A brief introduction to virtual microscopy is given. It is then developed how Virtual Microscopy will become an essential element in Medical Systems Biology. Concluding, it is shown how the individual talks integrate into the large field of virtual microscopy application development.

Microscopy is presently undergoing a transformation due to the wide introduction of digital image acquisition devices. Against this background, in the recent past we have focused our activities to the evolving field of automated morphometric image analysis in microscopy. This talk will give a brief overview on our previous and ongoing work for various life science disciplines.

What is needed to quantify alzheimer plaque on virtual slides, in addition to an algorithm, which reliably detects the corresponding objects? It is necessary to have an efficient work flow, which allows to analyse the huge amount of images and is to a large extent automated. The occurring problems and solutions due to the practical experience are presented.

The talk motivates the use of epithelial tissue cultures in virtual microscopy. Since the majority of tissues in humans comprise of epithelia, modelling of cornerstones of epithelial homeostasis forms the basis for prospective epithelial simulation. Such modelling is in turn indispensible for the understanding of epithelial physiology and pathophysiology in systems biological terms. To contribute to this goal, we have for the first time developed quantitative spatial profiles of biomarkers, indicating distinct stages of cell differentiation, a key element in epithelial homeostasis.

The rapidly growing collection of fruit fly embryo images makes automated image segmentation and classification an indispensable requirement for a large-scale analysis of in situ hybridization (ISH) gene expression patterns (GEP). We present here such an automated process flow for segmentation, classification, and clustering large-scale sets of Drosophila melanogaster GEP that is capable of dealing with most of the complications implicated in the images.

Tissue banks are essential platforms for modern biomedical and translational research. Beside tissue storage, distribution, project management and support in morphology based techniques, tissue banks offer the ideal platform for decentralized tissue based research projects. Virtual microscopy is an ideal tool to provide centralised scanning of decentral obtained stains, especially multi-tissue-arrays, data storage, and data analysis as well as decentralized access to data, their analyses as well as reference stains. These tools can be extended through variation of tissue technology (fluorescence techniques) and bioinformatics.

To provide accuarate information that enables Pathologists and Oncologists to determine the treatment of the disease (predictive) and patient outcome (prognostic)is critical. Therefore it needs additional quantitative biomarkers which correlate to clinical outcome. In this process "Intelligent Image Analysis" plays an important role and moves into the heart of R&DS and Diagnostics.

Quantitative analysis of the HER-2 immunostained cytoplasmic membrane of cancer cells in breast tissue represents an important challenge in digital pathology. Here we introduce a novel, highly discriminating algorithm for scoring HER-2, and thereby also present the widely applicable principles of image sampling, data management, and quantification in whole slide images provided by the Visiopharm Oncotopix™ software.

TissueGnostics (TG) develops new generations of hardware- and software-tools for imaging, analysis, diagnosis and documentation. New interactive forms of acquisition and analysis ensure control of result and quality.

With the development of Herceptin, accurate evaluation of HER-2 status has become particularly acute in the clinical laboratory. However, recent reports of discordance regarding interpretation of results has cast doubt on the reliability of HER-2 testing and clinical guidelines have suggested that image analysis could prove to be an effective tool for achieving consistent interpretation. Towards this goal we have developed a HER-2 image analysis algorithm that classifies cases based on both the intensity and continuity of membrane staining. The performance of the algorithm was validated across a large cohort of clinical cases and achieved 91% concordance with manual review by a pathologist, in addition to a lower reporting rate of equivocal cases. Agreement with gene amplification status as the gold standard was also found to be higher for the automated review, suggesting that this technology has great potential as a diagnostic support tool to increase the standardisation of HER-2 evaluation.

Whole slide imaging as an image acquisition technique includes more than the challenge to distribute microscopic images in a hospital.Different types of applications need to involve different groups of the medical domain on different scales.We will show an example for an administrative-medical workflow in the field of digital pathology.

Since Virtual Microscopy concept was published in 1997, it became an indispensable tool to support digital pathology applications. A key technology is slide scanning, and its advances enable to create a high resolution virtual slide in a few minutes. Furthermore, new features, such as virtual fluorescence slides and Z-stack for focusing, are available to expand the applications.

Two years ago at Erasmus Medical Center the conventional microscopy for students was replaced by virtual microscopy. The implementation including test phase, where server capacity, software capacity and optimal infrastructure were determined by simulation. Now, teachers spend more time on teaching then helping students to have the microscope set up properly and finding the histology of interest and can discuss the image while both can see the image, resulting in better interaction. Evaluation showed students also appreciate the virtual microscope technology far above conventional microscopy.

