2006 Breakout Sessions
A1 Challenges to Pathology Informatics in the Era of the Electronic Medical Record
Bruce A. Friedman, M.D
We are now entering an era in healthcare when the electronic medical record (EMR) will reign supreme as the central integrated clinical record, with the various supporting clinical systems like the laboratory information system (LIS) replicating data to it. This computing model raises a number of strategic questions and challenges for pathology informatics that will be discussed in detail in this presentation. On balance, the future for the field is bright for pathology informatics because of the increasing sophistication and complexity of laboratory test that will be routinely performed for patients and the increasing decentralization of testing venues. Some of the clouds on the horizon will also be discussed in the presentation including the need for more cost-effective delivery of services and the increasing influence of C-level executive in the management of clinical information in healthcare.
A2 Progress towards Clinical Implementation of Automated Whole Slide Imaging
John R. Gilbertson, MD, Jeffrey L. Fine, MD, and Johan Ho, MD
View Slide Presentation - Gilbertson
View Slide Presentation - Fine
View Slide Presentation - Ho
The past several years have seen significant technical advancement in the ability to image large numbers of slides automatically, rapidly and at high resolution. This advance will continue as imaging robot manufactures develop increasingly powerful designs within a general environment of ever more capable electronics and optics. The ability to routinely digitize entire slides should allow pathologists to better apply computer power and network connectivity to the study of pathologic morphology and eventually improve the efficiency of surgical pathology practice.
However, the implementation of large scale clinical whole slide imaging will require more than the installation of capable imaging robots. It will require significant changes to the workflows, mindset and informatics environment of the pathology department. The diagnostic utility of the images need to be validated and business plans tested, faculty need to be bought into the project and guidelines developed. A variety of supporting devices and services such storage facilities, networks, workstations (GUIs and monitors), bar-coding systems and the LIS will have to be either implemented de novo or modified, and data integration between these system will need to be developed. Large-scale clinical use of whole slide imaging will be a considerable effort that needs to be justified and paid for.
This talk will discuss ongoing efforts to implement large scale clinical applications in whole slide imaging with examples take largely but not exclusively from ongoing activities at the University of Pittsburgh. It will focus in five main areas:
- Capabilities of existing commercial imaging robots
- "Clinical" validation studies and faculty acceptance of WSI
- Implementation of supporting infrastructure such as bar coding, storage facilities, workstations and LIS modification
- Image and data integration;
- Costs and budget justification.
A3 The BioMOBY Interoperability System: Designing Bioinformatics Workflows Using Semantic Web Services
Mark D. Wilkinson, PhD
Though the number and breadth of bioinformatics data and analytical resources on the Web is increasing, it remains extremely difficult to "pipeline" these resources into analytical workflows. Most Web-based resources are provided only as Web pages, designed for human-readability. To automate the extraction of data requires custom programming of "screen scrapers", which themselves are quite fragile and prone to breakdown if the layout of the web page changes. Moreover, submission of data for analysis often requires re-formatting to suit the input requirements of the analytical interface, thus additional custom programming, or more often hand-editing and copy/pasting, is required.
The BioMOBY Interoperability Project was established in 2001, and is now becoming a widely used platform for provision of data and analytical services over the Web. BioMOBY defines an ontologically-based messaging system and a "yellow pages" registry of BioMOBY-compliant service providers. Together, these accomplish two fundamental tasks: 1) An end-user extensible ontology of valid data structures is defined, thus alleviating the need to cut/edit/paste when passing data from one Service to another, and 2) the registry can interpret in-hand data structures, and use these to discover and facilitate communication with Web Services that can operate that data. Thus BioMOBY can guide the user through the construction of an analytical pipeline that subsequently runs with no additional human intervention.
In this workshop we will provide a brief glimpse of the "guts" of the BioMOBY interoperability system in order to better understand how and why it works, and this will be followed by a hand-on session where a useful BioMOBY analytical workflow is created and run using Taverna – a powerful graphical workflow designer/editor created by the UK myGrid project.
