Objective assessment of image quality VI: Imaging in radiation therapy

Harrison H Barrett, Matthew A Kupinski, Stefan Müeller, Howard J. Halpern, John C. Morris, Roisin Dwyer

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

Earlier work on objective assessment of image quality (OAIQ) focused largely on estimation or classification tasks in which the desired outcome of imaging is accurate diagnosis. This paper develops a general framework for assessing imaging quality on the basis of therapeutic outcomes rather than diagnostic performance. By analogy to receiver operating characteristic (ROC) curves and their variants as used in diagnostic OAIQ, the method proposed here utilizes the therapy operating characteristic or TOC curves, which are plots of the probability of tumor control versus the probability of normal-tissue complications as the overall dose level of a radiotherapy treatment is varied. The proposed figure of merit is the area under the TOC curve, denoted AUTOC. This paper reviews an earlier exposition of the theory of TOC and AUTOC, which was specific to the assessment of image-segmentation algorithms, and extends it to other applications of imaging in external-beam radiation treatment as well as in treatment with internal radioactive sources. For each application, a methodology for computing the TOC is presented. A key difference between ROC and TOC is that the latter can be defined for a single patient rather than a population of patients.

Original languageEnglish (US)
Pages (from-to)8197-8213
Number of pages17
JournalPhysics in Medicine and Biology
Volume58
Issue number22
DOIs
StatePublished - Nov 21 2013

Fingerprint

Radiotherapy
ROC Curve
Computing Methodologies
Therapeutics
Area Under Curve
Radiation
Population
Neoplasms

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging
  • Radiological and Ultrasound Technology

Cite this

Objective assessment of image quality VI : Imaging in radiation therapy. / Barrett, Harrison H; Kupinski, Matthew A; Müeller, Stefan; Halpern, Howard J.; Morris, John C.; Dwyer, Roisin.

In: Physics in Medicine and Biology, Vol. 58, No. 22, 21.11.2013, p. 8197-8213.

Research output: Contribution to journalArticle

Barrett, Harrison H ; Kupinski, Matthew A ; Müeller, Stefan ; Halpern, Howard J. ; Morris, John C. ; Dwyer, Roisin. / Objective assessment of image quality VI : Imaging in radiation therapy. In: Physics in Medicine and Biology. 2013 ; Vol. 58, No. 22. pp. 8197-8213.
@article{35d7e5a9b6e74d058873d37e21427f01,
title = "Objective assessment of image quality VI: Imaging in radiation therapy",
abstract = "Earlier work on objective assessment of image quality (OAIQ) focused largely on estimation or classification tasks in which the desired outcome of imaging is accurate diagnosis. This paper develops a general framework for assessing imaging quality on the basis of therapeutic outcomes rather than diagnostic performance. By analogy to receiver operating characteristic (ROC) curves and their variants as used in diagnostic OAIQ, the method proposed here utilizes the therapy operating characteristic or TOC curves, which are plots of the probability of tumor control versus the probability of normal-tissue complications as the overall dose level of a radiotherapy treatment is varied. The proposed figure of merit is the area under the TOC curve, denoted AUTOC. This paper reviews an earlier exposition of the theory of TOC and AUTOC, which was specific to the assessment of image-segmentation algorithms, and extends it to other applications of imaging in external-beam radiation treatment as well as in treatment with internal radioactive sources. For each application, a methodology for computing the TOC is presented. A key difference between ROC and TOC is that the latter can be defined for a single patient rather than a population of patients.",
author = "Barrett, {Harrison H} and Kupinski, {Matthew A} and Stefan M{\"u}eller and Halpern, {Howard J.} and Morris, {John C.} and Roisin Dwyer",
year = "2013",
month = "11",
day = "21",
doi = "10.1088/0031-9155/58/22/8197",
language = "English (US)",
volume = "58",
pages = "8197--8213",
journal = "Physics in Medicine and Biology",
issn = "0031-9155",
publisher = "IOP Publishing Ltd.",
number = "22",

}

TY - JOUR

T1 - Objective assessment of image quality VI

T2 - Imaging in radiation therapy

AU - Barrett, Harrison H

AU - Kupinski, Matthew A

AU - Müeller, Stefan

AU - Halpern, Howard J.

AU - Morris, John C.

AU - Dwyer, Roisin

PY - 2013/11/21

Y1 - 2013/11/21

N2 - Earlier work on objective assessment of image quality (OAIQ) focused largely on estimation or classification tasks in which the desired outcome of imaging is accurate diagnosis. This paper develops a general framework for assessing imaging quality on the basis of therapeutic outcomes rather than diagnostic performance. By analogy to receiver operating characteristic (ROC) curves and their variants as used in diagnostic OAIQ, the method proposed here utilizes the therapy operating characteristic or TOC curves, which are plots of the probability of tumor control versus the probability of normal-tissue complications as the overall dose level of a radiotherapy treatment is varied. The proposed figure of merit is the area under the TOC curve, denoted AUTOC. This paper reviews an earlier exposition of the theory of TOC and AUTOC, which was specific to the assessment of image-segmentation algorithms, and extends it to other applications of imaging in external-beam radiation treatment as well as in treatment with internal radioactive sources. For each application, a methodology for computing the TOC is presented. A key difference between ROC and TOC is that the latter can be defined for a single patient rather than a population of patients.

AB - Earlier work on objective assessment of image quality (OAIQ) focused largely on estimation or classification tasks in which the desired outcome of imaging is accurate diagnosis. This paper develops a general framework for assessing imaging quality on the basis of therapeutic outcomes rather than diagnostic performance. By analogy to receiver operating characteristic (ROC) curves and their variants as used in diagnostic OAIQ, the method proposed here utilizes the therapy operating characteristic or TOC curves, which are plots of the probability of tumor control versus the probability of normal-tissue complications as the overall dose level of a radiotherapy treatment is varied. The proposed figure of merit is the area under the TOC curve, denoted AUTOC. This paper reviews an earlier exposition of the theory of TOC and AUTOC, which was specific to the assessment of image-segmentation algorithms, and extends it to other applications of imaging in external-beam radiation treatment as well as in treatment with internal radioactive sources. For each application, a methodology for computing the TOC is presented. A key difference between ROC and TOC is that the latter can be defined for a single patient rather than a population of patients.

UR - http://www.scopus.com/inward/record.url?scp=84896501804&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84896501804&partnerID=8YFLogxK

U2 - 10.1088/0031-9155/58/22/8197

DO - 10.1088/0031-9155/58/22/8197

M3 - Article

C2 - 24200954

AN - SCOPUS:84896501804

VL - 58

SP - 8197

EP - 8213

JO - Physics in Medicine and Biology

JF - Physics in Medicine and Biology

SN - 0031-9155

IS - 22

ER -