gms | German Medical Science

After very complete coverage of radiological image data had been achieved through establishing a hospital wide PACS (see [Ref. 1] and [Ref. 2]), a project for a unique comprehensive innovative archiving and distribution system for various kinds of medical image data and other types of large volume medical datasets (like bio-signals, movies, etc…) was started in the year 2000 - the Advanced Image Management (AIM). The main goal is to establish the multimedia electronic patient record through a tight integration of the AIM with the Clinical Information System based on Cerner Millennium (see [Ref. 3]). Establishing a central highly available, secure (data confidentiality) and highly performing archive with 90TB of RAID disk space for everything online access to image data from a unified viewing software distributed to 2.400 clinical workstations was the objective reached during the 3 year project.

Building on a high performance cluster of Linux based servers with 90TB RAID5 storage, the AIM as a central archive for DICOM and non-DICOM data archived either directly from imaging modalities (AIM as first-line archive) or from departmental imaging archives (AIM as second-line archive) has been established. Through an elaborate scheme of data access privileges, data privacy in a multi institutional setting is maintained. Legacy data (approx. 40TB) from existing PACS systems is migrated to the AIM. Workloads of up to 240GB of new data (corresponding to 320.000 image files) per day have to be coped with - probably the largest installation worldwide. Sophisticated interfacing to departmental PACS's to automatically keep patient and image data in sync when data administration is performed in the first-line archive ("managed data channel") have been developed. Through optimum tuning of server hardware and client server protocol for image data distribution, transfer rates of some 35 CT images per second are reached, allowing to easily handle even multislice CT exams. Data from the somnology labs (bio-signals and videos) is archived via AIM, with some 30GB per day. The archiving and distribution of PDF documents (eg. of ECGs) has been a major success.

[Abb. 1]

In contrast to similar approaches (see [Ref. 4]), we have implemented a multimodality archive also for non-DICOM datasets and a common viewing software which provides one common interface for image data access to all clinical departments, giving great organizational and financial benefits. The image data platform established with AIM will be constantly improved and expanded in the years to come, through the constant addition of new digital imaging systems throughout the hospital. It also serves as the basic infrastructure for a web portal system designed for secure report document and image distribution to external institutions and for data exchange in international scientific collaborations. Plans to utilized the AIM infrastructure as a comprehensive archiving and backup system for the CIS are being discussed and open new perspectives for the positioning of the AIM.