B1 Practical Survival Tactics for Pathology Data Management in the Era of Electronic Medical Records
Walter H. Henricks, MD
The increasing use of electronic medical records systems (EMRs) and computerized physician order entry (CPOE) in hospitals and in physician offices has profound implications for elements of information management that are crucial to successful operation of pathology laboratories. While anatomic and clinical pathology data are among the most important elements in the EMR, optimal presentation of pathology results and test request information is typically not a high priority in EMR systems or projects. This session illustrates challenges, pitfalls, and potential solutions, using real world examples, related to pathology information management in the context of EMRs.
After attending this session, participants should be able to:
- Describe factors driving the adoption of EMRs in healthcare
- Anticipate problems for their laboratories that may arise in test ordering and result reporting as a consequence of EMR and CPOE use.
- Develop a basis for greater involvement in EMR issues related to pathology data management.
- Identify opportunities for improvement in management of pathology data in their EMR system.
B2 Advances in Region-of-Interest-Based Query for Histopathology
Ulysses J. Balis, MD
View Slide Presentation
With a number of recent advances in complex data set theory, it now becomes possible to apply N-dimensional query approaches to the daunting problem of searching image database repositories based upon a spatial, image based (and not textual) search construct. This session will explore the underlying technology and some immediate uses made possible by its targeted application, within the domain of histopathology.
Multispectral Imaging - Beyond RGB
Michael Feldman, MD, PhD
The use of digital image acquisition systems has rapidly expanded in the last 5 years. Multiple vendors now offer systems that allow whole slide images to be acquired and used in research and diagnosis. Both static an whole slides imaging systems are not limited to just RGB color space and more recent advances have seen the introduction and adoption of multispectral image acquisition systems that allow for analysis of routine stains to be analyzed using greater color space. In addition, multispectral imaging systems allow for multiplexing of immunolgic and in situ hybridzation methodologies on tissues combined with histopathologic registration of target signals. Both of these technologies are well suited to the application of computer algorithms to facilitate data analysis.
In this breakout session, participants will learn about
a. Multispectral imaging system systems
b. Discuss different computer algorithms for analysis of whole slide images, static images and multispectra images
B3 HIPAA, Tissue Banking, and Data Aggregation: New Tools to Solve Old Problems
Rajiv Dhir, MD and Ashok Patel, MD
Research endeavors involve aggregation of biological materials and data. The tissue support facility is responsible for disbursing the required materials (biological samples and data) to researchers in a de-identified fashion. This need for de-identification has become even more significant with the advent of HIPAA regulations. In addition recent modifications of OHRP regulations provide opportunities for further streamlining, and enhancing, data aggregation for research purposes.
Maintaining the confidentiality of protected health information is very critical and must be addressed appropriately by every research resource. This session will discuss the basic aspects of HIPAA, introduce the concept of an honest broker, discuss basic and expanding data needs, introduce software tools for the tissue and data providers, and demonstrate an honest broker software tool being used routinely by the Tissue Bank and Research Registry teams involved in a collaborative honest broker service at the University of Pittsburgh.
This session will reveal methods for developing a Collaborative Honest Broker Service to support Tissue Banking, Research Information Services and Clinical Trials Research. These collaborative efforts result in the reduction of duplication of activities and streamline the process for de-identification and integration of pathology, tissue bank and cancer registry data. An in-depth overview of the development and management of this collaborative service, examples of processes used in de-identification and tools used to monitor the use of data from multiple data sources will be presented.
At the end of the session the participant should be able to:
- Describe biological specimen and data annotation in a large academic center
- Define HIPAA regulations and OHRP guidelines
- Define the functions of an Honest Broker
- Define software tools for honest broker activities
- Describe integrating and streamlining research workflow
C1 Acquisition and Storage of Digitized Whole Slide Images – The Radiology Experience
Edward M. Smith, ScD, FACNP and John S. Koller
Using radiology as a model, we will discuss the components of the information (data and image) acquisition and storage systems and relate them to whole slide imaging in pathology.
Acquiring the image is just one aspect of the information acquisition process since the correct patient demographics, required administrative data, imaging parameters and technique must be associated with the image. This requires bi-directional communication and synchrony of databases between the modality, multiple clinical data, e.g. HIS and RIS, and image, e.g. PACS, systems from different vendors. To accomplish this, standards must be established and adhered to by all components of the enterprise information system. For the most part this has been accomplished over many years and forced cooperation among vendors in radiology. Pathology is at the embryonic stage of this integration process, but can take advantage of the many misadventures and achievements of radiology.
