Evidence and Diagnostic Reporting in the IHE Context 1

Transcription

1 Computer-Assisted Radiology and Surgery Technical Report Evidence and Diagnostic Reporting in the IHE Context 1 Cor Loef, MsEE, Roel Truyen, MsEE Capturing clinical observations and findings during the diagnostic imaging process is increasingly becoming a critical step in diagnostic reporting. Standards developers notably HL7 and DICOM are making significant progress toward standards that enable exchanging clinical observations and findings among the various information systems of the healthcare enterprise. DICOM like the HL7 Clinical Document Architecture (CDA) uses templates and constrained, coded vocabulary (SNOMED, LOINC, etc.). Such a representation facilitates automated software recognition of findings and observations, intrapatient comparison, correlation to norms, and outcomes research. The scope of DICOM Structured Reporting (SR) includes many findings that products routinely create in digital form (measurements, computed estimates, etc.). In the Integrating the Healthcare Enterprise (IHE) framework, two Integration Profiles are defined for clinical data capture and diagnostic reporting: Evidence Document, and Simple Image and Numeric Report. This report describes these two DICOM SR based integration profiles in the diagnostic reporting process. Key Words. DICOM structured reporting; IHE; evidence document; simple image and numeric report. AUR, 2005 Capturing clinical data during the diagnostic imaging process is becoming essential for diagnostic reporting. Obstacles have been converting data into a digital coded format and the absence of standards. In the diagnostic imaging process, digital capture of findings has been routine for many years in cardiology, radiology, vascular, and obstetric/gynecologic specialties. While DICOM has achieved great success with images, findings interoperability has struggled with proprietary solutions. Fortunately, DICOM SR is removing this obstacle. In the meantime, the scope of digitally generated findings is increasing. New technology such as computer-aided diagnosis (CAD) and parametric imaging will increase the volume of clinical digital findings. Acad Radiol 2005; 12: From Philips Medical Systems, Veenpluis 4-6, Building QV-110, PO Box , 5680 DA Best, the Netherlands. Received December 22, 2004; revision requested January 14, 2005; revision received January 26; accepted January 27. Address correspondence to C.L. cor.loef@philips.com AUR, 2005 doi: /j.acra A key interoperability criterion is data comparability. Clinical software applications need to recognize evidence and findings and sort, collate, compare, and match discrete clinical findings to norms. Products that feature data comparability will enable radiologists to improve the productivity and quality of their performed tasks. CLINICAL APPLICATION: COMPUTER-AIDED DIAGNOSIS FOR PULMONARY NODULE An emerging clinical application that would benefit from structured and standardized ways to store and distribute this evidence is CAD. As an example, we can consider the Pulmonary Nodule CAD application (1), where the purpose is to assist the radiologist in locating pulmonary nodules that can cause cancer. Currently, a radiologist scans the CT volume manually to locate the nodules. However, this approach is time consuming, and there is no guarantee that all of the nod- 620

