Center of Excellence

Using the power of systems biology to decipher multifactorial diseases and address the complexity of biological systems - Interview with Diana Mechtcheriakova

Diana Mechtcheriakova (right) with two former PhD students, Felicitas Mungenast (left) and Anastasia Meshcheryakova (center)

Briefly describe your research interests and role at the Institute for Pathophysiology and Allergy Research, Centre for Pathophysiology, Infectiology, and Immunology at the Medical University of Vienna.

I am an Associate Professor at the Medical University of Vienna in Austria and head of the Molecular Systems Biology and Pathophysiology research group at the Institute of Pathophysiology and Allergy Research, Center for Pathophysiology, Infectiology, and Immunology. Due to my deep expertise, I am also head of the technology subunit at BioImaging Austria, and of the Systems Medicine subdivision at the Austrian Platform for Personalized Medicine.

My scientific directions are systems biology and systems medicine as part of personalized medicine, B-cell biology, and immuno-oncology. Special interests refer to the AID/APOBEC-associated biological events; architectural complexity, functionality, and clinical relevance of lymphoid structures; developing of systems biology algorithms for the discovery of novel disease-/immune-associated checkpoints and patients’ stratification strategies; and the cellular sphingolipid system in immunity and cancer.

What TissueGnostics solutions do you use at your institute and how do they help you with your research?

My research group has been using the power of TissueGnostics technologies and supervising the incorporation of new developments for almost 14 years. The cornerstone of the next-generation quantitative tissue image cytometry is the computerized microscopy-based TissueFAXS platform for digitalizing the whole-slide tissue sections upon staining in IHC or IF.

 The corresponding software packages, such as HistoQuest and TissueQuest, enable the analysis of digital images on the single cell level with the possibility for quantitative assessment across individual regions of interest, specified by a researcher, or across the entire digital tissue image. The analysis is further enforced by the unique properties of the StrataQuest software, which combines the information and extracts the knowledge based on both anatomical and cellular features of the analyzed tissue of the patient specimen.  Overall, the tissue-encrypted information gets computationally transformed into numerical values, representing the variables for the follow-up analyses. We use the staining-derived datasets for alignment with clinicopathological parameters including patient stratification into risk groups and biomarker nomination. The advanced TissueFAXS-based technological solution provides the image-derived data in omics format and is an integral part of our systems biology approach, the DIICO – from Digital Immune Imaging to Clinical Outcome algorithm, to address the complexity of biological systems. This is image-driven systems biology as part of personalized medicine.

 The DIICO strategy led us to the pioneering discovery of the formation of functionally active ectopic lymphoid structures at the metastasis in the liver of patients with colorectal cancer.

DIICO algorithm enabled us to transform tissue-encrypted information on immune cells and lymphoid structures into 24 variables for the assessment of their clinical relevance. We showed a strong prognostic effect of immune phenotype characteristics of lymphoid structures not only at the metastatic site but, surprisingly, also in normal colon mucosa of patients with metastatic colorectal cancer.