The first portion of this session will review the image acquisition process and associated data integration required for a valid image file as well as the purchasing standards for modalities and data systems required to acquire and store the image file. The current array of whole slide acquisition and digitizing systems will be reviewed from the standardization perspective.
The second portion of this session will discuss the storage and management of fixed and variable content files in radiology in the context of the healthcare enterprise. Areas to be discussed will include storage requirements, information availability, information life cycle management, redundancy, disaster recovery and business continuance from a radiology perspective and how this experience relates to the pathology business case.
At the conclusion of this session, attendees should be able to:
- Describe the steps and components required to acquire an image and integrate it with the necessary clinical information.
- List the obstacles that must be overcome to enhance workflow and productivity if whole slide digitization is to become routine in clinical pathology
- List the differences and requirements for managing and storing fixed and variable content files
- Explain the various components of the healthcare information storage and management system
- Explain the advantages of an enterprise storage system versus a departmental storage system
- Discuss the lessons pathology can learn from the experience radiology has gained from implementing a digital information environment available to the entire healthcare enterprise
D1 Proteomics Informatics – From Fundamentals to Applications
Francis Ouellette, Leonard Foster, David Wishart, Ronald Beavis and Raymond Ng
After attending this session, participants should be able to:
- Describe basic concepts, key principles, major challenges
- Describe lessons in designing proteomic experiments and in using informatics tools to probe proteomic data
- Discuss quality issues in and analytical approaches to plasma proteomic
D2 Case Studies in Informatics
This Session will consist of a series of 15 - 20 minute talks, each of which will be immediately followed by an audience question and answer session. The goal of the session is to provide attendees with insight into the evolution of technology and services available from corporate partners in the informatics community. Presentations will be in a case study format and will be delivered by a client or by a client in conjunction with a member of the firm. Elements that will be emphasized include success factors, return on investment, innovative or developmental aspects, problems solved, and vision for the future.
Case Study Schedule and Abstracts (PDF)
D3 Frontiers in Proteomics Informatics
Bruce McManus, MD, PhD, FRSC, Raymond Ng, PhD, Ronald Beavis, PhD, and David Wishart, PhD
The challenge of making large sets of annotated proteomics data available to the scientific community will be discussed. The efforts of informatics groups to produce large, useful, publicly available repositories of proteomics information will be compared and the current status of these projects will be reviewed.
Participants will become familiar with several web-based tools used to perform both peptide mass fingerprinting (PMF) and tandem mass peptide identification and/or de novo sequencing. The underlying algorithms and statistical principles used by these programs to process MS data and assess MS data quality will be discussed.
In this session we will present an overview of our experience in calibrating an ITRAQ platform. The calibration is not only important for verifying the reproducibility of the protein data collected, but is also valuable for determining the key parameters (e.g. fold change cutoff, false positive rate) for subsequent analysis.
After attending this session, participants should be able to:
- discuss strategies for amassing and using large proteomic datasets
- discuss new and practical horizons in proteomics informatics
- discuss use of tools to ensure quality of proteomic data analysis
E1 Progress toward the Interoperable Electronic Surgical Pathology Report
Gunther Schadow, MD, PhD
Healthcare information standards have been used successfully to submit Pathology data into the Electronic Medical Record (EMR) for decades. For chemistry laboratory data, interoperable data submissions works very well and is in fact the most widely used form of interoperable communication of structured EMR data. Conversely, electronic reporting in Surgical Pathology has traditionally been limited to electronic free text documents which are unsuitable to populate research databases, annotated tissue banks, or cancer registries. Recently interest has grown to develop interoperable data standard for pathology reports. The requirement by the American College of Surgeons that anatomic pathology report contain the "synoptic" abstracts defined by the College of American Pathologists (CAP) have helped the business case for interoperable pathology data. Thus, Pathology Laboratory Information System (LIS) vendors are adding functions for synoptic data entry templates, and public health agencies and cancer registries pilot implementations of structured data exchange.
These parties have come together in the HL7 standard organization to define everything that is needed for interoperable data exchange. This talk will give an update on the standardization activities. It will demonstrate the major components required for complete interoperability, including HL7 version 2 messages, HL7 version 3 information structures used for messaging and clinical documents, as well as the synoptic templates and the required terminology such as LOINC and SNOMED. The talk will also touch on approaches of implementing the standard with existing LIS and tissue banking systems. The goal is to prepare interested parties for the rapid implementation of the forthcoming final specifications in their practice.
