A Multi-Institutional Comparison of Dynamic Contrast-Enhanced Magnetic Resonance Imaging Parameter Calculations

Rachel B. Ger, Abdallah S.R. Mohamed, Musaddiq J. Awan, Yao Ding, Kimberly Li, Xenia J. Fave, Andrew L. Beers, Brandon Driscoll, Hesham Elhalawani, David A. Hormuth, Petra J.Van Houdt, Renjie He, Shouhao Zhou, Kelsey B. Mathieu, Heng Li, Catherine Coolens, Caroline Chung, James A. Bankson, Wei Huang, Jihong Wang & 12 others Vlad C. Sandulache, Stephen Y. Lai, Rebecca M. Howell, R. Jason Stafford, Thomas E. Yankeelov, Uulke A.Van Der Heide, Steven J. Frank, Daniel P. Barboriak, John D. Hazle, Laurence E. Court, Jayashree Kalpathy-Cramer, Clifton D. Fuller

Research output: Research - peer-reviewArticle

Abstract

Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) provides quantitative metrics (e.g. Ktrans, ve) via pharmacokinetic models. We tested inter-algorithm variability in these quantitative metrics with 11 published DCE-MRI algorithms, all implementing Tofts-Kermode or extended Tofts pharmacokinetic models. Digital reference objects (DROs) with known Ktrans and ve values were used to assess performance at varying noise levels. Additionally, DCE-MRI data from 15 head and neck squamous cell carcinoma patients over 3 time-points during chemoradiotherapy were used to ascertain Ktrans and ve kinetic trends across algorithms. Algorithms performed well (less than 3% average error) when no noise was present in the DRO. With noise, 87% of Ktrans and 84% of ve algorithm-DRO combinations were generally in the correct order. Low Krippendorff's alpha values showed that algorithms could not consistently classify patients as above or below the median for a given algorithm at each time point or for differences in values between time points. A majority of the algorithms produced a significant Spearman correlation in ve of the primary gross tumor volume with time. Algorithmic differences in Ktrans and ve values over time indicate limitations in combining/comparing data from distinct DCE-MRI model implementations. Careful cross-algorithm quality-assurance must be utilized as DCE-MRI results may not be interpretable using differing software.

LanguageEnglish (US)
Article number11185
JournalScientific Reports
Volume7
Issue number1
DOIs
StatePublished - Dec 1 2017

Fingerprint

Magnetic Resonance Imaging
Noise
Pharmacokinetics
Chemoradiotherapy
Tumor Burden
Software
Carcinoma, squamous cell of head and neck

ASJC Scopus subject areas

  • General

Cite this

Ger, R. B., Mohamed, A. S. R., Awan, M. J., Ding, Y., Li, K., Fave, X. J., ... Fuller, C. D. (2017). A Multi-Institutional Comparison of Dynamic Contrast-Enhanced Magnetic Resonance Imaging Parameter Calculations. Scientific Reports, 7(1), [11185]. DOI: 10.1038/s41598-017-11554-w

A Multi-Institutional Comparison of Dynamic Contrast-Enhanced Magnetic Resonance Imaging Parameter Calculations. / Ger, Rachel B.; Mohamed, Abdallah S.R.; Awan, Musaddiq J.; Ding, Yao; Li, Kimberly; Fave, Xenia J.; Beers, Andrew L.; Driscoll, Brandon; Elhalawani, Hesham; Hormuth, David A.; Houdt, Petra J.Van; He, Renjie; Zhou, Shouhao; Mathieu, Kelsey B.; Li, Heng; Coolens, Catherine; Chung, Caroline; Bankson, James A.; Huang, Wei; Wang, Jihong; Sandulache, Vlad C.; Lai, Stephen Y.; Howell, Rebecca M.; Stafford, R. Jason; Yankeelov, Thomas E.; Heide, Uulke A.Van Der; Frank, Steven J.; Barboriak, Daniel P.; Hazle, John D.; Court, Laurence E.; Kalpathy-Cramer, Jayashree; Fuller, Clifton D.

In: Scientific Reports, Vol. 7, No. 1, 11185, 01.12.2017.