2 Academic Radiology, Vol 12, No 5, May 2005 EVIDENCE AND DIAGNOSTIC REPORTING IN IHE CONTEXT Figure 1. Lung nodule detection paradigm. ules are properly detected. A CAD method usually takes less time than manual detection but still does not have 100% reliability. To increase the efficiency and reliability of the results, a combination of the two methods is used. With this two-pass approach, the volume is first scanned manually and afterwards an automatic detection is performed and compared (Fig. 1). The result of each examination is a list of nodules. The nodules are characterized with their dimensions, volume, structure, position, etc. Besides the numerical data, images can be generated to visualize nodules (Fig. 2). Findings from the CAD system and the radiologist have to be compared, even if they were generated on different systems. This illustrates the need for a structured and standardized way of storing these findings and their evidence. Figure 3 shows the proposed structure of a typical report for the lung assessment application. The report is a tree of items with cross-references. Each item can be an image or a container consisting of one or more items. Some of the containers can hold heterogeneous items (e.g., Nodule or Report in Fig. 3), whereas other containers can contain a list of similar items (e.g., Nodule list in Fig. 3). DICOM STRUCTURED REPORTING (SR) DICOM SR (2) uses hierarchically organized coded vocabularies to express clinical data such as the report of the lung nodule assessment. SR templates constrain vocabulary and its hierarchical organization. Judiciously chosen atomic concepts can express a much larger scope of more complex, composite concepts such as clinical findings (often called postcoordination). The template specifies the concept codes and how they are structurally organized. Software requires a precise recipe to compose and recognize findings. Templates provide these recipes and specify the essential information required to assemble the finding in the report document. The cells in the template are directly tied to the clinical application. The report consumer must reassemble the finding from the cells in the template. DICOM SR templates facilitate expressing much more that just the finding: whether the value is derived as an average; links of the image-to-image cursors, or to other findings; and other context information that may influence clinical judgment. These relationships may point to other content items in the SR content tree. IHE INTEGRATION PROFILE: EVIDENCE DOCUMENT The Evidence Documents Profile (3) defines interoperable ways for observations, measurements, results, and other procedure details recorded in the course of carrying out a procedure step to be output by devices, such as acquisition systems and other workstations; to be stored and managed by archival systems; and to be retrieved and presented or used by display and reporting systems. This allows detailed nonimage information, such as measurements, CAD results, procedure logs, etc., to be made available as input to the process of generating a diagnostic report either as additional evidence for the reporting physician or, in some cases, for selected items in the Evidence Document to be included in the diagnostic report. The IHE Scheduled Workflow Profile and/or the IHE Post Processing Workflow Profile manage the production of Evidence Documents. As such, the Evidence Document Profile is required to be supported together with one or the other of those two profiles. Evidence Documents represent one of the inputs to the reporting process and may provide details that get included in diagnostic reports described in the Simple Image and Numeric Report Profile. Actors and Transactions Figure 4 shows the actors directly involved in the Evidence Documents Integration Profile and the relevant transactions between them. Process Flow As with other Evidence Objects, Evidence Documents are usually created by the system operator and used by the reading physician in the process of creat- 621

3 LOEF AND TRUYEN Academic Radiology, Vol 12, No 5, May 2005 Figure 2. Different views of a nodule. Figure 3. Proposed structure for a report of the lung assessment application. ing a Diagnostic Report, either by reviewing or interpreting the Evidence Document contents or by copying selected parts into the Report. Evidence Documents represent the uninterpreted information that is primarily managed and used inside an imaging department, although distribution outside the imaging department might be of interest as well. In contrast, the diagnostic reports described in the Simple Image and Numeric Reports Profile represent the interpreted information that is the primary output of the imaging department and are available for wide distribution. Figure 4. Evidence Document Actor Diagram (3). Evidence Report Content and Structure Evidence Documents belong to the family of Evidence Objects that also includes Images, Presentation States, and Key Image Notes. These are objects generated as a result of performing procedure steps on systems in a clinical department. The DICOM SOP classes are Basic Text, Enhanced SR, Comprehensive SR, Chest CAD SR, and Mammography CAD SR. The key characteristic of these reports is that coded numeric measurements are included (Fig. 5). A number of DICOM Supplements are currently under development: OB-GYN Ultrasound Procedure Reports, 622

