The optimum practice of transplantation pathology requires considerable experience and accrual of a “critical mass” of cases.  Regular feedback during discussions with clinical colleagues is mandatory.  Transplantation pathology also requires considerable infrastructural support for sophisticated laboratory testing used to monitor graft-versus-host reactions, uncommon infections and tumors.  Consequently, most leading transplant programs are located within regional tertiary care centers.  But in most instances, follow-up care and donor organ procurement take place at local facilities that usually have little experience with transplantation-related patient management issues; they rely on the major transplant centers for guidance.  Even in these larger centers, expertise in Transplantation Pathology often lags behind Transplantation Surgery.

This current state of affairs weakens outcome-based research in transplantation.  Very few large transplantation centers accrue enough patients on a yearly basis to effectively evaluate new treatment regimens.  Most large transplant centers carry out between 50 and 100 operations per organ per year and a significant percentage of these patients may be excluded from consideration in individual trials because of co-existent conditions.  Consequently, multi-center studies are needed to achieve the statistical power necessary for meaningful analysis and confident conclusions.  Since transplant pathology-generated data originating from multiple centers frequently are used as endpoints in such trials, the need for standardization, precision and quality assurance of this information is clearly of key importance for successful outcome.

We have made several attempts to address the needs for accurate and standardized diagnoses in transplantation pathology.

1)         Maintaining an “open door” policy for training and continuing education of visiting pathologist(s) from new centers that have a need for experience in this area. 

2)         Creating an Internet site dedicated to Transplantation Pathology: Transplant Pathology Internet Services (TPIS; http://tpis.upmc.edu; Appendix A).  It provides image-rich didactic reference materials; standard grading systems; and topic and organ-based list server e-mail discussion groups. 

3)         Providing a forum for anonymous electronic posting of images from interesting and/or difficult cases to any pathologist with a digital camera and internet access ( http://136.142.177.165:8080/~caseconf ) under the umbrella of TPIS. 

4)         Developing a web-based Training, Quality Assurance and Coordination of Pathology Data for Clinical Trials module that is specifically designed for multicenter trials ( http://136.142.177.165/ ).

5)         Developed a private client-server based Telepathology System (Appendix B) that is currently being used for direct professional consultation on specific transplant (and occasional nontransplant) patients from sites in Palermo, Italy and Kyushu, Japan.  We are currently considering invitations to extend these services to transplant centers in Amristar, India; and Tehran, Iran.

Costs for initial setup of all of the hardware, software, and programming and maintenance of these services are partially covered through a combination of grants, corporate sponsorship and private funds.  Content for the site is provided by the staff of Transplantation Pathology at the University of Pittsburgh Medical Center.  However, given the growth of these activities, the increased demand for services and lack of specifically dedicated personnel, it has become increasingly difficult to maintain a robust level of services.  This proposal addresses the creation of a formal organizational structure for an internet-based transplant pathology consortium that fills existing needs for clinical service, quality assurance, multi-center studies, and outcomes-based research. 

Existing Components of the Proposed Consortium and Relevant Medical, Computing and Multi-Center Study Experience

Several components of the proposed transplantation pathology consortium already exist.  The following is a brief description of the individual modules, and relevant medical and computing experience of the participants.  A more detailed description of each of the modules is given in the appendices.  The plan is to combine these functional components into an integrated, adaptable non-profit organization dedicated to improved patient care, standardization, and outcomes-based research in transplantation.

Transplant Pathology Internet Services

Transplant Pathology Internet Services (TPIS) is an interactive Web-site (See http://tpis.upmc.edu and Appendix A) intended to serve as both a collaborative tool for transplantation physicians and as an educational resource for healthcare professionals interested in the field of transplantation medicine with an emphasis on pathology.  It serves as a base for reference material in transplantation pathology and as an interactive platform for focused discussion and dissemination of information.  The image-laden didactic reference materials include all of the internationally accepted grading and staging systems for acute and chronic allograft rejection, complete with illustrative image examples of each grade/stage and e-mail list servers for organ or disease-specific group discussions.  Related grading systems, such as those used for chronic hepatic autoimmune hepatitis and posttransplant lymphoproliferative disorders, are also provided.

