: factor analysis of digital image analysis data. Diagnostic Pathology 2012.

     http://www.diagnosticpathology.org/content/pdf/1746-1596-7-27.pdf

     

    [4] Laurinavicius A, Green AR, Laurinaviciene A, Smailyte G, Ostapenko V, Meskauskas R, Ellis IO: Ki67/SATB1 ratio is an independent prognostic factor of overall survival in patients with early hormone receptor-positive invasive ductal breast carcinoma. Oncotarget 2015.

     http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4747395/pdf/oncotarget-06-41134.pdf

     

    [5] Laurinavicius A, Plancoulaine B, Laurinaviciene A, Herlin P, Meskauskas R, Baltrusaityte I, Besusparis J, Dasevicius D, Elie N, Iqbal Y, Bor C, Ellis IO: A methodology to ensure and improve accuracy of Ki67 labelling index estimation by automated digital image analysis in breast cancer tissue. Breast Cancer Research 2014.

     http://www.breast-cancer-research.com/content/pdf/bcr3639.pdf

     

    [6] Laurinaviciene A, Plancoulaine B, Baltrusaityte I, Meskauskas R, Besusparis J, Lesciute-Krilaviciene D, Raudeliunas D, Iqbal Y, Herlin P, Laurinavicius A: Digital immunohistochemistry platform for the staining variation monitoring based on integration of image and statistical analyses with laboratory information system. Diagn Pathol 2014.

     http://www.ncbi.nlm.nih.gov/pubmed/25565007

    http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4305968/pdf/1746-1596-9-S1-S10.pdf

     

    [7] Plancoulaine B, Laurinaviciene A, Meskauskas R, Baltrusaityte I, Besusparis J, Herlin P, Laurinavicius A: Digital immunohistochemistry wizard: image analysis-assisted stereology tool to produce reference data set for calibration and quality control. Diagn Pathol 2014.

     http://www.ncbi.nlm.nih.gov/pubmed/25565221

    http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4305978/pdf/1746-1596-9-S1-S8.pdf

     

    [8] Besusparis J, Plancoulaine B, Rasmusson A, Augulis R, Green AR, Ellis IO, Laurinaviciene A, Herlin P, Laurinavicius A: Impact of tissue sampling on accuracy of Ki67 immunohistochemistry evaluation in breast cancer. Diagnostic Pathology 2016.

     https://pubmed.ncbi.nlm.nih.gov/27576949/

     

    [9] Plancoulaine B, Laurinaviciene A, Herlin P, Besusparis J, Meskauskas R, Baltrusaityte I, Iqbal Y, Laurinavicius A: A methodology for comprehensive breast cancer Ki67 labeling index with intra-tumor heterogeneity appraisal based on hexagonal tiling of digital image analysis data. Virchows Archiv 2015.

     https://link.springer.com/article/10.1007/s00428-015-1865-x

     

    [10] Laurinavicius A, Plancoulaine B, Herlin P, Laurinaviciene A: Comprehensive Immunohistochemistry: Digital, Analytical and Integrated. Pathobiology 2016.

     <Go to ISI>://WOS:000375025100012

    https://www.karger.com/Article/Pdf/442389

     

    [11] Laurinavicius A, Plancoulaine B, Rasmusson A, Besusparis J, Augulis R, Meskauskas R, Herlin P, Laurinaviciene A, Muftah AAA, Miligy I, Aleskandarany M, Rakha EA, Green AR, Ellis IO: Bimodality of intratumor Ki67 expression is an independent prognostic factor of overall survival in patients with invasive breast carcinoma. Virchows Archiv 2016.

     https://link.springer.com/article/10.1007%2Fs00428-016-1907-z

     

    [12] Zilenaite D, Rasmusson A, Augulis R, Besusparis J, Laurinaviciene A, Plancoulaine B, Ostapenko V, Laurinavicius A: Independent Prognostic Value of Intratumoral Heterogeneity and Immune Response Features by Automated Digital Immunohistochemistry Analysis in Early Hormone Receptor-Positive Breast Carcinoma. Front Oncol 2020.

     https://www.ncbi.nlm.nih.gov/pubmed/32612954

    https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7308549/pdf/fonc-10-00950.pdf

     

    [13] Zilenaite-Petrulaitiene D, Rasmusson A, Besusparis J, Valkiuniene RB, Augulis R, Laurinaviciene A, Plancoulaine B, Petkevicius L, Laurinavicius A: Intratumoral heterogeneity of Ki67 proliferation index outperforms conventional immunohistochemistry prognostic factors in estrogen receptor-positive HER2-negative breast cancer. Virchows Arch 2024.

