INSERM, UMR, Paris
University René Descartes, Paris
Department of Pathology
Harvard Medical School and Beth Israel
Deconess Medical Center, Boston, MA
Marcial García Rojo
Department of Pathology
Hospital General de Ciudad Real
Department of Pathology
Charité Universitätsmedizin Berlin
Pathology Informatics Case Comprehensive Cancer Center
Case Western Reserve University
Anatomic pathology is a diagnostic medical discipline that establishes disease identification, pathogenesis and prognosis (eg, tumour typing and prognostic factor analysis in cancer) by use of both basic and modern techniques applied to tissues, cells, subcellular and molecular
Traditional pathology information systems collect and manage data on a complex workflow from specimen collection to report editing. In recent years, pathologists have begun to use images to document various steps in the diagnostic process. Because imaging and traditional laboratory information systems (LIS) are not usually integrated, information and image acquisition often involves time-consuming double data entry. Orders, images and reports are spread out over different systems that do not interoperate. Although standards bodies HL7(1) and DICOM (Digital Imaging and Communications in Medicine)(2) are working to provide integration solutions, they do not address the specific needs of pathology. Moreover, messages supplied via the HL7 or DICOM usually contain a large number of optional data fields, so that being DICOM- or HL7-compliant does not imply direct integration.
The goal of the Integrating Healthcare Enterprise (IHE) project is to specify how data standards should be implemented to meet specific healthcare needs, and to make systems integration more efficient and less expensive.(3) Based on the findings of working groups that include users and manufacturers, IHE defines “integration profiles” that are real-world situations describing exchange of information, called “transactions”, from various functional components of a distributed healthcare environment, called “actors”.
IHE provides implementation guides for transactions using DICOM or HL7. The annual definition cycle of new profiles by users and suppliers, ending in the organisation of international platforms of interoperability tests (“connectathons”), confers its unique efficiency, transforming basic standards into “plug and play” solutions.
If errors in, or extensions to, existing standards are identified, IHE’s policy is to report them to HL7 or DICOM for resolution within their conformance and standards evolution strategy.
In 1996, the Association for the Development of Informatics in Cytology and Pathology (ADICAP), working in collaboration with software producers, proposed a European de-facto standard for image folders.(4) The aim is now to promote the use of international standards in the development of information systems in pathology.
In 2005, ADICAP, the Spanish Society of Anatomic Pathology (SEAP) and the Spanish Health Informatics Society (SEIS), working with the Group Promoting the Modernisation of Hospital Information Systems in France (GMSIH), launched the IHE pathology initiative in Europe. At the same time, IHE-Japan and the USA’s Laboratory Digital Imaging Project (LDIP) were working on a model for a pathology information system.
US, European and Japanese groups agreed that, although specific DICOM objects are defined for pathology, extensions are necessary since the DICOM model did not initially describe specimens in sufficient detail. A DICOM working group (DI-COM WG26) was set up in September 2000 to address these issues.(5) Meanwhile, the HL-7 Anatomic Pathology Special Interest Group (SIG) was established to investigate the complex relationships between specimens, observations, protocols and documents in anatomic pathology. There are currently ballots taking place on the HL-7 specimen model.
Modelling anatomic pathology workflow
Ten pathologists and haematologists, five informatics technology specialists and eight software manu-facturers volunteered to take part in 12 working sessions on HL7 between September 2005 and June 2007. The international working group reviewed the literature on order forms and pathology reports in order to describe the main requirements for the structure and its content of orders, imaging folders and reports. When they had defined pathologists’ needs and created a corresponding integration profile (pathology worklow), they identified the actors and transactions involved.
The diagnostic process in anatomic pathology (Figure 1) differs from the one that takes place in the clinical laboratory since it relies extensively on visual morphology and spatial relationships. This makes imaging a potentially important tool. It differs from radiology, however, in that the subject of an image is usually a specimen (not a patient). In addition, a single diagnostic study may involve multiple images of multiple related specimens (parts, blocks and/or slides) captured with many different types of imaging equipment. In some cases, an image can include more than one specimen or even tissue from more than one patient (eg, tissue microarrays).
A “pathology workflow” (PWF) profile was proposed to establish the continuity and integrity of basic pathology data acquired for examinations being ordered for an identified inpatient or outpatient. It focuses on the main workflow transactions of ordering, reporting and imaging. The PWF involves seven actors and specifies six transactions to maintain the consistency of ordering information and specimen management information. For each transaction, the workgroup proposes the use of the most suitable format – HL7 (v2 or V3, including clinical document architecture [CDA]) or DICOM.
