Machine learning improves classification of preclinical models of pancreatic cancer with chemical exchange saturation transfer MRI

Joshua M. Goldenberg, Julio Cárdenas-Rodríguez, Mark "Marty" Pagel

Research output: Contribution to journalArticle

Abstract

Purpose: We sought to assess whether machine learning-based classification approaches can improve the classification of pancreatic tumor models relative to more simplistic analysis methods, using T1 relaxation, CEST, and DCE MRI. Methods: The T1 relaxation time constants, % CEST at five saturation frequencies, and vascular permeability constants from DCE MRI were measured from Hs 766 T, MIA PaCa-2, and SU.86.86 pancreatic tumor models. We used each of these measurements as predictors for machine learning classifier algorithms. We also used principal component analysis to reduce the dimensionality of entire CEST spectra and DCE signal evolutions, which were then analyzed using classification methods. Results: The T1 relaxation time constants, % CEST amplitudes at specific saturation frequencies, and the relative Ktrans and kep values from DCE MRI could not classify all three tumor types. However, the area under the curve from DCE signal evolutions could classify each tumor type. Principal component analysis was used to analyze the entire CEST spectrum and DCE signal evolutions, which predicted the correct tumor model with 87.5% and 85.1% accuracy, respectively. Conclusions: Machine learning applied to the entire CEST spectrum improved the classification of the three tumor models, relative to classifications that used % CEST values at single saturation frequencies. A similar improvement was not attained with machine learning applied to T1 relaxation times or DCE signal evolutions, relative to more simplistic analysis methods.

Original languageEnglish (US)
JournalMagnetic Resonance in Medicine
DOIs
StateAccepted/In press - Jan 1 2018
Externally publishedYes

Fingerprint

Pancreatic Neoplasms
Neoplasms
Principal Component Analysis
Capillary Permeability
Area Under Curve
Machine Learning

Keywords

  • CEST
  • DCE. machine learning
  • MRI
  • MRI
  • pancreatic cancer imaging
  • PCA

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging

Cite this

Machine learning improves classification of preclinical models of pancreatic cancer with chemical exchange saturation transfer MRI. / Goldenberg, Joshua M.; Cárdenas-Rodríguez, Julio; Pagel, Mark "Marty".

In: Magnetic Resonance in Medicine, 01.01.2018.

Research output: Contribution to journalArticle

@article{1eceeeb1750d47e199fe06339b55fdaf,
title = "Machine learning improves classification of preclinical models of pancreatic cancer with chemical exchange saturation transfer MRI",
abstract = "Purpose: We sought to assess whether machine learning-based classification approaches can improve the classification of pancreatic tumor models relative to more simplistic analysis methods, using T1 relaxation, CEST, and DCE MRI. Methods: The T1 relaxation time constants, {\%} CEST at five saturation frequencies, and vascular permeability constants from DCE MRI were measured from Hs 766 T, MIA PaCa-2, and SU.86.86 pancreatic tumor models. We used each of these measurements as predictors for machine learning classifier algorithms. We also used principal component analysis to reduce the dimensionality of entire CEST spectra and DCE signal evolutions, which were then analyzed using classification methods. Results: The T1 relaxation time constants, {\%} CEST amplitudes at specific saturation frequencies, and the relative Ktrans and kep values from DCE MRI could not classify all three tumor types. However, the area under the curve from DCE signal evolutions could classify each tumor type. Principal component analysis was used to analyze the entire CEST spectrum and DCE signal evolutions, which predicted the correct tumor model with 87.5{\%} and 85.1{\%} accuracy, respectively. Conclusions: Machine learning applied to the entire CEST spectrum improved the classification of the three tumor models, relative to classifications that used {\%} CEST values at single saturation frequencies. A similar improvement was not attained with machine learning applied to T1 relaxation times or DCE signal evolutions, relative to more simplistic analysis methods.",
keywords = "CEST, DCE. machine learning, MRI, MRI, pancreatic cancer imaging, PCA",
author = "Goldenberg, {Joshua M.} and Julio C{\'a}rdenas-Rodr{\'i}guez and Pagel, {Mark {"}Marty{"}}",
year = "2018",
month = "1",
day = "1",
doi = "10.1002/mrm.27439",
language = "English (US)",
journal = "Magnetic Resonance in Medicine",
issn = "0740-3194",
publisher = "John Wiley and Sons Inc.",

