Abstract

Purpose: The new classification announced by the World Health Organization in 2016 recognized five molecular subtypes of diffuse gliomas based on isocitrate dehydrogenase (IDH) and 1p/19q genotypes in addition to histologic phenotypes. We aim to determine whether clinical MRI can stratify these molecular subtypes to benefit the diagnosis and monitoring of gliomas. Experimental Design: The data from 456 subjects with gliomas were obtained from The Cancer Imaging Archive. Overall, 214 subjects, including 106 cases of glioblastomas and 108 cases of lower grade gliomas with preoperative MRI, survival data, histology, IDH, and 1p/19q status were included. We proposed a three-level machine-learning model based on multimodal MR radiomics to classify glioma subtypes. An independent dataset with 70 glioma subjects was further collected to verify the model performance. Results: The IDH and 1p/19q status of gliomas can be classified by radiomics and machine-learning approaches, with areas under ROC curves between 0.922 and 0.975 and accuracies between 87.7% and 96.1% estimated on the training dataset. The test on the validation dataset showed a comparable model performance with that on the training dataset, suggesting the efficacy of the trained classifiers. The classification of 5 molecular subtypes solely based on the MR phenotypes achieved an 81.8% accuracy, and a higher accuracy of 89.2% could be achieved if the histology diagnosis is available. Conclusions: The MR radiomics-based method provides a reliable alternative to determine the histology and molecular subtypes of gliomas.

Original languageEnglish
Pages (from-to)4429-4436
Number of pages8
JournalClinical Cancer Research
Volume24
Issue number18
DOIs
Publication statusPublished - Sep 15 2018

Fingerprint

Glioma
Isocitrate Dehydrogenase
Histology
Phenotype
Glioblastoma
ROC Curve
Area Under Curve
Research Design
Genotype
Datasets
Neoplasms

ASJC Scopus subject areas

  • Oncology
  • Cancer Research

Cite this

Machine learning–based radiomics for molecular subtyping of gliomas. / Lu, C.-F.; Hsu, F.-T.; Hsieh, K.L.-C.; Kao, Y.-C.J.; Cheng, S.-J.; Hsu, J.B.-K.; Tsai, P.-H.; Chen, R.-J.; Huang, C.-C.; Yen, Y.; Chen, C.-Y.

In: Clinical Cancer Research, Vol. 24, No. 18, 15.09.2018, p. 4429-4436.