E2 Robotic Telepathology: Practical Applications
Bruce E. Dunn, MD
Telepathology involves the transmission and viewing of video and static images for the purposes of rendering primary or consultative diagnoses by pathologists at a distance. Since mid-1996 we have operated a routine telepathology service using a commercially available, robotic, hybrid dynamic store-and-forward system (HDSF) between the Iron Mountain (MI) and Milwaukee (WI) Veterans Affairs Medical Centers which are separated by a distance of over 220 miles. Pathologists located in Milwaukee can control slide movement, adjust the focus and change the objectives and lighting intensity on the robotically controlled microscope located in Iron Mountain. This service was implemented as a cost-effective alternative to maintain frozen section diagnostic capability and to provide more timely routine pathology service upon the retirement of a single on-site pathologist. A medical technologist, who is also board certified as a pathologist assistant, serves to dissect specimens and generate annotated images of gross specimens when needed. Our initial feasibility study of 100 cases published in 1997 demonstrated that for routine surgical pathology cases the accuracy of robotic HDSF pathology was not significantly different from that of light microscopy compared to consensus diagnoses. Subsequent work published in 1999 involving our first 2200 cases validated these observations. To date, we have rendered over 8,000 primary diagnoses by robotic telepathology. In general, clinician satisfaction with telepathology services has been high.
After attending the session, participants should be able to:
- Define practical applications of robotic telepathology.
- Describe the importance for surgical pathologists working at a distance to be able to view gross tissue specimens using video and static images.
- Describe key enabling features required to maintain a successful robotic telepathology program.
Advantages and Disadvantages of the Robotic Telepathology System
Paul Weisz-Carrington, MD
Objectives: 1) Review Robotic Telepathology System at the McGuire Veterans Affaire Medical Center
2) Experience with the system and system justification
3) Critique of the system
E3 National HL7 Standards for Electronic Pathology Reporting to Cancer Registries
Eric B. Durbin
Electronic pathology (E-Path) reporting is playing an increasingly important role in cancer registry data acquisition and use. E-Path reporting has been shown to reduce the registry workload while improving data accuracy and timeliness. Reducing the delay from the cancer pathology report to the cancer registry’s capture of a case greatly increases the registry’s potential to support pathology and oncology research. E-Path reporting offers similar benefits to cancer registries participating in tissue banking. Increased utilization of E-Path reporting has demonstrated the need for a national data transmission standard. To address this need the North American Association of Central Cancer Registries (NAACCR) formed the HL7 E-Path Transmissions Work Group in 2003. The work group consists of representatives from pathology laboratory information system vendors, cancer registry software vendors, pathology labs, cancer registries and government agencies. As a result of this effort, the “Implementation Guide for Transmission of Laboratory-Based Reports to Cancer Registries using Version 2.3.1 of the Health Level Seven (HL7) Standard Protocol” was officially published in Standards for Cancer Registries, Volume V, Pathology Laboratory Electronic Reporting by NAACCR in 2006. We are pleased to report that the first pilot project with a large pathology lab has validated the specifications of this standard. The NAACCR E-Path Standard and Work Group progress to date will be presented.
The attendee will learn about:
- NAACCR’s efforts to define and establish the national HL7 standard for electronic pathology reporting.
- The NAACCR HL7 Version 2.3.1 E-Path Standard, including details about the use of HL7 ORU^RO1 Unsolicited Observation Messages, specified message segments, requirements and example messages.
- The experiences of implementing E-Path reporting with the NAACCR standard.
- Continuing Work Group efforts to extend the standard to accommodate synoptic pathology reporting and tissue banking.
- The role of the NAACCR HL7 E-Path standard in the work of a newly formed HL7 Pathology Special Interest Group.
CDC-NPCR Pilot Projects Using SNOMED CT Encoded CAP Cancer Checklists
Kenneth A. Gerlach, MPH, CTR, BS
The Centers for Disease Control and Prevention (CDC) has administered the National Program of Cancer Registries (NPCR) since 1994. CDC-NPCR supports central registries and promotes the use of registry data in 45 states, the District of Columbia, and the territories of Puerto Rico, the Republic of Palau, and the Virgin Islands. NPCR complements NCI's Surveillance, Epidemiology, and End Results (SEER) registry program. Together, NPCR and the SEER program collect cancer data for the entire U.S. population.