}

TY - JOUR

T1 - Machine learning improves classification of preclinical models of pancreatic cancer with chemical exchange saturation transfer MRI

AU - Goldenberg, Joshua M.

AU - Cárdenas-Rodríguez, Julio

AU - Pagel, Mark "Marty"

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Purpose: We sought to assess whether machine learning-based classification approaches can improve the classification of pancreatic tumor models relative to more simplistic analysis methods, using T1 relaxation, CEST, and DCE MRI. Methods: The T1 relaxation time constants, % CEST at five saturation frequencies, and vascular permeability constants from DCE MRI were measured from Hs 766 T, MIA PaCa-2, and SU.86.86 pancreatic tumor models. We used each of these measurements as predictors for machine learning classifier algorithms. We also used principal component analysis to reduce the dimensionality of entire CEST spectra and DCE signal evolutions, which were then analyzed using classification methods. Results: The T1 relaxation time constants, % CEST amplitudes at specific saturation frequencies, and the relative Ktrans and kep values from DCE MRI could not classify all three tumor types. However, the area under the curve from DCE signal evolutions could classify each tumor type. Principal component analysis was used to analyze the entire CEST spectrum and DCE signal evolutions, which predicted the correct tumor model with 87.5% and 85.1% accuracy, respectively. Conclusions: Machine learning applied to the entire CEST spectrum improved the classification of the three tumor models, relative to classifications that used % CEST values at single saturation frequencies. A similar improvement was not attained with machine learning applied to T1 relaxation times or DCE signal evolutions, relative to more simplistic analysis methods.

AB - Purpose: We sought to assess whether machine learning-based classification approaches can improve the classification of pancreatic tumor models relative to more simplistic analysis methods, using T1 relaxation, CEST, and DCE MRI. Methods: The T1 relaxation time constants, % CEST at five saturation frequencies, and vascular permeability constants from DCE MRI were measured from Hs 766 T, MIA PaCa-2, and SU.86.86 pancreatic tumor models. We used each of these measurements as predictors for machine learning classifier algorithms. We also used principal component analysis to reduce the dimensionality of entire CEST spectra and DCE signal evolutions, which were then analyzed using classification methods. Results: The T1 relaxation time constants, % CEST amplitudes at specific saturation frequencies, and the relative Ktrans and kep values from DCE MRI could not classify all three tumor types. However, the area under the curve from DCE signal evolutions could classify each tumor type. Principal component analysis was used to analyze the entire CEST spectrum and DCE signal evolutions, which predicted the correct tumor model with 87.5% and 85.1% accuracy, respectively. Conclusions: Machine learning applied to the entire CEST spectrum improved the classification of the three tumor models, relative to classifications that used % CEST values at single saturation frequencies. A similar improvement was not attained with machine learning applied to T1 relaxation times or DCE signal evolutions, relative to more simplistic analysis methods.

KW - CEST

KW - DCE. machine learning

KW - MRI

KW - MRI

KW - pancreatic cancer imaging

KW - PCA

UR - http://www.scopus.com/inward/record.url?scp=85053544198&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85053544198&partnerID=8YFLogxK

U2 - 10.1002/mrm.27439

DO - 10.1002/mrm.27439

M3 - Article

C2 - 30277270

AN - SCOPUS:85053544198

JO - Magnetic Resonance in Medicine

JF - Magnetic Resonance in Medicine

SN - 0740-3194

ER -