Quantitative imaging: Correlating image features with the segmentation accuracy of PET based tumor contours in the lung

Perry Johnson, Lori A. Young, Narottam Lamichhane, Vivek Patel, Felix M. Chinea, Fei Yang

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

The purpose of this study was to investigate the correlation between image features extracted from PET images and the accuracy of manually drawn lesion contours in the lung. Such correlations are interesting in that they could potentially be used in predictive models to help guide physician contouring. In this work, 26 synthetic PET datasets were created using an anthropomorphic phantom and Monte Carlo simulation. Manual contours of simulated lesions were provided by 10 physicians. Contour accuracy was quantified using five commonly used similarity metrics which were then correlated with several features extracted from the images. Features were sub-divided into three groups using intensity, geometry, and texture as categorical descriptors. When averaged among the participants, the results showed relatively strong correlations with complexity and contrastI (r ≥ 0.65, p < 0.001), and moderate correlations with several other image features (r ≥ 0.5, p < 0.01). The predictive nature of these correlations was improved through stepwise regression and the creation of multi-feature models. Imaging features were also correlated with the standard deviation of contouring error in order to investigate inter-observer variability. Several features were consistently identified as influential including integral of mean curvature and complexity. These relationships further the understanding as to what causes variation in the contouring of PET positive lesions.

Original languageEnglish (US)
Pages (from-to)257-262
Number of pages6
JournalRadiotherapy and Oncology
Volume123
Issue number2
DOIs
StatePublished - May 1 2017

Fingerprint

Physicians
Lung
Observer Variation
Neoplasms
Datasets

Keywords

  • Image texture
  • Positron emission tomography
  • Quantitative imaging
  • Target delineation

ASJC Scopus subject areas

  • Hematology
  • Oncology
  • Radiology Nuclear Medicine and imaging

Cite this

Quantitative imaging : Correlating image features with the segmentation accuracy of PET based tumor contours in the lung. / Johnson, Perry; Young, Lori A.; Lamichhane, Narottam; Patel, Vivek; Chinea, Felix M.; Yang, Fei.

In: Radiotherapy and Oncology, Vol. 123, No. 2, 01.05.2017, p. 257-262.

Research output: Contribution to journalArticle

Johnson, Perry ; Young, Lori A. ; Lamichhane, Narottam ; Patel, Vivek ; Chinea, Felix M. ; Yang, Fei. / Quantitative imaging : Correlating image features with the segmentation accuracy of PET based tumor contours in the lung. In: Radiotherapy and Oncology. 2017 ; Vol. 123, No. 2. pp. 257-262.
@article{451928b8a63a4ba3b4d548ecc18aa909,
title = "Quantitative imaging: Correlating image features with the segmentation accuracy of PET based tumor contours in the lung",
abstract = "The purpose of this study was to investigate the correlation between image features extracted from PET images and the accuracy of manually drawn lesion contours in the lung. Such correlations are interesting in that they could potentially be used in predictive models to help guide physician contouring. In this work, 26 synthetic PET datasets were created using an anthropomorphic phantom and Monte Carlo simulation. Manual contours of simulated lesions were provided by 10 physicians. Contour accuracy was quantified using five commonly used similarity metrics which were then correlated with several features extracted from the images. Features were sub-divided into three groups using intensity, geometry, and texture as categorical descriptors. When averaged among the participants, the results showed relatively strong correlations with complexity and contrastI (r ≥ 0.65, p < 0.001), and moderate correlations with several other image features (r ≥ 0.5, p < 0.01). The predictive nature of these correlations was improved through stepwise regression and the creation of multi-feature models. Imaging features were also correlated with the standard deviation of contouring error in order to investigate inter-observer variability. Several features were consistently identified as influential including integral of mean curvature and complexity. These relationships further the understanding as to what causes variation in the contouring of PET positive lesions.",
keywords = "Image texture, Positron emission tomography, Quantitative imaging, Target delineation",
author = "Perry Johnson and Young, {Lori A.} and Narottam Lamichhane and Vivek Patel and Chinea, {Felix M.} and Fei Yang",
year = "2017",
month = "5",
day = "1",
doi = "10.1016/j.radonc.2017.03.008",
language = "English (US)",
volume = "123",
pages = "257--262",
journal = "Radiotherapy and Oncology",
issn = "0167-8140",
publisher = "Elsevier Ireland Ltd",
number = "2",

}

TY - JOUR

T1 - Quantitative imaging

T2 - Correlating image features with the segmentation accuracy of PET based tumor contours in the lung

AU - Johnson, Perry

AU - Young, Lori A.

AU - Lamichhane, Narottam

AU - Patel, Vivek

AU - Chinea, Felix M.

AU - Yang, Fei

PY - 2017/5/1

Y1 - 2017/5/1

N2 - The purpose of this study was to investigate the correlation between image features extracted from PET images and the accuracy of manually drawn lesion contours in the lung. Such correlations are interesting in that they could potentially be used in predictive models to help guide physician contouring. In this work, 26 synthetic PET datasets were created using an anthropomorphic phantom and Monte Carlo simulation. Manual contours of simulated lesions were provided by 10 physicians. Contour accuracy was quantified using five commonly used similarity metrics which were then correlated with several features extracted from the images. Features were sub-divided into three groups using intensity, geometry, and texture as categorical descriptors. When averaged among the participants, the results showed relatively strong correlations with complexity and contrastI (r ≥ 0.65, p < 0.001), and moderate correlations with several other image features (r ≥ 0.5, p < 0.01). The predictive nature of these correlations was improved through stepwise regression and the creation of multi-feature models. Imaging features were also correlated with the standard deviation of contouring error in order to investigate inter-observer variability. Several features were consistently identified as influential including integral of mean curvature and complexity. These relationships further the understanding as to what causes variation in the contouring of PET positive lesions.

AB - The purpose of this study was to investigate the correlation between image features extracted from PET images and the accuracy of manually drawn lesion contours in the lung. Such correlations are interesting in that they could potentially be used in predictive models to help guide physician contouring. In this work, 26 synthetic PET datasets were created using an anthropomorphic phantom and Monte Carlo simulation. Manual contours of simulated lesions were provided by 10 physicians. Contour accuracy was quantified using five commonly used similarity metrics which were then correlated with several features extracted from the images. Features were sub-divided into three groups using intensity, geometry, and texture as categorical descriptors. When averaged among the participants, the results showed relatively strong correlations with complexity and contrastI (r ≥ 0.65, p < 0.001), and moderate correlations with several other image features (r ≥ 0.5, p < 0.01). The predictive nature of these correlations was improved through stepwise regression and the creation of multi-feature models. Imaging features were also correlated with the standard deviation of contouring error in order to investigate inter-observer variability. Several features were consistently identified as influential including integral of mean curvature and complexity. These relationships further the understanding as to what causes variation in the contouring of PET positive lesions.

KW - Image texture

KW - Positron emission tomography

KW - Quantitative imaging

KW - Target delineation

UR - http://www.scopus.com/inward/record.url?scp=85018539960&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85018539960&partnerID=8YFLogxK

U2 - 10.1016/j.radonc.2017.03.008

DO - 10.1016/j.radonc.2017.03.008

M3 - Article

C2 - 28433412

AN - SCOPUS:85018539960

VL - 123

SP - 257

EP - 262

JO - Radiotherapy and Oncology

JF - Radiotherapy and Oncology

SN - 0167-8140

IS - 2

ER -