Research output: Research - peer-reviewArticle

Ger, RB, Mohamed, ASR, Awan, MJ, Ding, Y, Li, K, Fave, XJ, Beers, AL, Driscoll, B, Elhalawani, H, Hormuth, DA, Houdt, PJV, He, R, Zhou, S, Mathieu, KB, Li, H, Coolens, C, Chung, C, Bankson, JA, Huang, W, Wang, J, Sandulache, VC, Lai, SY, Howell, RM, Stafford, RJ, Yankeelov, TE, Heide, UAVD, Frank, SJ, Barboriak, DP, Hazle, JD, Court, LE, Kalpathy-Cramer, J & Fuller, CD 2017, 'A Multi-Institutional Comparison of Dynamic Contrast-Enhanced Magnetic Resonance Imaging Parameter Calculations' Scientific Reports, vol 7, no. 1, 11185. DOI: 10.1038/s41598-017-11554-w
Ger RB, Mohamed ASR, Awan MJ, Ding Y, Li K, Fave XJ et al. A Multi-Institutional Comparison of Dynamic Contrast-Enhanced Magnetic Resonance Imaging Parameter Calculations. Scientific Reports. 2017 Dec 1;7(1). 11185. Available from, DOI: 10.1038/s41598-017-11554-w
Ger, Rachel B. ; Mohamed, Abdallah S.R. ; Awan, Musaddiq J. ; Ding, Yao ; Li, Kimberly ; Fave, Xenia J. ; Beers, Andrew L. ; Driscoll, Brandon ; Elhalawani, Hesham ; Hormuth, David A. ; Houdt, Petra J.Van ; He, Renjie ; Zhou, Shouhao ; Mathieu, Kelsey B. ; Li, Heng ; Coolens, Catherine ; Chung, Caroline ; Bankson, James A. ; Huang, Wei ; Wang, Jihong ; Sandulache, Vlad C. ; Lai, Stephen Y. ; Howell, Rebecca M. ; Stafford, R. Jason ; Yankeelov, Thomas E. ; Heide, Uulke A.Van Der ; Frank, Steven J. ; Barboriak, Daniel P. ; Hazle, John D. ; Court, Laurence E. ; Kalpathy-Cramer, Jayashree ; Fuller, Clifton D./ A Multi-Institutional Comparison of Dynamic Contrast-Enhanced Magnetic Resonance Imaging Parameter Calculations. In: Scientific Reports. 2017 ; Vol. 7, No. 1.
@article{9f944b8be7834c4bab33aa5c62fa4648,
title = "A Multi-Institutional Comparison of Dynamic Contrast-Enhanced Magnetic Resonance Imaging Parameter Calculations",
abstract = "Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) provides quantitative metrics (e.g. Ktrans, ve) via pharmacokinetic models. We tested inter-algorithm variability in these quantitative metrics with 11 published DCE-MRI algorithms, all implementing Tofts-Kermode or extended Tofts pharmacokinetic models. Digital reference objects (DROs) with known Ktrans and ve values were used to assess performance at varying noise levels. Additionally, DCE-MRI data from 15 head and neck squamous cell carcinoma patients over 3 time-points during chemoradiotherapy were used to ascertain Ktrans and ve kinetic trends across algorithms. Algorithms performed well (less than 3% average error) when no noise was present in the DRO. With noise, 87% of Ktrans and 84% of ve algorithm-DRO combinations were generally in the correct order. Low Krippendorff's alpha values showed that algorithms could not consistently classify patients as above or below the median for a given algorithm at each time point or for differences in values between time points. A majority of the algorithms produced a significant Spearman correlation in ve of the primary gross tumor volume with time. Algorithmic differences in Ktrans and ve values over time indicate limitations in combining/comparing data from distinct DCE-MRI model implementations. Careful cross-algorithm quality-assurance must be utilized as DCE-MRI results may not be interpretable using differing software.",
author = "Ger, {Rachel B.} and Mohamed, {Abdallah S.R.} and Awan, {Musaddiq J.} and Yao Ding and Kimberly Li and Fave, {Xenia J.} and Beers, {Andrew L.} and Brandon Driscoll and Hesham Elhalawani and Hormuth, {David A.} and Houdt, {Petra J.Van} and Renjie He and Shouhao Zhou and Mathieu, {Kelsey B.} and Heng Li and Catherine Coolens and Caroline Chung and Bankson, {James A.} and Wei Huang and Jihong Wang and Sandulache, {Vlad C.} and Lai, {Stephen Y.} and Howell, {Rebecca M.} and Stafford, {R. Jason} and Yankeelov, {Thomas E.} and Heide, {Uulke A.Van Der} and Frank, {Steven J.} and Barboriak, {Daniel P.} and Hazle, {John D.} and Court, {Laurence E.} and Jayashree Kalpathy-Cramer and Fuller, {Clifton D.}",
year = "2017",
month = "12",
doi = "10.1038/s41598-017-11554-w",
volume = "7",
journal = "Scientific Reports",
issn = "2045-2322",
publisher = "Nature Publishing Group",
number = "1",

}

TY - JOUR

T1 - A Multi-Institutional Comparison of Dynamic Contrast-Enhanced Magnetic Resonance Imaging Parameter Calculations

AU - Ger,Rachel B.

