TY - JOUR
T1 - Classification of brain tumor isocitrate dehydrogenase status using MRI and deep learning
AU - Nalawade, Sahil
AU - Murugesan, Gowtham K.
AU - Vejdani-Jahromi, Maryam
AU - Fisicaro, Ryan A.
AU - Bangalore Yogananda, Chandan G.
AU - Wagner, Ben
AU - Mickey, Bruce
AU - Maher, Elizabeth
AU - Pinho, Marco C.
AU - Fei, Baowei
AU - Madhuranthakam, Ananth J.
AU - Maldjian, Joseph A.
N1 - Publisher Copyright:
© 2019 Society of Photo-Optical Instrumentation Engineers (SPIE).
PY - 2019/10/1
Y1 - 2019/10/1
N2 - Isocitrate dehydrogenase (IDH) mutation status is an important marker in glioma diagnosis and therapy. We propose an automated pipeline for noninvasively predicting IDH status using deep learning and T2-weighted (T2w) magnetic resonance (MR) images with minimal preprocessing (N4 bias correction and normalization to zero mean and unit variance). T2w MR images and genomic data were obtained from The Cancer Imaging Archive dataset for 260 subjects (120 high-grade and 140 low-grade gliomas). A fully automated two-dimensional densely connected model was trained to classify IDH mutation status on 208 subjects and tested on another held-out set of 52 subjects using fivefold cross validation. Data leakage was avoided by ensuring subject separation during the slice-wise randomization. Mean classification accuracy of 90.5% was achieved for each axial slice in predicting the three classes of no tumor, IDH mutated, and IDH wild type. Test accuracy of 83.8% was achieved in predicting IDH mutation status for individual subjects on the test dataset of 52 subjects. We demonstrate a deep learning method to predict IDH mutation status using T2w MRI alone. Radiologic imaging studies using deep learning methods must address data leakage (subject duplication) in the randomization process to avoid upward bias in the reported classification accuracy.
AB - Isocitrate dehydrogenase (IDH) mutation status is an important marker in glioma diagnosis and therapy. We propose an automated pipeline for noninvasively predicting IDH status using deep learning and T2-weighted (T2w) magnetic resonance (MR) images with minimal preprocessing (N4 bias correction and normalization to zero mean and unit variance). T2w MR images and genomic data were obtained from The Cancer Imaging Archive dataset for 260 subjects (120 high-grade and 140 low-grade gliomas). A fully automated two-dimensional densely connected model was trained to classify IDH mutation status on 208 subjects and tested on another held-out set of 52 subjects using fivefold cross validation. Data leakage was avoided by ensuring subject separation during the slice-wise randomization. Mean classification accuracy of 90.5% was achieved for each axial slice in predicting the three classes of no tumor, IDH mutated, and IDH wild type. Test accuracy of 83.8% was achieved in predicting IDH mutation status for individual subjects on the test dataset of 52 subjects. We demonstrate a deep learning method to predict IDH mutation status using T2w MRI alone. Radiologic imaging studies using deep learning methods must address data leakage (subject duplication) in the randomization process to avoid upward bias in the reported classification accuracy.
KW - convolutional neural network
KW - deep learning
KW - isocitrate dehydrogenase
KW - magnetic resonance imaging
KW - segmentation
KW - tumor classification
UR - http://www.scopus.com/inward/record.url?scp=85077496941&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85077496941&partnerID=8YFLogxK
U2 - 10.1117/1.JMI.6.4.046003
DO - 10.1117/1.JMI.6.4.046003
M3 - Article
C2 - 31824982
AN - SCOPUS:85077496941
SN - 2329-4302
VL - 6
JO - Journal of Medical Imaging
JF - Journal of Medical Imaging
IS - 4
M1 - 046003
ER -