The purpose of this presentation is to describe two pilot projects to implement a new means of collecting and transmitting pathology information to cancer registries using the SNOMED CT encoded College of American Pathologists (CAP) cancer checklists and HL7 messages. In 2001 CDC-NPCR funded two states, California and Ohio, for a pilot project with a focus on colon and rectum cancers. In 2004, CDC-NPCR initiated a second pilot project for cancers of the breast, prostate, and melanoma of the skin. Three states have been selected to participate in this project: California, Maine, and Pennsylvania.
This first project ended in 2005. Participating pathology laboratories used an electronic version of the checklists and the information was formatted into a project standard HL7 message using codes in the Logical Observation Identifiers Names and Codes (LOINC) database as the questions and SNOMED CT codes as the answers. The second project for cancers of the breast, prostate and melanoma is currently in the implementation phase. Project participants include representatives from NPCR and hospital cancer registries, hospital anatomical pathology (AP) laboratories, AP laboratory information system (LIS) vendors, SNOMED International, and HL7 messaging consultants. These participants have formed two work groups: messaging and evaluation. The Messaging Work Group is currently meeting to define the common HL7 message data items (or segments) and the associated structure of the CAP checklist data. The Evaluation Work Group is currently meeting to establish evaluation measures.
F1 Practical APLIS-based Structured/Synoptic Reporting
David L Booker, MD and Anil Parwani, MD, PhD
Synoptic Reporting is gaining acceptance in the anatomic pathology community, but these reports continue to be free-text based rather than structured because Anatomic Pathology Laboratory Information Systems (APLIS’s) generally do no support structured synoptic reporting. There is increasing demand for such support in pathology practice and research; for example, since January 1, 2004 the American College of Surgeons Commission on Cancer (ACS CoC) has required that its accredited institutions demonstrate that they report required data elements from the College of American Pathologists (CAP) Cancer Protocols (http://www.cap.org/apps/docs/cancer_protocols/protocols_index.html). Because current APLIS’s incorporate these data elements in blocks of unstructured text rather than in discrete data elements, they are not amenable to automated analysis or data sharing.
Dr. Parwani will discuss the use of a “synoptic reporting” tool (Cerner CoPath Plus) used at UPMC to incorporate the CAP checklists as discrete data elements, allowing for storage of data elements in a relational table within the APLIS. Synoptic worksheets were created for select organ systems and malignancies such as prostate, melanoma, breast and lung. He will discuss the experience at UPMC in creating the synoptic tools with the overall goal of promoting easy data sharing between the APLIS, Cancer Registry and other Clinical and Research systems in Oncology.
Dr. Booker will discuss the obstacles to the use of such systems and elements required for systems to become successful. He will demonstrate the benefits of structured/synoptic reporting tools (SoftPath Synoptic Reporting, Soft Computer Consultants) designed for use in general surgical pathology (not restricted to Cancer Protocols). Finally, he will describe current efforts to incorporate Synoptic Reporting into Health Level Seven (HL7) standards and technologies and the need for Common Data Elements (CDE’s) to promote data sharing and interoperability of such systems.
Objectives:
- Synoptic Worksheet entry in the LIS and its incorporation into existing Pathology Reports
- Synoptic data entry as methods of producing standardized reports, leading to improved pathology reports
- Experience of production of synoptic data entry worksheets and comparison to CAP checklists
- Obstacles to use and elements of successful systems
- Uses of synoptic reporting other than with Cancer Protocols
- Common Data Elements (CDE’s) and HL7
- Questions and Answers
F2 A Working Group of Our Own (DICOM WG-26)
Bruce A. Beckwith, MD
The goal of this portion of the digital imaging standards session is to give an update on the state of DICOM with regards to pathology images. A major new development is the establishment of a working group within DICOM dedicated to Pathology. We will consider the current state of the DICOM standard with regard to pathologic images and how some departments are using DICOM compliant files and tools to for image transfer, storage and retrieval. The mission and current work of the DICOM Pathology working group will be summarized along with the relationship of this effort to the API’s LDIP project and the recently created HL7 Pathology group. At the conclusion of this session, the participants will have a better understanding of the progress and interrelationship of the relevant electronic imaging standards efforts underway in the Pathology domain.