AU - Mohamed,Abdallah S.R.

AU - Awan,Musaddiq J.

AU - Ding,Yao

AU - Li,Kimberly

AU - Fave,Xenia J.

AU - Beers,Andrew L.

AU - Driscoll,Brandon

AU - Elhalawani,Hesham

AU - Hormuth,David A.

AU - Houdt,Petra J.Van

AU - He,Renjie

AU - Zhou,Shouhao

AU - Mathieu,Kelsey B.

AU - Li,Heng

AU - Coolens,Catherine

AU - Chung,Caroline

AU - Bankson,James A.

AU - Huang,Wei

AU - Wang,Jihong

AU - Sandulache,Vlad C.

AU - Lai,Stephen Y.

AU - Howell,Rebecca M.

AU - Stafford,R. Jason

AU - Yankeelov,Thomas E.

AU - Heide,Uulke A.Van Der

AU - Frank,Steven J.

AU - Barboriak,Daniel P.

AU - Hazle,John D.

AU - Court,Laurence E.

AU - Kalpathy-Cramer,Jayashree

AU - Fuller,Clifton D.

PY - 2017/12/1

Y1 - 2017/12/1

N2 - Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) provides quantitative metrics (e.g. Ktrans, ve) via pharmacokinetic models. We tested inter-algorithm variability in these quantitative metrics with 11 published DCE-MRI algorithms, all implementing Tofts-Kermode or extended Tofts pharmacokinetic models. Digital reference objects (DROs) with known Ktrans and ve values were used to assess performance at varying noise levels. Additionally, DCE-MRI data from 15 head and neck squamous cell carcinoma patients over 3 time-points during chemoradiotherapy were used to ascertain Ktrans and ve kinetic trends across algorithms. Algorithms performed well (less than 3% average error) when no noise was present in the DRO. With noise, 87% of Ktrans and 84% of ve algorithm-DRO combinations were generally in the correct order. Low Krippendorff's alpha values showed that algorithms could not consistently classify patients as above or below the median for a given algorithm at each time point or for differences in values between time points. A majority of the algorithms produced a significant Spearman correlation in ve of the primary gross tumor volume with time. Algorithmic differences in Ktrans and ve values over time indicate limitations in combining/comparing data from distinct DCE-MRI model implementations. Careful cross-algorithm quality-assurance must be utilized as DCE-MRI results may not be interpretable using differing software.

AB - Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) provides quantitative metrics (e.g. Ktrans, ve) via pharmacokinetic models. We tested inter-algorithm variability in these quantitative metrics with 11 published DCE-MRI algorithms, all implementing Tofts-Kermode or extended Tofts pharmacokinetic models. Digital reference objects (DROs) with known Ktrans and ve values were used to assess performance at varying noise levels. Additionally, DCE-MRI data from 15 head and neck squamous cell carcinoma patients over 3 time-points during chemoradiotherapy were used to ascertain Ktrans and ve kinetic trends across algorithms. Algorithms performed well (less than 3% average error) when no noise was present in the DRO. With noise, 87% of Ktrans and 84% of ve algorithm-DRO combinations were generally in the correct order. Low Krippendorff's alpha values showed that algorithms could not consistently classify patients as above or below the median for a given algorithm at each time point or for differences in values between time points. A majority of the algorithms produced a significant Spearman correlation in ve of the primary gross tumor volume with time. Algorithmic differences in Ktrans and ve values over time indicate limitations in combining/comparing data from distinct DCE-MRI model implementations. Careful cross-algorithm quality-assurance must be utilized as DCE-MRI results may not be interpretable using differing software.

UR - http://www.scopus.com/inward/record.url?scp=85029332732&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85029332732&partnerID=8YFLogxK

U2 - 10.1038/s41598-017-11554-w

DO - 10.1038/s41598-017-11554-w

M3 - Article

VL - 7

JO - Scientific Reports

T2 - Scientific Reports

JF - Scientific Reports

SN - 2045-2322

IS - 1

M1 - 11185

ER -