A predictive model for distinguishing radiation necrosis from tumour progression after gamma knife radiosurgery based on radiomic features from MR images

Zijian Zhang, Jinzhong Yang, Angela Ho, Wen Jiang, Jennifer Logan, Xin Wang, Paul D. Brown, Susan L. McGovern, Nandita Guha-Thakurta, Sherise D. Ferguson, Xenia Fave, Lifei Zhang, Dennis Mackin, Laurence E. Court, Jing Li

Research output: Contribution to journalArticle

23 Citations (Scopus)

Abstract

Objectives To develop a model using radiomic features extracted from MR images to distinguish radiation necrosis from tumour progression in brain metastases after Gamma Knife radiosurgery. Methods We retrospectively identified 87 patients with pathologically confirmed necrosis (24 lesions) or progression (73 lesions) and calculated 285 radiomic features from four MR sequences (T1, T1 post-contrast, T2, and fluid-attenuated inversion recovery) obtained at two follow-up time points per lesion per patient. Reproducibility of each feature between the two time points was calculated within each group to identify a subset of features with distinct reproducible values between two groups. Changes in radiomic features from one time point to the next (delta radiomics) were used to build a model to classify necrosis and progression lesions. Results A combination of five radiomic features from both T1 post-contrast and T2 MR images were found to be useful in distinguishing necrosis from progression lesions. Delta radiomic features with a RUSBoost ensemble classifier had an overall predictive accuracy of 73.2% and an area under the curve value of 0.73 in leave-one-out cross-validation. Conclusions Delta radiomic features extracted from MR images have potential for distinguishing radiation necrosis from tumour progression after radiosurgery for brain metastases.

Original languageEnglish (US)
Pages (from-to)2255-2263
Number of pages9
JournalEuropean Radiology
Volume28
Issue number6
DOIs
StatePublished - Nov 24 2018
Externally publishedYes

Fingerprint

Radiosurgery
Necrosis
Radiation
Neoplasms
Neoplasm Metastasis
Brain
Area Under Curve

Keywords

  • Brain metastases
  • Delta radiomic features
  • Gamma knife radiosurgery
  • MRI
  • Radiation necrosis

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging

Cite this

A predictive model for distinguishing radiation necrosis from tumour progression after gamma knife radiosurgery based on radiomic features from MR images. / Zhang, Zijian; Yang, Jinzhong; Ho, Angela; Jiang, Wen; Logan, Jennifer; Wang, Xin; Brown, Paul D.; McGovern, Susan L.; Guha-Thakurta, Nandita; Ferguson, Sherise D.; Fave, Xenia; Zhang, Lifei; Mackin, Dennis; Court, Laurence E.; Li, Jing.

In: European Radiology, Vol. 28, No. 6, 24.11.2018, p. 2255-2263.

Research output: Contribution to journalArticle

Zhang, Z, Yang, J, Ho, A, Jiang, W, Logan, J, Wang, X, Brown, PD, McGovern, SL, Guha-Thakurta, N, Ferguson, SD, Fave, X, Zhang, L, Mackin, D, Court, LE & Li, J 2018, 'A predictive model for distinguishing radiation necrosis from tumour progression after gamma knife radiosurgery based on radiomic features from MR images', European Radiology, vol. 28, no. 6, pp. 2255-2263. https://doi.org/10.1007/s00330-017-5154-8
Zhang, Zijian ; Yang, Jinzhong ; Ho, Angela ; Jiang, Wen ; Logan, Jennifer ; Wang, Xin ; Brown, Paul D. ; McGovern, Susan L. ; Guha-Thakurta, Nandita ; Ferguson, Sherise D. ; Fave, Xenia ; Zhang, Lifei ; Mackin, Dennis ; Court, Laurence E. ; Li, Jing. / A predictive model for distinguishing radiation necrosis from tumour progression after gamma knife radiosurgery based on radiomic features from MR images. In: European Radiology. 2018 ; Vol. 28, No. 6. pp. 2255-2263.
@article{e9a9866800e34cba89d7b2dc23b35a05,
title = "A predictive model for distinguishing radiation necrosis from tumour progression after gamma knife radiosurgery based on radiomic features from MR images",
abstract = "Objectives To develop a model using radiomic features extracted from MR images to distinguish radiation necrosis from tumour progression in brain metastases after Gamma Knife radiosurgery. Methods We retrospectively identified 87 patients with pathologically confirmed necrosis (24 lesions) or progression (73 lesions) and calculated 285 radiomic features from four MR sequences (T1, T1 post-contrast, T2, and fluid-attenuated inversion recovery) obtained at two follow-up time points per lesion per patient. Reproducibility of each feature between the two time points was calculated within each group to identify a subset of features with distinct reproducible values between two groups. Changes in radiomic features from one time point to the next (delta radiomics) were used to build a model to classify necrosis and progression lesions. Results A combination of five radiomic features from both T1 post-contrast and T2 MR images were found to be useful in distinguishing necrosis from progression lesions. Delta radiomic features with a RUSBoost ensemble classifier had an overall predictive accuracy of 73.2{\%} and an area under the curve value of 0.73 in leave-one-out cross-validation. Conclusions Delta radiomic features extracted from MR images have potential for distinguishing radiation necrosis from tumour progression after radiosurgery for brain metastases.",
keywords = "Brain metastases, Delta radiomic features, Gamma knife radiosurgery, MRI, Radiation necrosis",
author = "Zijian Zhang and Jinzhong Yang and Angela Ho and Wen Jiang and Jennifer Logan and Xin Wang and Brown, {Paul D.} and McGovern, {Susan L.} and Nandita Guha-Thakurta and Ferguson, {Sherise D.} and Xenia Fave and Lifei Zhang and Dennis Mackin and Court, {Laurence E.} and Jing Li",
year = "2018",
month = "11",
day = "24",
doi = "10.1007/s00330-017-5154-8",
language = "English (US)",
volume = "28",
pages = "2255--2263",
journal = "European Radiology",
issn = "0938-7994",
publisher = "Springer Verlag",
number = "6",