Informatics Laboratory Digital Imaging Project
Jules J. Berman, PhD, MD
The Laboratory Digital Imaging Project (LDIP) is a three year effort (that began in October, 2004) to provide pathologists and image vendors with a way of conveying well-specified clinical and research information within pathology image files. The LDIP committee has over 40 members and is composed of software vendors, service pathologists, informaticians, image experts, and researchers who have a stake in developing a comprehensive and complete image specification. The LDIP image specification will enhance the utility of pathology images, will enable pathology image data to be integrated with data derived from diverse biomedical data sets, and will support portability to standard image formats, including DICOM. The purpose of this session's presentation is to review LDIP's progress to date, and to invite discussion of future goals for the LDIP committee.
Adoption of Standards Can Enhance Productivity and Workflow for Digitizing Whole Slide Images – the Radiology Experience
Steven C. Horii, MD, FACR, FSCAR
In this session, we will discuss how radiology pioneered standards to increase productivity, enhance workflow and reduce errors in a digital environment that extends beyond radiology to the entire healthcare enterprise. We will then discuss the importance standards will play if whole slide imaging is to participate clinically in the digital environment.
The first standard developed for digital imaging in radiology was Digital Imaging and Communication in Medicine (DICOM). This standard has evolved over a twenty plus year time-frame and is in its third and hopefully final form. It started out as a standard for cable connectors so a device from one vendor could connect with that of another. The second phase was device to device connectivity so that data could be transmitted and used between devices from different vendors. The current state of DICOM is to ensure connectivity and interoperability over an enterprise network for devices “speaking” DICOM.
This presents the problem that the imaging devices, speaking DICOM need to communicate and exchange information with clinical data systems that use a different standard, Health Level 7, (HL7). The solution to this disparity was to develop interfaces that translated various HL7 “dialects” so the data system could understand each other and systems “speaking” DICOM could understand systems “speaking” HL7. This latter stage has proven to be very time consuming and costly.
About five years ago, a joint effort between the vendors and academia, initiated the technical framework called Integrating the Healthcare Enterprise (IHE). The objective of IHE was to promote interoperability and enhanced workflow between systems comprised of components from different vendors by developing Integration Profiles for specific clinical operations. These integration profiles have been developed and tested for cardiology and other specialties outside of radiology.
At the conclusion of this session, attendees should be able to:
- List the standards currently used in radiology and explain how they increase productivity, enhance workflow and reduce errors.
- Explain how pathology will benefit from utilizing the lessons learned from radiology’s experience in the evolution of standards based practice.
F3 Disease Surveillance - Applying Pathology Informatics to Public Health
Raymond D. Aller, MD
Pathology has led numerous aspects of development of clinical informatics – less known is that it has also pioneered many of the underlying technologies and standards that today make possible establishment of the public health disease surveillance network. Pathologists' comfort and familiarity with building and maintaining large analytical databases (such as the laboratory information system) prepares us to contribute to the current public health effort in bioterrorism preparedness and response. We will review the many mechanisms for use of data – produced as a byproduct of clinical care – in detecting disease outbreaks in our communities.
At the conclusion of this presentation, the participant will be able to:
- Describe at least three types of data from clinical medicine that are being used to detect disease outbreaks.
- Discuss four or more non-technical barriers delaying connection of clinical data sources to the public health database.
- List at least three types of unique contributions, critical to the establishment of public health disease surveillance, that originated in pathology.
- Not make excuses about why his/her laboratory isn't pursuing connection with their local public health disease surveillance system.
G1 A Population-Based Laboratory Information Strategy
Michael D. D. McNeely, MD
There will be an ever-increasing need for laboratory results to be knowledge-based: to be interpreted, to guide treatment, and to smoothly integrate with the medical record.
Canada Health Infoway is a government of Canada project whose goal is to have electronic medical records (EMR) for 80% of Canada’s population by 2010. The Provincial Laboratory Information Solution is a BC project to provide a unified database of all laboratory results produced in the province. These two projects are at an early stage but eventually (phase III-IV) will incorporate knowledge support.
The presentation will, by way of a review, discuss the potential for these initiatives to carry forward existing programs involving laboratory utilization control, risk management, chronic disease management, telepathology, epidemiology, genominformatics, and sample management.
