Visual Computing Methods for Radiotherapy Planning | TU Wien

Information

Publication Type: PhD-Thesis
Workgroup(s)/Project(s): not specified
Date: 2023
Date (Start): 2019
Date (End): 5. September 2022
Second Supervisor: Katja Bühler
Open Access: yes
1st Reviewer: Wolfgang Birkfellner
2nd Reviewer: Bernhard Preim
Rigorosum: 14. February 2023
First Supervisor: Eduard Gröller
Pages: 180
Keywords: medicine, radiology, medical physics, oncology

Abstract

Radiotherapy (RT) is one of the major curative approaches for cancer. It is a complex andrisky treatment approach, which requires precise planning, prior to the administrationof the treatment. Visual Computing (VC) is a fundamental component of RT planning,providing solutions in all parts of the process — from imaging to delivery.VC employs elements from computer graphics and image processing to create meaningful,interactive visual representations of medical data, and it has become an influentialfield of research for many advanced applications like radiation oncology. InteractiveVC approaches represent a new opportunity to integrate knowledgeable experts andtheir cognitive abilities in exploratory processes, which cannot be conducted by solelyautomatized methods.Despite the significant technological advancements of RT over the last decades, thereare still many challenges to address. In RT planning medical doctors need to consider avariety of information sources for anatomical and functional target volume delineation.The validation and inspection of the defined target volumes and the resulting RT planis a complex task, especially in the presence of moving target areas as it is the case fortumors of the chest and the upper abdomen, for instance, caused by breathing motion.Handling RT planning and delivery-related uncertainties, especially in the presence oftumor motion, is essential to improve the efficiency of the treatment and the minimizationof side effects.This dissertation contributes to the handling of RT planning related uncertainties byproposing novel VC methods. Quantification and visualization of these types of uncer-tainties will be an essential part of the presented methods, and aims at improving the RTworkflow in terms of delineation and registration accuracy, margin definitions and theinfluence of these uncertainties onto the dosimetric outcome. The publications presentedin this thesis address key aspects of the RT treatment planning process, where humaninteraction is required, and VC has the potential to improve the treatment outcome.First, major requirements for a multi-modal visualization framework are defined andimplemented with the aim to improve motion management by including 4D imageinformation. The visualization framework was designed to provide medical doctors withthe necessary visual information to improve the accuracy of tumor target delineationsand the efficiency of RT plan evaluation.xiii Furthermore, the topic of deformable image registration (DIR) accuracy is addressed inthis thesis. DIR has the potential to improve modern RT in many aspects, includingvolume definition, treatment planning, and image-guided adaptive RT. However, mea-suring DIR accuracy is difficult without known ground truth, but necessary before theintegration in the RT workflow. Visual assessment is an important step towards clinicalacceptance. A visualization framework is proposed, which supports the exploration andthe assessment of DIR accuracy. It offers different interaction and visualization featuresfor exploration of candidate regions to simplify the process of visual assessment, andthereby improve and contribute to its adequate use in RT planning.Finally, the topic of healthy tissue sparing is addressed with a novel visualization approachto interactively explore RT plans, and identify regions of healthy tissue, which can bespared further without compromising the treatment goals defined for tumor targets. Forthis, overlap volumes of tumor targets and healthy organs are included in the RT planevaluation process, and the initial visualization framework is extended with quantitativeviews. This enables quantitative properties of the overlap volumes to be interactivelyexplored, to identify critical regions and to steer the visualization for a detailed inspectionof candidates.All approaches were evaluated in user studies covering the individual visualizations andtheir interactions regarding helpfulness, comprehensibility, intuitiveness, decision-makingand speed, and if available using ground truth data to prove their validity.