The RWTH Aachen University offers medical students since June 2006 the histology of the practical courses as virtual slides through Internet. We analyzed the acceptance of the web-based digital microscopy by evaluation of questionnaires filled by students and also by analysis of the number of visits on the server. We found that students not only appreciate the possibility of learning microscopy through the Web, but also use the system intensively previous examination. Moreover, the integration of the virtual histology into a video podcast containing macroscopical images, drawings etc. has improved the quality of the lessons and thus the acceptance of clinical pathology as an attractive subject.

Pathorama is a publicly accessible modular e-learning platform including a searchable virtual slides atlas with more than 600 virtual slides. Various contributors feed categorized metadata including the slide URL into the central slide database while the slides are stored on servers at the contributing institutes. Pathorama is used at all levels of pathology education.

The department of Histopathology at University College London Hospital has now had a Hamamatsu nanozoomer for approximately one year. This presentation will describe our first years experience with virtual microscopy in a mixed clinical/research environment. Potential clinical and research applications of this technology will be discussed together with the factors that may hinder more generalised utilisation of this technology.

One of the most important application fields for virtual microscopy are telepathology and teleconsultation. We have used virtual microscopy for both purposes in a clinical trial with outside hospitals and report here on the use of telepathology for frozen tissue diagnosis and teleconsultation for difficult cases that require a second opinion. In our experience, telepathology is equivalent to conventional microscopy for frozen sections if no evaluation of the macroscopic aspects of the specimen is required. This applies e.g. to the frozen section diagnoses in sentinel lymph node biopsy of breast cancer. We will present here information on the prerequisites and results in terms of sensitivity and specificity of the method.

The histological assessment of human tissue is the key point for detection and treatment of cancer. We employ randomized detection trees and survival statistics to automate and objectify some of the most crucial tasks in modern pathology. I am going to present a framework for biomarker quantification on tissue microarrays from renal clear cell cancer patients.

The tipping point is a term commonly used to describe that moment when ideas, trends and behaviors cross a threshold, tip and spread like wildfire resulting in a complete shift in our thinking and practice. We are currently on the verge of such a tipping point in microscopy. Virtual Microscopy is radically altering our perception of microscopy.

Virtual microscopy is widely accepted in Pathology for educational purpose and tele-consultation but is far from the routine use in surgical pathology due to the technical requirements and some limitations. What are the basic requirements for routine use? Which scenarios are useful and cost-effective? Which tools we need? Answers will be given, algorithms will be presented and open questions will be referred.

Urokinase plasminogen activator (uPA) and its inhibitor (PAI-1) are clinically approved tumor markers in primary breast cancer with strong prognostic impact. Automated scanning, followed by computer-assisted quantitative scoring of uPA and PAI-1 expression assessed by immunohistochemistry is an interesting alternative to the established ELISA measurements, recently recommended for clinical application in the ASCO guidelines 2007.

In colorectal cancer (CRC) quantitative cell counts based on immune cell surface molecules (CD3, CD8, Granzyme B, and CD45RO) have been shown to be a potentially better predictor for patient outcome than the classical TNM system. Using virtual microscopy we have studied the intratumoral heterogeneity of the investigated immune cell surface molecules. Computational simulation of Tissue Microarrays (TMA) from the virtual slides shows that intratumoral heterogeneity of all measured surface molecules was so high that a reliable measurement of the surface molecules for an individual patient can only be achieved by microscopy of the full tissue sections but not from TMAs. We suggest that also studies for the definition of potential biomarker cut-offs -and their later use in individual patient’s predictions-based on quantitative cell counts should include quantitative assessments of intratumoral heterogeneity. This can only be reliably achieved by measurement of full tissue slides using virtual microscopy.

The standardization of IHC is a task with high priority. Different domains (medical, tissue handling, reports) are covered by this process. The task of IHE (Integrating the Healthcare Enterprise) is to identify the necessary actors, transactions and content elements of messages for an integrated communication between automatons and Pathology Laboratory Information Systems. The IHE presented the Technical Framework as a first step for modeling of processes in Pathology.

Current cervical cancer screening is based on the detection of abnormal cells in cytological specimens prepared from cervical smears. Virtual microscopy has the potential to improve this process at several levels. Web-based evaluation systems allow cytologists to analyze specimens without having the glass slide at the same physical location. Archiving of digital microscopy images preserves slides under optimal conditions and saves storage space. Novel biomarkers allow for automated detection of abnormal cells, increasing sensitivity and reducing evaluation time.