In addition, any pathologist with a digital camera and internet access can anonymously submit digital images from difficult or interesting cases for educational review and discussion. Posting images at this site automatically sends an e-mail containing the anonymous clinical history to the internet community of pathologists and other interested medical professionals who subscribe without cost to this particular service. Users can then log into the site using a standard web browser, view the associated images and comment on the case, if so motivated.  This rudimentary telepathology system (adapted e-mail list server and chat box modified to function as an image server) is functional and useful as a teaching exercise, but the public format inhibits interactions and precludes discussion of specific patients.

Seed funding for TPIS was provided by the College of American Pathologists Foundation Scholars Program and the TPIS administration remains committed to the specific aims outlined in our mission statement (see Appendix A).  These aims emphasize collaborative efforts among academic transplant centers and specifically include: 1) Regular online case conferences of unusual and enlightening cases from multiple transplant centers; 2) Special topics resources moderated by physicians from multiple transplant centers whom we anticipate will evolve into an editorial board for electronic publishing of collaborative study findings; and 3) Original transplant pathology material that can be hypertext linked from within a Web-based customizable e-mail server and database.

Endpoints analyses in transplant pathology are frequently based on a number of histomorphologic tissue features such as architectural distortion and the development of fibrosis.  Currently, assessment of various parameters is based on the subjective opinion of local pathologists who rely on “semi-quantitative” analyses.  A simple, user-friendly, web-based automated image analysis software would be an ideal solution to the problem of standardization of interpretation.  It could be used, for example, to objectively assess morphometric endpoints such as the severity of fibrosis and architectural distortion in clinical trials.

We have already created such a system, which is a modification of the NIH-supported image analysis software (ImageJ) offered on the internet.  Our programmers have adapted this software to the specific needs of transplantation pathologists.  It performs a variety of functions designed by our group to fit the needs of transplantation pathologists.  The software is specifically designed to quantify percent area (chosen by the pathologist), including fibrosis and inflammation, provide information of area shapes (e.g. portal tracts in developing cirrhosis) and severity of lumen narrowing (e.g. transplant vasculopathy).  The original NIH software can be examined at http://rsb.info.nih.gov/ij/docs/index.html.  The source code of this software is freely available and lends itself to the addition of plug-in modules to solve specific image processing or image analysis problems.  The first plug-in module we have written, called “Ratio Calculator” was designed to allow the rapid calculation of the percentage of fibrous tissue in a biopsy (http://tpis.upmc.edu/tpis/ImageJ).  The program runs from any standard web browser.  The user simply calls up image files from his/her own computer for analysis and the program will automatically segregate the tissue sample from the non-tissue region of the photograph.  The user then selects a typical area of the liver to serve as “background”.  The program stores this information.  The user then selects a representative spot of interest, the “foreground”.  The program automatically determines the ratios of red/green/blue and selects regions of similar or greater color saturation as contrasted with background liver.  The selected area is falsely colored to provide feedback to the user as to what areas have been selected.  The process is iterative so the user can repeat or undo the selection.  A ratio of selected area to total liver area is calculated and presented to the user.

Telepathology Support for Mediterranean Institute for Transplantation and Advanced Specialized Therapies

Transplant pathologists at the University of Pittsburgh Medical Center in Pittsburgh, PA have created a Web-based interactive “Virtual Pathology Division” that includes the transplant pathologist at the Mediterranean Institute for Transplantation and Advanced Specialized Therapies (ISMETT) in Palermo, Italy through a static or “store and forward” telepathology system.  The main purpose of the Telepathology system is to provide rapid second opinion consultation between the ISMETT pathologist, Dr. Marta Minervini, and transplant pathologist consultants at the University of Pittsburgh.  To date, more than 100 cases have been posted and evaluated.  A complete description of the system and our experience with its use is provided in the attached manuscript, which has been accepted for publication to Human Pathology (Appendix B).  The link also allows pathologists from both centers to participate in unknown slide and quality assurance conferences.