4 Academic Radiology, Vol 12, No 5, May 2005 EVIDENCE AND DIAGNOSTIC REPORTING IN IHE CONTEXT Figure 5. Example of Evidence Report structure with Numeric Measurements (3). Table 1 Evidence Document Templates Template ID TID 4000 TID 5000 TID 3500 TID 4100 TID 5100 Vascular Template Name Mammography CAD Document Root Template OB-GYN Ultrasound Procedure Report Hemodynamics Report Chest CAD SR Document Root Template Ultrasound Procedure Report Template voice recognition, and specialized reporting packages, by separating the functions of reporting into discrete actors for creation, management, storage, and viewing. Separating these functions while defining transactions to exchange the reports between them enables a vendor to include one or more of these functions in an actual system. Actors and Transactions Figure 6 shows the actors directly involved in the Simple Image and Numeric Report Integration Profile and the relevant transactions between them. Chest CAD SR, Catheterization Lab SR, and Vascular Ultrasound SR. The templates used in these Evidence Documents are listed in Table 1. IHE INTEGRATION PROFILE: SIMPLE IMAGE AND NUMERIC REPORT The Simple Image and Numeric Report Integration Profile facilitates the growing use of digital dictation, Process Flow In the initial stage of a typical diagnostic reporting, a reading physician records the diagnosis by generating a draft DICOM Structured Report object, which is submitted to the report manager. once a report is sent to the report manager, the report creator relinquishes control of the report to the report manager. Reports are processed and modified by the report manager. This involves adding and changing report data as well as verifying draft reports. At any time, the report manager can transmit reports to the report repository for external access, but at a minimum the final report must be sent to the 623

5 LOEF AND TRUYEN Academic Radiology, Vol 12, No 5, May 2005 Figure 6. Simple Image and Numeric Report Actor Diagram (3). Figure 7. Simple Image Report Pattern (3). report repository. A report creator can effectively amend a report by submitting a new SR SOP Instance. The report repository provides permanent storage of DICOM Structured Reports. It also allows reports to be queried and retrieved throughout the enterprise by report readers. A report reader provides a user interface to view DICOM Structured Reports that it retrieves from the report repository or external report repository access. The External Report Repository Access is a gateway to obtain other enterprise department reports, such as laboratory and pathology, from within the Imaging department. DICOM Structured Reports are queried and retrieved by a report reader from the External Report Repository Access. The Enterprise Report Repository receives diagnostic reports in HL7 format. Simple Image and Numeric Report Content and Structure Reports exchanged have a simple structure attractive to many imaging departments: a title, an observation context, and one or more sections, each with a heading, ob- 624

6 Academic Radiology, Vol 12, No 5, May 2005 EVIDENCE AND DIAGNOSTIC REPORTING IN IHE CONTEXT servation context, text, image references, and optionally coded measurements (see Fig. 7 for a Simple Image Report). Some elements can also be coded to facilitate computer searches. Such reports can be input to the formal diagnostic report, thus avoiding reentry of information. The created reports conform to the DICOM Basic Text SR Information Object Definition (IOD). If numeric values are required in the report, then the report creator creates a report that conforms to the DICOM Enhanced SR IOD. Reports created also conform to the template TID Standards are available now to capture the clinical findings during the diagnostic imaging process and to structure and encode these data in such a way that they can be shared, automatically processed, and compared with norms. Templates have been defined by professional societies, which define the relevant information and structure for a specific clinical domain, such as OB-GYN Ultrasound Procedure Report and Chest CAD Report, and many more will emerge in the coming years. Clinical software applications based on these definitions reduce the effort for the radiologist to capture the clinical findings for the diagnostic report with high quality. The IHE initiative promotes the broad deployment in the medical industry of the standardized clinical domain templates for evidence documents and reports, which allows this information to be shared and used for research to foster the progress of evidence-based medicine. CONCLUSION REFERENCES 1. Wiemker-R, Rogalla-P, Zwartkruis-A, Blaffert-T. Computer-aided lung nodule detection on high-resolution CT data. SPIE Med Imaging 2002; 4684(pt 1 3): DICOM PS 3.3 Chapter 17: SR Document Modules. National Electrical Manufacturers Association, Rosslyn, VA IHE Radiology Technical Framework, Revision 5.5, Vol. 1: Integration Profiles. Available at Copyright ACC, HIMSS, RSNA , figures reprinted with permission. 625