ESTRO 2025 - Abstract Book

S2410

Interdisciplinary – Other

ESTRO 2025

4160

Digital Poster The deployment of ProKnow for cloud-based clinical research in radiotherapy Vasiliki A Anagnostatou 1 , Martin Knauer 2 , Tobias Winderl 1 , Michael Reiner 1 , Zakaria Bahammou 1 , Reinhard Thasler 2 , Claus Belka 1 , Stefanie Corradini 1 , Stephan Schönecker 1 1 Department of Radiation Oncology, LMU University Hospital, Munich, Germany. 2 Center for Medical Data Integration and Analysis, LMU University Hospital, Munich, Germany Background and Objectives With the expansion of cloud computing, Picture Archiving and Communications in Radiotherapy (RT-PACS) systems have become available as cloud-based solutions, not only for storage but also for viewing and analysis of data for clinical research. The aim of this work is to present a deep and easy-to-use implementation of the Elekta Proknow cloud-based RT-PACS system within our department for peer review, plan analysis, studying of metrics and the discovery of trends. The first scientific use cases are presented to highlight the ProKnow potential. Methods and tools An absolute prerequisite for uploading RT-Plans on cloud-based RT-PACS systems is the de-identification of the DICOM data prior to the upload. This is accomplished using a ProKnow Dicom Agent anonymization service, installed locally on a virtual server running on the hospital’s network. The service anonymizes and uploads any Dicom data, which reaches a “watched” folder either via network drive mapping (copy-paste) or directly from the treatment planning system (DICOM TCP/IP transfer) to a pre-defined ProKnow workspace. We can access ProKnow not only through the user interface, but also through Python scripts to extract information from the uploaded data, such as RT-structures or DVH curves, to calculate and store metrics as well as to upload clinical data. Results/Scientific use cases ProKnow was used for a retrospective feasibility study of an isotoxic dose-escalated radiotherapy concept for glioblastoma. Furthermore, to ensure protocol-compliant irradiation planning for preparation of a prospective dose-escalation trial, we conducted a dummy run with 10 collaborating institutes in Germany. Whenever a plan was uploaded, a script “watching” the ProKnow workspace was calculating the Equivalent Uniform Dose (EUD) by accessing the DVH Curve. In addition, RT-structures were automatically downloaded and the Dice Score and Hausdorff Distance were calculated. The results were set as custom metrics in ProKnow. Conclusion/Outlook Major drawback of the currently implemented de-identification process for DICOM files is that retrospective cloud based analysis is limited to data anonymized at a certain fixed time. Derived data/analysis results can neither be correlated with local patient records, nor is it possible to subsequently add clinical data for further analyses. We therefore collaborate with the MeDIC LMU (Data Integration Center) of the LMU University Hospital for development and implementation of an automated de-identification and pseudonymization process via a trusted third party service. Then it will be possible at any time to merge clinical data in local DIZ databases with de-identified data.

Keywords: Research, Clinical Data, Peer Review

4219

Digital Poster Harmonizing oncology data: a pan-Canadian strategy for radiotherapy data standards

Amanda Caissie 1,2,3 , Monique Ashe 4 , Jean-Pierre Bissonnette 5,3,6 , Erika Brown 7 , Renata Chmielewski 6 , Carol-Anne Davis 1 , Caitlin Gillan 2,8 , Eric Gutierrez 6 , Nareesa Ishmail 6 , Kristi MacKenzie 3,9 , Brian Liszewski 2,6,3 , Michelle Nielsen 2,3,5 , Jason Pantarotto 6,10 , Teri Stuckless 3,11 , Kathleen Surry 7,12