The order for the pathological examination is communicated between the order placer (of the order entry system) and the order filler (of the laboratory information system). Order form quality assurance provides a list of mandatory items: order identification, order date and time, identification of the ordering physician and the ordering care department (including call-back telephone number), patient identification (eg, PID, name, visit number), identification of the patient’s care unit (if different from the ordering care department), order priority (date and time when results are expected), etc.(6) Each order may contain one or more requested procedures, which may be reported by different pathologists. A requested procedure is defined as a unit of work resulting in one report with associated codified, billable acts.
Two transactions (placer order management and filler order management) contain all the messages required between the order placer and the order filler for the management of the lifecycle of the order.
In pathology, the image folder (study) is defined at case level. Image acquisition may require different modalities (eg, gross imaging, microscopic imaging). A new series of images is created whenever an imaging procedure step is performed on a new specimen or by a new type of equipment.
Thanks to the transaction worklist provided, the image acquisition system asks the laboratory information system for patient demographics and information about the specimens that are to be imaged. The transaction procedure schedule and update allows the order filler to send the image manager and report manager information about the imaging procedure or procedure update.
Since 1993, the Association of Directors of Anatomic and Surgical Pathology has published recommendations for pathologic reporting.(7) A generic model of a structured report can be derived from these templates. Furthermore, studies of report quality assurance provide lists of mandatory items and stress the positive role of checklists to enhance the reporting process.(8) According to “evidence-based pathology”, only features that are reproducible and relevant – with a demonstrated diagnostic or prognostic significance – should be reported in descriptions with corresponding evidence available.(9,10)
A pathology report is a clinical document and should include all the relevant features (eg, author, date/time, patient ID). It might include findings and the evidence for them, expressed in structured data, free text, images, etc. Observations should be tied to specific specimens. Procedures should to be tied to laboratory manuals, etc. The HL7 anatomic pathology special interest group is working on a CDA model for a pathology report that supports this level of complexity. A crucial issue is to identify a technical solution to handle templates of structured reports, including findings and associated evidence.
The report management transaction details changes to the report and report status and changes (corrections and cancellations) between order fillers to the enterprise report repository.
The modality image stored transaction is based on two transactions already defined in the radiology domain and allows an acquisition modality or an evidence creator sending acquired or generated images to the image archive. Acquisition modality or evidence creator asks the image manager to confirm ownership for the specified DICOM objects (images, key image notes, evidence documents, or any combination of them) stored in the image archive, thereby allowing the sender to delete those objects.
Defining a specimen identification
Before beginning this work, DICOM had several information object definitions. These are useful in anatomic pathology, including taking VL (visible light) photographic images (XC) for gross specimens and VL slide coordinate microscopic images (SM) for slide-based microscopic imaging. However, the IODs (information objects definitions) did not have a strong mechanism for describing the specimen being imaged or associating a particular specimen with a particular image.
Following creation of the DICOM WG-26, seven IHE-Pathology-DICOM WG26 working sessions were organised between September 2005 and June 2007 on the DICOM specimen identification and description module.
Specimen identification (DICOM supplement 122)
DICOM supplement 122 is available for public comment. The module defines formal DICOM attributes for the identification and description of laboratory specimens when specimens are the subject of a DICOM image.
The module is focused on the specimen and laboratory attributes necessary to understand and interpret the image. These include those that:
- Identify (specify) the specimen (within a given institution and across institutions).
- Identify and describe the container in which the specimen resides.
- Describe specimen collection, sampling and processing.
- Describe the specimen or its ancestors when these descriptions help with the interpretation of the image.
Attributes that convey diagnostic opinions or interpretations are not within the scope of the specimen module. The DICOM specimen module does not seek to replace or mirror the pathologist’s report.
A specimen is a physical object (or a collection of objects) that the laboratory considers to be a single discrete, uniquely identified unit that is the subject of one or more steps in the laboratory diagnostic workflow. Sampling can create new child specimens (tissue “dice(s)” or sections). These are full specimens in their own right (they have unique identifiers and are direct subjects in one or more steps in the laboratory’s diagnostic workflow). The potential for child specimens, which can be created from existing specimens by sampling, extends the common definition of the word specimen, which is usually limited to the original object received for examination (eg, a part from surgery).