It should be noted that the telepathology system was developed in-house ”from scratch” because no existing commercial system could meet our requirements and schedule.  This approach also provided the ability to adapt the software specifically for the needs of transplantation pathology; for example, in this area of medicine review of previous biopsies from the same patients and access to standard reference material often facilitates the process of diagnosis.  Workflow requirements for the systems were: 1) 24 hours/7days week coverage; 2) real time frozen section support; 3) support for templated organ transplant data elements. Significantly, communication bandwidth was limited to 33.5 kbps through a private network between ISMETT and UPMC (with modem access directly to the pathologists’ homes).  Bandwidth limitations forced a decision to implement a static image telepathology system that could be used for both real time interactive sessions and store and forward.  The system has the capability to be integrated to dynamic, higher resolution image and/or remote control system in the future.  However, at this time, the system utilizes the “lowest common denominator” approach to enable access from the greatest number of users, some of whom may not have access to higher speed internet connections.

Important features of the Telepathology System include:

1.      Software modules designed to mimic the standard workflow for “on-site” transplantation pathology, with links to previous cases from the same patients, which enable an assessment of the effect of therapy on rejection and other causes of allograft injury.

2.      Ability to designate “consultant” pathologist by the referral pathologist, and electronically page consultant in real time for urgent cases.

3.      Clinical and pathology information fields specialized for transplantation, with drop down template menus for digital standardization.  Standardized data is then stored in a database for analysis.  Histological features examined during a normal workflow are captured as digital data, obviating the need to “re-review” images/slides for research purposes, unless so desired.

4.      Real time chat capability that allows pathologists to “discuss” cases interactively.  All conversations are stored in a database as a part of the case folder.

5.      Patient and site confidentiality maintained by user session authentication based on the role-based user profile securely stored inside the system. Included in this role definition are the “sites” to which a user is authorized. This allows the central team of consulting pathologists to see cases from all remote sites, but appropriately restricts remote users to seeing only those cases from their particular site. This functionality makes the system easily scaleable for larger groups while respecting the confidentiality requirements of each site.

6.      All comments, preliminary diagnoses and final diagnoses are tagged with the identity of the author and the time and date of the comment. All data is stored and protected with a multiple tier data backup mechanism. Data is never removed from a case once entered and committed to by the pathologist. Future plans include the use of a PKI (public key infrastructure) to digitally sign and message digest all images and final diagnosis text, though this is not yet required by regulation or policy.

7.      The overall system was designed to foster adherence to telepathology guidelines as issued by the American Telepathology Association (Appendix C).

8.      Real clinical cases and images can be made anonymous “electronically”.  This facilitates educational use of interesting cases and provides a “built-in honest broker” system to satisfy confidentiality requirements for national and local institutional review committees when conducting research on clinical cases.

Offsite Training, Quality Assurance and Coordination of Pathology Data for Clinical Trials

            Pathologists of TPIS have conducted and/or coordinated offsite training, quality assurance and pathology data collection for 3 separate clinical trials in which the incidence and severity of acute liver allograft rejection or hepatitis was used as an endpoint.  Prior to initiation of each study, a designated “central pathologist” (a TPIS staff member) would coordinate a “training session” during which “projection” microscopes were used to show glass slides on a closed circuit television monitor.  Thereafter, a limited number of glass slides were circulated to the pathologists in the room to view at their rented microscopes.  During these sessions, study pathologists reviewed typical examples of the various grades of acute and chronic rejection and other post-transplant complications, such as viral hepatitis.