     https://www.ncbi.nlm.nih.gov/pubmed/38217716

    https://link.springer.com/article/10.1007/s00428-024-03737-4

     

    [14] Rasmusson A, Zilenaite D, Nestarenkaite A, Augulis R, Laurinaviciene A, Ostapenko V, Poskus T, Laurinavicius A: Immunogradient indicators for anti-tumor response assessment by automated tumor-stroma interface zone detection. Am J Pathol 2020.

     https://www.sciencedirect.com/science/article/pii/S0002944020301267#appsec1 https://www.ncbi.nlm.nih.gov/pubmed/32194048

     

    [15] Nestarenkaite A, Fadhil W, Rasmusson A, Susanti S, Hadjimichael E, Laurinaviciene A, Ilyas M, Laurinavicius A: Immuno-Interface Score to Predict Outcome in Colorectal Cancer Independent of Microsatellite Instability Status. Cancers (Basel) 2020.

     https://www.ncbi.nlm.nih.gov/pubmed/33050344

    https://res.mdpi.com/d_attachment/cancers/cancers-12-02902/article_deploy/cancers-12-02902-v2.pdf

     

    [16] Stulpinas R, Zilenaite-Petrulaitiene D, Rasmusson A, Gulla A, Grigonyte A, Strupas K, Laurinavicius A: Prognostic Value of CD8+ Lymphocytes in Hepatocellular Carcinoma and Perineoplastic Parenchyma Assessed by Interface Density Profiles in Liver Resection Samples. Cancers (Basel) 2023.

     https://www.ncbi.nlm.nih.gov/pubmed/36672317

    https://mdpi-res.com/d_attachment/cancers/cancers-15-00366/article_deploy/cancers-15-00366-v2.pdf?version=1673921305

     

    [17] Zilenaite-Petrulaitiene D, Rasmusson A, Valkiuniene RB, Laurinaviciene A, Petkevicius L, Laurinavicius A: Spatial distributions of CD8 and Ki67 cells in the tumor microenvironment independently predict breast cancer-specific survival in patients with ER+HER2- and triple-negative breast carcinoma. PLoS One 2024.

     https://www.ncbi.nlm.nih.gov/pubmed/39576843

    https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0314364&type=printable

     

    [18] Radziuviene G, Rasmusson A, Augulis R, Grineviciute RB, Zilenaite D, Laurinaviciene A, Ostapenko V, Laurinavicius A: Intratumoral Heterogeneity and Immune Response Indicators to Predict Overall Survival in a Retrospective Study of HER2-Borderline (IHC 2+) Breast Cancer Patients. Front Oncol 2021.

     https://www.ncbi.nlm.nih.gov/pubmed/34858854

    https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8631965/pdf/fonc-11-774088.pdf

     

    [19] Morkunas M, Zilenaite D, Laurinaviciene A, Treigys P, Laurinavicius A: Tumor collagen framework from bright-field histology images predicts overall survival of breast carcinoma patients. Sci Rep 2021.

     https://www.ncbi.nlm.nih.gov/pubmed/34326378

    https://www.nature.com/articles/s41598-021-94862-6.pdf

     

    [20] Stulpinas R, Morkunas M, Rasmusson A, Drachneris J, Augulis R, Gulla A, Strupas K, Laurinavicius A: Improving HCC Prognostic Models after Liver Resection by AI-Extracted Tissue Fiber Framework Analytics. Cancers (Basel) 2023.

     https://www.ncbi.nlm.nih.gov/pubmed/38201532

    https://mdpi-res.com/d_attachment/cancers/cancers-16-00106/article_deploy/cancers-16-00106.pdf?version=1703409446

     

    [21] Laurinavicius A, Rasmusson A, Plancoulaine B, Shribak M, Levenson R: Machine-learning-based evaluation of intratumoral heterogeneity and tumor-stroma interface for clinical guidance. Am J Pathol 2021.

     https://www.ncbi.nlm.nih.gov/pubmed/33895120

    https://ajp.amjpathol.org/article/S0002-9440(21)00165-6/pdf

     

    [22] Yousif M, van Diest PJ, Laurinavicius A, Rimm D, van der Laak J, Madabhushi A, Schnitt S, Pantanowitz L: Artificial intelligence applied to breast pathology. Virchows Arch 2021.

     https://www.ncbi.nlm.nih.gov/pubmed/34791536

    https://link.springer.com/content/pdf/10.1007/s00428-021-03213-3.pdf

     

     

     

        Privacy & Cookie Policy
    (c) 2016 All rigths reserved. Solution: TEXUS

    Our website uses cookies. We use these cookies for statistical and marketing purposes.

    If you consent to the use of these cookies, please click “accept” and continue browsing our website.

    Agree
    Settings