Specimen containers play an important role in laboratory diagnostic processes. In most, but not all, process steps, specimens are held in containers, and the container often carries its specimen’s identification (ID). Sometimes the container becomes intimately involved with the specimen (eg, a paraffin block), and in some situations (such as examining tissue under the microscope) the container (the slide and cover slip) becomes part of the optical path.
Containers have identifiers that are important in laboratory operations and in some imaging processes (such as whole-slide imaging). The DICOM specimen module distinguishes the container ID from the specimen ID, making them different data elements. In many laboratories where there is one specimen per container, the value of the specimen ID and container ID will be same (Figure 2). However, there are use cases in which there is more than one specimen in a container. In this case, the value of the container ID and the specimen ID will be different — as in the case of a tissue microarray (TMA) (see Figure 3).
TMA is a collection of specimens embedded in the same paraffin block. Typically, a TMA block is created by taking a small core of tissue from many different tissue blocks and re-embedding them in a new block. Slides created from this TMA block therefore have small fragments of many different tissues, all of which may be processed at the same time, under the same conditions, using the chosen technique. Typically, these are used in research (Figure 3).
Quality assessment studies in anatomic pathology show that each of the different steps from specimen collection and processing to report editing may be a source of errors and that information systems integration supports error reduction.(11–13) This work done within the framework of the IHE project consisted of defining the needs of systems integration in anatomic pathology, especially between imaging systems and traditional pathology information systems.
The result is a new integration profile “pathology workflow” proposed for a first IHE cycle in anatomic pathology and described in the Pathology Technical Framework (volume 1 and 2).(14) This integration profile addresses the specificity of the anatomic pathology workflow in relation to laboratory and radiology workflows and defines the structure and content of cases, orders, image folders and reports. There was an issue about how to make explicit the links between information entities (eg, orders, image folder, reports) and real-world objects (eg, specimens and containers). HL7 and DICOM-based transactions were proposed to implement the integration profile.
The anatomic pathology specimen model proposed within the DICOM supplement 122 was found to be consistent with the Specimen Domain Messaging Information Model (Specimen D-MIM) currently being balloted as part of the HL7 reference information model (RIM). The Specimen D-MIM was designed for clinical laboratory specimens but is sufficiently generic to be useful for anatomic pathology specimens.
Although the main output of the anatomic pathology workflow is a timely and clear report of a diagnostic opinion, images will increasingly be treated as evidence to textual reports. DICOM is a convenient format for image archiving and communication within the anatomic pathology department. HL7 CDA seems to be the most suitable format for integration to the electronic healthcare record, but solutions available to link textual items to DICOM images in an HL7 CDA document must be clarified. Our view is that software manufacturers will implement the IHE “pathology workflow” integration profile and make integration available in daily practice.
- HL7. www.hl7.org
- DICOM 3. www.nema.org
- IHE. http://ihe. univ-rennes1.fr/whatisihe
- Klossa J, Cordier JC, Flandrin G, Got C, Hémet J. A European de facto standard for image folders applied to telepathology and teaching. Int J Med Inform 1998;48:207-16.
- NEMA. DICOM strategic document, version 5.6, Oct 24, 2006. p. 36-8. http://medical.nema.org/DICOM
- Burnett L, Chesher D, Mudaliar Y. Improving the quality of information on pathology request forms. Ann Clin Biochem 2004;41(Pt 1):53-6.
- Association of Directors of Anatomic and Surgical Pathology. Consultations in surgical pathology. Am J Surg Pathol 1993;17:743-5.
- Campbell F, Griffiths DF. Quantitative audit of the content of histopathology reports. J Clin Pathol 1994;47:360-1.
- Fleming KA. Evidence-based pathology. J Pathol 1996;179(2):127-8.
- Fleming KA. Evidence-based cellular pathology. Lancet 2002;359(9312):1149-50.
- Cowan DF. Quality assurance in anatomic pathology. An information system approach. Arch Pathol Lab Med 1990;114(2):129-34.
- Goldstein NS. Diagnostic errors in surgical pathology. Clin Lab Med 1999;19(4):43-56.
- Zarbo RJ, Meier FA, Raab SS. Error detection in anatomic pathology. Arch Pathol Lab Med 2005;129:1237-45.
- IHE Pathology. www.ihe.net