}

TY - JOUR

T1 - A predictive model for distinguishing radiation necrosis from tumour progression after gamma knife radiosurgery based on radiomic features from MR images

AU - Zhang, Zijian

AU - Yang, Jinzhong

AU - Ho, Angela

AU - Jiang, Wen

AU - Logan, Jennifer

AU - Wang, Xin

AU - Brown, Paul D.

AU - McGovern, Susan L.

AU - Guha-Thakurta, Nandita

AU - Ferguson, Sherise D.

AU - Fave, Xenia

AU - Zhang, Lifei

AU - Mackin, Dennis

AU - Court, Laurence E.

AU - Li, Jing

PY - 2018/11/24

Y1 - 2018/11/24

N2 - Objectives To develop a model using radiomic features extracted from MR images to distinguish radiation necrosis from tumour progression in brain metastases after Gamma Knife radiosurgery. Methods We retrospectively identified 87 patients with pathologically confirmed necrosis (24 lesions) or progression (73 lesions) and calculated 285 radiomic features from four MR sequences (T1, T1 post-contrast, T2, and fluid-attenuated inversion recovery) obtained at two follow-up time points per lesion per patient. Reproducibility of each feature between the two time points was calculated within each group to identify a subset of features with distinct reproducible values between two groups. Changes in radiomic features from one time point to the next (delta radiomics) were used to build a model to classify necrosis and progression lesions. Results A combination of five radiomic features from both T1 post-contrast and T2 MR images were found to be useful in distinguishing necrosis from progression lesions. Delta radiomic features with a RUSBoost ensemble classifier had an overall predictive accuracy of 73.2% and an area under the curve value of 0.73 in leave-one-out cross-validation. Conclusions Delta radiomic features extracted from MR images have potential for distinguishing radiation necrosis from tumour progression after radiosurgery for brain metastases.

AB - Objectives To develop a model using radiomic features extracted from MR images to distinguish radiation necrosis from tumour progression in brain metastases after Gamma Knife radiosurgery. Methods We retrospectively identified 87 patients with pathologically confirmed necrosis (24 lesions) or progression (73 lesions) and calculated 285 radiomic features from four MR sequences (T1, T1 post-contrast, T2, and fluid-attenuated inversion recovery) obtained at two follow-up time points per lesion per patient. Reproducibility of each feature between the two time points was calculated within each group to identify a subset of features with distinct reproducible values between two groups. Changes in radiomic features from one time point to the next (delta radiomics) were used to build a model to classify necrosis and progression lesions. Results A combination of five radiomic features from both T1 post-contrast and T2 MR images were found to be useful in distinguishing necrosis from progression lesions. Delta radiomic features with a RUSBoost ensemble classifier had an overall predictive accuracy of 73.2% and an area under the curve value of 0.73 in leave-one-out cross-validation. Conclusions Delta radiomic features extracted from MR images have potential for distinguishing radiation necrosis from tumour progression after radiosurgery for brain metastases.

KW - Brain metastases

KW - Delta radiomic features

KW - Gamma knife radiosurgery

KW - MRI

KW - Radiation necrosis

UR - http://www.scopus.com/inward/record.url?scp=85035128944&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85035128944&partnerID=8YFLogxK

U2 - 10.1007/s00330-017-5154-8

DO - 10.1007/s00330-017-5154-8

M3 - Article

C2 - 29178031

AN - SCOPUS:85035128944

VL - 28

SP - 2255

EP - 2263

JO - European Radiology

JF - European Radiology

SN - 0938-7994

IS - 6

ER -