These conferences were extremely difficult to coordinate because of scheduling conflicts among as many as 20 pathologists from North America and Europe, even when attempts were made to “piggyback” training session on to national or international conferences.  Consequently, the TPIS team developed a Web-based training and testing module that could be utilized for offsite training, quality assurance testing and continuing educational purposes (Appendix B).  In essence, a study pathologist logs in to a secure internet site and applies for a username and password, which are restricted to study members.  Once in the site, the study pathologist has access to didactic training materials, if s/he determines it is necessary to review materials prior to quality assurance testing.  Quality assurance testing consists of “typical” case-based examples of allograft biopsies illustrating the various grades of rejection and other post-transplant complications.  The study pathologist reviews the cases and fills in histological templates specific to the organ system and coded diagnosis(-es), which are entered into a database for analysis.  When the pathologist has successfully completed the training session, e-mail notification is immediately sent to all interested parties.  This system can also be used for periodic quality assurance re-testing, periodic study updates and conducting reliability studies.

Interactions with the Center for Pathology Informatics

The Center for Pathology Informatics at the University of Pittsburgh Medical Center is the first center of its kind in the country. Pathology informatics is a developing discipline that focuses on the management and analysis of clinical and research pathology data using modern computing and communications techniques. Since its inception as the Division of Pathology Informatics in 1991, the Center has focused on providing high-quality training experiences for pathologists at all levels.  By sharing experience and discussing common problems, interactions with the Center for Pathology Informatics will facilitate development of the Transplant Telepathology Consortium.

The Center for Pathology Informatics is responsible for managing and improving UPMC laboratory information services and systems in anatomic pathology (Dynamic Healthcare Technology, Inc.'s CoPath PicsPlus), clinical pathology (Sunquest Information Systems), and the pathology imaging system (TriPath Imaging, Inc.'s Autocyte Image Management System). The Center also manages its own teleconsultation, image-enhanced reporting, image indexing and distance education systems. These systems are integral to medical decision making at UPMC and are responsible for over half of the data in the UPMC's electronic patient record. The Division is also responsible for academic and administrative computing in the Department of Pathology, and supports a robust developmental and academic program.

• The AP Laboratory Information Services group (APLIS) provides computing resources and support for all of the divisions comprising Anatomic Pathology. These divisions include surgical pathology, autopsy pathology, and transplant pathology throughout the UPMC Health System. The Division of Anatomic Pathology is growing, with particular emphasis on oncogenesis and genitourinary, hematologic, head and neck, pulmonary, bone and soft tissue, orthopedic, endocrine, cytologic, and transplant pathology.

  In addition to providing development, training and support, the APLIS Team supports Dynamic Healthcare's CoPath PicsPlus for UPMC's large network of Anatomic Pathology departments. The APLIS team is active in many other clinical developmental projects, including voice recognition, document imaging, and gross/microscopic image capture. The team also provides assistance to the annual conference, Advancing Pathology Informatics, Imaging and the Internet (APIII), in Pittsburgh.

• The CP Laboratory Information Services group (CPLIS) provides computing resources and support for all of the divisions of Clinical Pathology. These divisions include clinical chemistry, immunopathology and tissue typing, microbiology, and hematopathology throughout the UPMC. In addition to providing development, training, and support, the CPLIS team supports Sunquest Information Systems for UPMC's network of Clinical Pathology departments in eight affiliated hospitals.

  The CPLIS team is actively developing and implementing a process control program for Point of Care Testing Systems, interacting with the labs to implement robotic control modules into the LIS and lab workflow. CPLIS works with the LIS vendor to evaluate new products and is serving as an alpha and beta test site for future product development.

• Developmental Pathology Informatics is the research and development arm of the Center. It focuses on cutting-edge developmental and research programs, with an emphasis on next-generation pathology reporting methods, databases, imagery, and bioinformatics. The Center has substantial grant support in the areas of next-generation Internet test beds, imaging research, collaborative human tissue banking, development of tissue banking information systems and bioinformatics support for the molecular reclassification of prostate cancer.