Additional Files and Images

Additional images and videos

Additional files

Dissertation

Weblinks

BibTeX

@phdthesis{schlachter-2022-vcm,
  title =      "Visual Computing Methods for Radiotherapy Planning",
  author =     "Matthias Schlachter",
  year =       "2023",
  abstract =   "Radiotherapy (RT) is one of the major curative approaches
               for cancer. It is a complex andrisky treatment approach,
               which requires precise planning, prior to the
               administrationof the treatment. Visual Computing (VC) is a
               fundamental component of RT planning,providing solutions in
               all parts of the process — from imaging to delivery.VC
               employs elements from computer graphics and image processing
               to create meaningful,interactive visual representations of
               medical data, and it has become an influentialfield of
               research for many advanced applications like radiation
               oncology. InteractiveVC approaches represent a new
               opportunity to integrate knowledgeable experts andtheir
               cognitive abilities in exploratory processes, which cannot
               be conducted by solelyautomatized methods.Despite the
               significant technological advancements of RT over the last
               decades, thereare still many challenges to address. In RT
               planning medical doctors need to consider avariety of
               information sources for anatomical and functional target
               volume delineation.The validation and inspection of the
               defined target volumes and the resulting RT planis a complex
               task, especially in the presence of moving target areas as
               it is the case fortumors of the chest and the upper abdomen,
               for instance, caused by breathing motion.Handling RT
               planning and delivery-related uncertainties, especially in
               the presence oftumor motion, is essential to improve the
               efficiency of the treatment and the minimizationof side
               effects.This dissertation contributes to the handling of RT
               planning related uncertainties byproposing novel VC methods.
               Quantification and visualization of these types of
               uncer-tainties will be an essential part of the presented
               methods, and aims at improving the RTworkflow in terms of
               delineation and registration accuracy, margin definitions
               and theinfluence of these uncertainties onto the dosimetric
               outcome. The publications presentedin this thesis address
               key aspects of the RT treatment planning process, where
               humaninteraction is required, and VC has the potential to
               improve the treatment outcome.First, major requirements for
               a multi-modal visualization framework are defined
               andimplemented with the aim to improve motion management by
               including 4D imageinformation. The visualization framework
               was designed to provide medical doctors withthe necessary
               visual information to improve the accuracy of tumor target
               delineationsand the efficiency of RT plan evaluation.xiii
               Furthermore, the topic of deformable image registration
               (DIR) accuracy is addressed inthis thesis. DIR has the
               potential to improve modern RT in many aspects,
               includingvolume definition, treatment planning, and
               image-guided adaptive RT. However, mea-suring DIR accuracy
               is difficult without known ground truth, but necessary
               before theintegration in the RT workflow. Visual assessment
               is an important step towards clinicalacceptance. A
               visualization framework is proposed, which supports the
               exploration andthe assessment of DIR accuracy. It offers
               different interaction and visualization featuresfor
               exploration of candidate regions to simplify the process of
               visual assessment, andthereby improve and contribute to its
               adequate use in RT planning.Finally, the topic of healthy
               tissue sparing is addressed with a novel visualization
               approachto interactively explore RT plans, and identify
               regions of healthy tissue, which can bespared further
               without compromising the treatment goals defined for tumor
               targets. Forthis, overlap volumes of tumor targets and
               healthy organs are included in the RT planevaluation
               process, and the initial visualization framework is extended
               with quantitativeviews. This enables quantitative properties
               of the overlap volumes to be interactivelyexplored, to
               identify critical regions and to steer the visualization for
               a detailed inspectionof candidates.All approaches were
               evaluated in user studies covering the individual
               visualizations andtheir interactions regarding helpfulness,
               comprehensibility, intuitiveness, decision-makingand speed,
               and if available using ground truth data to prove their
               validity.",
  pages =      "180",
  address =    "Favoritenstrasse 9-11/E193-02, A-1040 Vienna, Austria",
  school =     "Research Unit of Computer Graphics, Institute of Visual
               Computing and Human-Centered Technology, Faculty of
               Informatics, TU Wien ",
  keywords =   "medicine, radiology, medical physics, oncology",
  URL =        "https://www.cg.tuwien.ac.at/research/publications/2023/schlachter-2022-vcm/",
}