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

Autor: Julio Cárdenas-Rodríguez, Mark D. Pagel, Joshua M. Goldenberg
Rok vydání: 2018
Předmět:
Normal Distribution
Mice
SCID
Machine learning
computer.software_genre
Sensitivity and Specificity
Article
030218 nuclear medicine & medical imaging
Machine Learning
Mice
03 medical and health sciences
0302 clinical medicine
Cell Line
Tumor
Pancreatic cancer
Image Interpretation
Computer-Assisted
Image Processing
Computer-Assisted
medicine
Animals
Humans
Chemical Exchange Saturation Transfer MRI
Radiology
Nuclear Medicine and imaging
Tumor type
Hypoxia
Mathematics
Principal Component Analysis
Learning classifier system
business.industry
Spin–lattice relaxation
Reproducibility of Results
Image Enhancement
medicine.disease
Magnetic Resonance Imaging
Pancreatic Neoplasms
Area Under Curve
Principal component analysis
Classification methods
Female
Artificial intelligence
business
computer
Algorithms
030217 neurology & neurosurgery
Curse of dimensionality
Zdroj: Magnetic Resonance in Medicine. 81:594-601
ISSN: 0740-3194
DOI: 10.1002/mrm.27439
Popis: 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 T(1) relaxation, Chemical Exchange Saturation Transfer (CEST), and dynamic contrast-enhanced (DCE) MRI. METHODS: The T(1) relaxation time constants, % CEST at five saturation frequencies, and vascular permeability constants from DCE MRI were measured from Hs 766T, 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 (PCA) to reduce the dimensionality of entire CEST spectra and DCE signal evolutions, which were then analyzed using classification methods. RESULTS: The T(1) relaxation time constants, % CEST amplitudes at specific saturation frequencies, and the relative K(trans) and k(ep) values from DCE MRI could not classify all three tumor types. PCA was used to analyze entire CEST spectra, which predicted the correct tumor model with 87.5% accuracy. However, the AUC from DCE signal evolutions could classify each tumor type. PCA 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 T(1) relaxation times or DCE signal evolutions, relative to more simplistic analysis methods.
Databáze: OpenAIRE
Externí odkaz: