Differences in radiotherapy delivery and outcome due to contouring variation

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

Gross tumor volume (GTV) delineation is central for radiotherapy planning. It provides the basis of the clinical target volume and, ultimately, the planning target volume which is used for dose optimization. Manual GTV delineations are prone to intra- and inter-observer variation and automatic segmentation methods also produce different results. There is no consensus on how to account for the contouring uncertainty, but has been suggested to incorporate it into the planning target volume (PTV) margin. Current recipes for the PTV margin are based on normal distribution assumptions and are more suitable for setup and execution errors. In this study we use the GTV delineations made by 6 experienced clinicians to create delineation-specific dose plans. These dose plans are then used to calculate theoretic tumor control probabilities (TCP) differences between delineations. The results show that current margin recipes are inadequate for maintaining the same TCP despite manual delineation variation. New methods to account for delineation variation should be developed.

Original languageEnglish
Title of host publicationClinical Image-Based Procedures : From Planning to Intervention - International Workshop, CLIP 2012, Held in Conjunction with MICCAI 2012, Revised Selected Papers
Number of pages8
PublisherSpringer Verlag
Publication date2013
Pages122-129
ISBN (Print)9783642380785
DOIs
Publication statusPublished - 2013
EventInternational Workshop on Clinical Image-Based Procedures: From Planning to Intervention, CLIP 2012 - Held in Conjunction with MICCAI 2012 - Nice, France
Duration: 5 Oct 20125 Oct 2012

Conference

ConferenceInternational Workshop on Clinical Image-Based Procedures: From Planning to Intervention, CLIP 2012 - Held in Conjunction with MICCAI 2012
LandFrance
ByNice
Periode05/10/201205/10/2012
Sponsorexocad GmbH, MedCom GmbH
SeriesLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume7761 LNCS
ISSN0302-9743

    Research areas

  • Contouring observer variation, Dose planning, Lung tumor, Radiotherapy, Tumor control probability

ID: 357363967