Contouring in transition: perceptions of AI-based autocontouring by radiation oncologists and medical physicists in German-speaking countries
Vorbach SM, Putz F, Ganswindt U, Janssen S, Grohmann M, Knippen S, Heinemann F, Shafie RA, Peeken JC (2025)
Publication Type: Journal article
Publication year: 2025
Journal
DOI: 10.1007/s00066-025-02403-1
Abstract
Background: Artificial intelligence (AI)-based autocontouring software has the potential to revolutionize radiotherapy planning. In recent years, several AI-based autocontouring solutions with many advantages have emerged; however, their clinical use raises several challenges related to implementation, quality assurance, validation, and training. The aim of this study was to investigate the current use of AI-based autocontouring software and the associated expectations and hopes of radiation oncologists and medical physicists in German-speaking countries. Methods: A digital survey consisting of 24 questions including single-choice, multiple-choice, free-response, and five-point Likert scale rankings was conducted using the online tool umfrageonline.com (enuvo GmbH, Pfäffikon SZ, Switzerland). Results: A total of 163 participants completed the survey, with approximately two thirds reporting use of AI-based autocontouring software in routine clinical practice. Of the users, 92% found the software helpful in clinical practice. More than 90% reported using AI solutions to contour organs at risk (OARs) in the brain, head and neck, thorax, abdomen, and pelvis. The majority (88.8%) reported time savings in OAR delineation, with approximately 41% estimating savings of 11–20 min per case. However, nearly half of the respondents expressed concern about the potential degradation of resident training in sectional anatomy understanding. Of respondents, 60% would welcome guidelines for implementation and use of AI-based contouring aids from their respective radiation oncology societies. Respondents’ free-text comments emphasized the need for careful monitoring and postprocessing of AI-delivered autocontours as well as concerns about overreliance on AI and its impact on the development of young physicians’ contouring and planning skills. Conclusion: Artificial intelligence-based autocontouring software shows promise for integration into radiation oncology workflows, with respondents recognizing its potential for time saving and standardization. However, successful implementation will require ongoing education and curriculum adaptation to ensure AI enhances, rather than replaces, clinical expertise.
Authors with CRIS profile
Involved external institutions
Helios Hospital Schwerin
Germany (DE)
Universitätsklinikum Freiburg
Germany (DE)
Universitätsklinikum Göttingen
Germany (DE)
Klinikum rechts der Isar
Germany (DE)
Medizinische Universität Innsbruck
Austria (AT)
Universität zu Lübeck
Germany (DE)
Universitätsklinikum Hamburg-Eppendorf (UKE)
Germany (DE)
Germany (DE)
Universitätsklinikum Freiburg
Germany (DE)
Universitätsklinikum Göttingen
Germany (DE)
Klinikum rechts der Isar
Germany (DE)
Medizinische Universität Innsbruck
Austria (AT)
Universität zu Lübeck
Germany (DE)
Universitätsklinikum Hamburg-Eppendorf (UKE)
Germany (DE)
How to cite
APA:
Vorbach, S.M., Putz, F., Ganswindt, U., Janssen, S., Grohmann, M., Knippen, S.,... Peeken, J.C. (2025). Contouring in transition: perceptions of AI-based autocontouring by radiation oncologists and medical physicists in German-speaking countries. Strahlentherapie und Onkologie. https://doi.org/10.1007/s00066-025-02403-1
MLA:
Vorbach, Samuel M., et al. "Contouring in transition: perceptions of AI-based autocontouring by radiation oncologists and medical physicists in German-speaking countries." Strahlentherapie und Onkologie (2025).
BibTeX: Download