Skip to main navigation Skip to search Skip to main content

QA for Auto-Contouring: Technology, Bias, and “The Human in the Loop”

Activity: Talk, presentation or poster › Scientific talk or presentation › Online

Description

Auto-contouring is one of the most dynamic areas of application for AI-based systems in radiation oncology. However, as these systems are rapidly integrated into clinical practice, the requirements for quality assurance, validation, and evaluation of the quality of their results are also increasing. Uniform, widely established standards are still under development. The 2024 guidance document “Auto-contouring in radiotherapy” published by the Royal College of Radiologists addresses these challenges comprehensively.

In addition to technical aspects of quality assurance, the document places particular emphasis on the human factor: What impact do perceptual effects and cognitive biases have on the assessment of AI-generated contours? How can quality assurance take into account the clinical review process as well as the technical performance of the algorithms?

Current concepts of QA for auto-contouring, with particular emphasis on the human factor and clinical decision-making, are discussed.
Period23 Jun 2026
Event titleAutomatisierte Konturierung in der Strahlentherapie:
KI-basierte Tools, Workflows und Qualitätssicherung: Österreichische Gesellschaft für Medizinische Physik , Arbeitskreis Strahlentherapie
Event typeWorkshop
Degree of RecognitionNational