1. Mungenast F.*, Meshcheryakova A.*, Beer A., et al. and Mechtcheriakova D. The Immune Phenotype of Isolated Lymphoid Structures in Non-Tumorous Colon Mucosa Encrypts the Information on Pathobiology of Metastatic Colorectal Cancer. Cancers 2020. doi: 10.3390/cancers12113117. This study gives novel clinically relevant insights into the role of lymphoid structures in cancer pathobiology gained using quantitative tissue image cytometry as an innovative technological solution within the newly developed DIICO algorithm.
2. Meshcheryakova A., Svoboda M., Jaritz M., et al. and Mechtcheriakova D. Interrelations of sphingolipid and lysophosphatidate signaling with immune system in ovarian cancer. Computational and Structural Biotechnology Journal. 2019. doi.org/10.1016/j.csbj.2019.04.004. The first study addressing joint impact of sphingolipid/lysophosphatidate systems on ovarian cancer pathobiology in conjunction with local immune response implementing a newly developed holistic approach – the MuSiCO algorithm. Digital pathology using the TissueFAXS platform is one of the modules of MuSiCO.
3. Meshcheryakova A., Tamandl D., Bajna E., et al. and Mechtcheriakova D. B cells and ectopic follicular structures: novel players in anti-tumor programming with prognostic power for patients with metastatic colorectal cancer. PLoS One. 2014. doi: 10.1371/journal.pone.0099008. Pioneering work interrelating B cells and lymphoid structures with clinical outcome of patients with metastatic colorectal cancer; proposal of novel patient stratification models based on biomarkers discovered using whole-slide computerized quantitative microscopy.
4. Meshcheryakova A., Zimmermann P., Ecker R., et al and Mechtcheriakova D. An Integrative MuSiCO Algorithm: From the Patient-Specific Transcriptional Profiles to Novel Checkpoints in Disease Pathobiology. In: Rajewsky N., Jurga S., Barciszewski J. (eds) Systems Biology. RNA Technologies. 2018. Springer, Cham DOI https://doi.org/10.1007/978-3-319-92967-5_18. This book chapter gives a detailed description of the multi-modular integrative MuSiCO algorithm applicable for deciphering the pathomechanisms of complex multifactorial diseases.
5. Meshcheryakova A., Mungenast F., Ecker R., Mechtcheriakova D. Tissue image cytometry. Imaging Modalities for Biological and Preclinical Research: A Compendium, Volume 1. 2021. (pp. I.2.h-1-I.2.h-10): IOP Publishing. doi.org/10.1088/978-0-7503-3059-6ch14. The book chapter gives detailed insights into digital tissue image cytometry in basic and translational research.

Interview with Isabella Ellinger

Isabella Ellinger

Briefly describe your research interests and role at the Institute for Pathophysiology and Allergy Research, Centre for Pathophysiology, Infectiology, and Immunology at the Medical University of Vienna.

I am Head of the Placenta Pathophysiology Research Group and Deputy Head of the Division of Cellular and Molecular Pathophysiology at the Institute of Pathophysiology and Allergy Research (IPA at the Center for Pathophysiology, Infectiology & Immunology, MedUni Vienna).

I am a researcher (first mission) and a teacher (second mission) and I aim to contribute to the three main cornerstones of the academic system of MedUni Vienna: research, teaching, and patient care. I feel connected to the UN Sustainable Development Goals (SDGs), especially Goal 3.

Sustainable Development Goal 3 calls for "ensuring healthy lives and promoting well-being for all at all ages," with a focus on reproductive, maternal, newborn, and child health. In this context, an important aspect of my research deals with understanding the functions of the placenta. The placenta attempts to ensure proper development of the fetus and thus is also critical to the development of the adult phenotype of the offspring. It can adapt when confronted with adverse environmental conditions, such as maternal malnutrition or overnutrition, exposure to drugs or environmental pollutants, or other adverse conditions such as hypoxia. However, when the placenta's adaptive capacity is exceeded, the fetus is compromised, which can subsequently impact the fetus' lifelong health. Given some of the major health problems of our time, such as malnutrition on the one hand and increasing obesity rates on the other, as well as the many environmental pollutants known today, I believe that deciphering the physiological and pathophysiological functions of the placenta is extremely important for our future.

Another focus of Goal 3 of the SDGs is universal access to basic health services. Histopathologic image analysis, for example, is necessary for the diagnosis of malignant lesions. However, even for experienced pathologists, the diagnostic process is not trivial. Diagnostic agreement between specialists averages only 75%. In addition, there is a shortage of pathologists in many parts of the world. These limitations have led to the development of computer-aided diagnosis (CAD) systems based on automated image analysis algorithms. As second opinion systems, CAD systems are expected to reduce the workload of specialists, improve diagnostic efficiency, and help reduce costs. In addition, automated image analysis is a Big Data analysis tool of great interest for biomedical research. MedUni Vienna has a special focus on precision medicine and Big Data analysis. I am interested in using and developing automated image analysis in ongoing and future research projects to identify new biomarkers, better understand specific diseases, and advance personalized medicine.