At the conclusion of this presentation, participants should be able to:
- Discuss and have contact information for two significant population-based initiatives, which involve the integration of laboratory information and the EMR.
- Explain the potential for knowledge-based computer applications in the clinical laboratory, many of which can be initiated with “low tech” approaches.
G2 Pathology as Integrative Research Biology
Robert D. Cardiff, M.D., PhD
Research data, collected by the entire scientific team, are now provided in digitized imaging format. These data need to be captured, stored and retrieved for analysis by the pathologist who is the one scientist trained to recognize the structure of disease and thus integrate structure and function. Whole slide imaging provides the pathologist with an unprecedented tool for sharing, integrating and analyzing research images. In this session, examples of whole slide imaging research, as well as the strengths and weaknesses of current imaging application, will be discussed.
Mining the Images: Extracting Meaningful Information from Microscopy Images
Tony Pan, MS
Pathology and laboratory imaging are currently experiencing a digital revolution, similar to the revolution radiology has gone through in the past 20 years. Increasingly, more and more departments are moving towards whole slide imaging and electronic record keeping. The large amount of digitally captured image data creates a necessity for computer-based analysis and large-scale data processing and management.
In this talk, we will illustrate the utility of computer-based analysis and data management through several real-world image analysis applications, such as virtual mouse placenta phenotyping, computer-aided clinical decision support for neuroblastoma, and mouse brain confocal quantitative imaging. These applications engage a wide range of medical image analysis techniques including post-acquisition image correction, pixel and tissue classification and segmentation, 3D reconstruction and visualization, and feature quantification. Distributed data management and processing using clusters and grid computing will also be discussed in the context of these applications.
After attending the session, participants will have a better understanding of
- the principles and benefits of automated or semi-automated image analysis systems,
- the components and requirements of an end-to-end image analysis application,
- the current capabilities and challenges of the microscopy image analysis.
G3 caTIES - Transforming Text and Promoting Partnership in the caBIG Tissue Banking Workspace
Rebecca Crowley MD, MS and Kevin Mitchell, MS
The Cancer Text Information Extraction System (caTIES) project is ongoing research that utilizes and contributes to the caBIG Tissue Bank and Pathology Tools (TBPT) workspace. caTIES focuses on two important challenges of biomedical informatics: information extraction from free text and access to tissue. As one of the earliest grid-enabled caBIG applications - caTIES v 2 has also been grappling with the challenges of the new research roadmap. As we evolve from isolated silos of data towards collaborative data sharing networks, our systems must balance the needs of patients, clinicians, researchers, developers, and organizations. This break-out session will focus on the technology needed for such a social transformation, along five critical dimensions:(1) tissue banking collaboration, (2) grid trust fabric, (3) concept coding and recoding, (4) data stewardship, data sharing and honest brokering, and (5) interoperability within a grid community. We will describe some of the barriers we’ve breeched and other challenges that lie ahead. As part of this breakout, we will also provide a demonstration of the caTIES system and describe its current reach and capacity.
Online Access to Archival Tissue Samples – The Harvard Virtual Specimen Locator
Bruce Beckwith, MD
Pathology departments contain a wealth of archived tissue samples which have associated clinical information in the form of pathology reports and other data residing in hospital information systems. In 2001, the National Cancer Institute funded a feasibility study (the Shared Pathology Informatics Network or SPIN) with the goal of designing a system to allow researchers search de-identified clinical information from multiple institutions in order to locate human tissue samples available for research. As a member of the SPIN project, we have extended the prototype tools developed in SPIN and implemented a virtual bio-repository called the Virtual Specimen Locator (VSL). The VSL is a peer-to-peer network in which each participating tissue repository maintains local control of its information and samples. This Harvard-wide resource has been approved by four Institutional Review Boards and currently includes information on over 600,000 specimens from four different hospital pathology departments. The search tool is web based and can display statistical information or, with appropriate authentication, individual specimen information including portions of the de-identified pathology report. We are currently performing proof on concept studies and plan to extend the scope to include both frozen and specialty tissue banks as well as clinical outcome data. At the conclusion of this session, the participants will have a better understanding of the opportunities and challenges provided by distributed systems for sharing pathology specimens and data.