Room: AAPM ePoster Library
Purpose: Legacy clinical data buried into archives often become virtually lost after the evolution of technology and software, such as upgrade or switch of treatment planning system (TPS). We developed a general tool that automatically converts the legacy archives to standard DicomRT and exports to the centralized RT-PACS system, making them widely accessible to clinicians and researchers through any DICOM-compatible software.
Methods: Our first project addresses the tape archives of Pinnacle TPS, which compresses data of many patients and/or machines into a single tarball (*.tar). From untarred archive, the “plan.trial” of each plan is parsed with corresponding contours (plan.roi) and beam-dose (plan.Trial.binary.#) identified. The prescriptions, fractionation-group, and beam control-points are parsed and converted to the RTPlan modules. The “plan.roi” is parsed and converted to DICOM coordinates and saved as the RTStruct. All beam-doses are accumulated into plan-dose with normalization defined in “plan.trial” and stored as RTDose. The CT volume is converted into standard DICOM-CT. The CT, RTPlan, RTDose and RTStruct are organized into “patient/study/series/image” hierarchy with DICOM uids assigned and cross-linked. Finally, the DICOM files are automatically pushed into RT-PACS, which is accessible by any DICOM compatible software. The whole procedure, from “untar” to “DICOM-pushing”, is built into a streamlined process and runs as Windows service.
Results: The software was validated through apple-to-apple comparison of plan (image, contour, dose and beams) in the original Pinnacle TPS, Eclipse TPS and Dicoman after conversion and import. A total of 353 archives (4.20TB) between 2009 and 2016 have been automatically extracted and converted, consisting of 4395 patients with 5890 plans. It would take years of human power if done manually through the TPS.
Conclusions: ArchiveDicomizer is an essential tool to bring years of “buried” clinical data back into clinical/research life by converting them to standardized format and stored in centralized PACS system.