What TissueGnostics solutions do you use at your institute and how do they help you with your research?

Researchers at IPA are fortunate to have a TissueFAXS Plus scanning software-controlled microscope for fluorescence and brightfield imaging and the latest versions of the image analysis software products HistoQuest, TissueQuest, and StrataQuest.

Since 2008, TG and I have been collaborating in EU and nationally funded research projects, developing algorithms for structure recognition, classification, and segmentation. Since 2016, we are mainly interested in using machine learning-based algorithms, especially Deep Learning (DL)-based algorithms, which have shown excellent performance for many image analysis tasks such as object recognition, image classification, and semantic and instance segmentation. While TG has the computer scientists, I contribute relevant research questions and provide histological samples along with data essential for the development, training, and validation of DL-based algorithms. To date, we have worked on DL-based algorithms, e.g., for skin lesion classification, breast cancer histological image classification, microscopic pollen image classification, food ulcer segmentation, and most importantly, nucleus segmentation and classification (see reference 1). This DL-based nuclear segmentation algorithm was among the top 10 methods in the MoNuSAC challenge leaderboard (see reference 2), was the top-ranked method in the MoNuSAC post-challenge leaderboard, and was integrated into the latest StrataQuest software release.

Furthermore, in our collaboration, we have also prepared relevant Big Data sets (such as the CryoNuSeg data set, see reference 3) with concomitant ground truth data further enabling the training and validation of nuclei segmentation algorithms.

More recently, members from my group, and Robert Nica from TissueGnostics, have also developed image analysis algorithms using immunofluorescently labeled placental tissue sections to be able to study the structure and protein expression of human and rat placenta in situ, with the goal of further deciphering the physiological and pathophysiological functions of the placenta.

What are your most important reference publications?

1. Mahbod Amirreza, Schaefer Gerald, Dorffner Georg, Hatamikia Sepideh, Ecker Rupert,  and Ellinger Isabella. A dual decoder U-Net-based model for nuclei instance segmentation in hematoxylin and eosin-stained histological images. Front Med (Lausanne). 2022 Nov 11;9:978146. doi: 10.3389/fmed.2022.978146. PMID: 36438040; PMCID: PMC9691672. IF:5.058             
2. Verma Ruchika, Kumar N, Patil A, Kurian NC, Rane S, Graham S, Vu QD, Zwager M, Raza SEA, Rajpoot N, Wu X, Chen H, Huang Y, Wang L, Jung H, Brown GT, Liu Y, Liu S, Jahromi SAF, Khani AA, Montahaei E, Baghshah MS, Behroozi H, Semkin P, Rassadin A, Dutande P, Lodaya R, Baid U, Baheti B, Talbar S, Mahbod Amirreza, Ecker Rupert, Ellinger Isabella, Luo Z, Dong B, Xu Z, Yao Y, Lv S, Feng M, Xu K, Zunair H, Hamza AB, Smiley S, Yin TK, Fang QR, Srivastava S, Mahapatra D, Trnavska L, Zhang H, Narayanan PL, Law J, Yuan Y, Tejomay A, Mitkari A, Koka D, Ramachandra V, Kini L, Sethi A. MoNuSAC2020: A Multi-organ Nuclei Segmentation and Classification Challenge. IEEE Trans Med Imaging. 2021 Dec;40(12):3413-3423. doi: 10.1109/TMI.2021.3085712. Epub 2021 Nov 30. PMID: 34086562. IF:11.037
3. Mahbod Amirrza, Schaefer Gerald, Bancher Benjamin, Löw Christine, Dorffner Georg, Ecker Rupert, and Ellinger Isabella. CryoNuSeg: A dataset for nuclei instance segmentation of cryosectioned H&E-stained histological images. Comput Biol Med. 2021 May;132:104349. doi: 10.1016/j.compbiomed.2021.104349. Epub 2021 Mar 22. PMID: 3377 IF:6.698