Research output: Contribution to journalArticle

@article{2dbb2c8ff5194840855196a99ed396db,
title = "Machine learning–based radiomics for molecular subtyping of gliomas",
abstract = "Purpose: The new classification announced by the World Health Organization in 2016 recognized five molecular subtypes of diffuse gliomas based on isocitrate dehydrogenase (IDH) and 1p/19q genotypes in addition to histologic phenotypes. We aim to determine whether clinical MRI can stratify these molecular subtypes to benefit the diagnosis and monitoring of gliomas. Experimental Design: The data from 456 subjects with gliomas were obtained from The Cancer Imaging Archive. Overall, 214 subjects, including 106 cases of glioblastomas and 108 cases of lower grade gliomas with preoperative MRI, survival data, histology, IDH, and 1p/19q status were included. We proposed a three-level machine-learning model based on multimodal MR radiomics to classify glioma subtypes. An independent dataset with 70 glioma subjects was further collected to verify the model performance. Results: The IDH and 1p/19q status of gliomas can be classified by radiomics and machine-learning approaches, with areas under ROC curves between 0.922 and 0.975 and accuracies between 87.7{\%} and 96.1{\%} estimated on the training dataset. The test on the validation dataset showed a comparable model performance with that on the training dataset, suggesting the efficacy of the trained classifiers. The classification of 5 molecular subtypes solely based on the MR phenotypes achieved an 81.8{\%} accuracy, and a higher accuracy of 89.2{\%} could be achieved if the histology diagnosis is available. Conclusions: The MR radiomics-based method provides a reliable alternative to determine the histology and molecular subtypes of gliomas.",
author = "C.-F. Lu and F.-T. Hsu and K.L.-C. Hsieh and Y.-C.J. Kao and S.-J. Cheng and J.B.-K. Hsu and P.-H. Tsai and R.-J. Chen and C.-C. Huang and Y. Yen and C.-Y. Chen",
note = "Export Date: 27 October 2018 CODEN: CCREF Correspondence Address: Chen, C.-Y.; Department of Radiology, School of Medicine, College of Medicine, Taipei Medical University, 250 Wuxing Street, Taiwan; email: sandy0932@gmail.com References: Eckel-Passow, J.E., Lachance, D.H., Molinaro, A.M., Walsh, K.M., Decker, P.A., Sicotte, H., Glioma groups based on 1p/19q, IDH, and TERT promoter mutations in tumors (2015) N Engl J Med, 372, pp. 2499-2508; Brat, D.J., Verhaak, R.G., Aldape, K.D., Yung, W.K., Salama, S.R., Cooper, L.A., Comprehensive, integrative genomic analysis of diffuse lower-grade gliomas (2015) N Engl J Med, 372, pp. 2481-2498; Ceccarelli, M., Barthel, F.P., Malta, T.M., Sabedot, T.S., Salama, S.R., Murray, B.A., Molecular profiling reveals biologically discrete subsets and pathways of progression in diffuse glioma (2016) Cell, 164, pp. 550-563; Louis, D.N., Ohgaki, H., Wiestler, O.D., Cavenee, W.K., (2016) World Health Organization Histological Classification of Tumours of The Central Nervous System, , Lyon, France: International Agency for Research on Cancer; Louis, D.N., Perry, A., Reifenberger, G., von Deimling, A., Figarella-Branger, D., Cavenee, W.K., The 2016 World Health Organization classification of tumors of the central nervous system: A summary (2016) Acta Neuropathol, 131, pp. 803-820; Diehn, M., Nardini, C., Wang, D.S., McGovern, S., Jayaraman, M., Liang, Y., Identification of noninvasive imaging surrogates for brain tumor gene-expression modules (2008) Proc Natl Acad Sci U S A, 105, pp. 5213-5218; Lambin, P., Rios-Velazquez, E., Leijenaar, R., Carvalho, S., Van Stiphout, R.G., Granton, P., Radiomics: Extracting more information from medical images using advanced feature analysis (2012) Eur J Cancer, 48, pp. 441-446; Aerts, H.J., Velazquez, E.R., Leijenaar, R.T., Parmar, C., Grossmann, P., Carvalho, S., Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach (2014) Nat Commun, 5, p. 4006; Hsieh, K.L.-C., Chen, C.-Y., Lo, C.-M., Quantitative glioma grading using transformed gray-scale invariant textures of MRI (2017) Computers Biol Med, 83, pp. 102-108; Hsieh, K.L.-C., Lo, C.-M., Hsiao, C.-J., Computer-aided grading of gliomas based on local and global MRI features (2017) Computer Methods Prog Biomed, 139, pp. 31-38; Zhang, B., Chang, K., Ramkissoon, S., Tanguturi, S., Bi, W.L., Reardon, D.A., Multimodal MRI features predict isocitrate dehydrogenase genotype in high-grade gliomas (2017) Neuro-Oncol, 19, pp. 109-117; Hsieh, K., Chen, C., Lo, C., Radiomic model for predicting mutations in the isocitrate dehydrogenase gene in glioblastomas (2017) Oncotarget, 8, pp. 45888-45897; Kickingereder, P., Burth, S., Wick, A., Gotz, M., Eidel, O., Schlemmer, H.P., Radiomic profiling of glioblastoma: Identifying an imaging predictor of patient survival with improved performance over established clinical and radiologic risk models (2016) Radiology, 280, pp. 880-889; Clark, K., Vendt, B., Smith, K., Freymann, J., Kirby, J., Koppel, P., The Cancer Imaging Archive (TCIA): Maintaining and operating a public information repository (2013) J Digit Imaging, 26, pp. 1045-1057; Scarpace, L., Mikkelsen, T., Cha, S., Rao, S., Tekchandani, S., Gutman, D., Radiology data from the cancer genome atlas glioblastoma multiforme [TCGA-GBM] collection (2016) The Cancer Imaging Archive; Pedano, N., Flanders, A.E., Scarpace, L., Mikkelsen, T., Eschbacher, J.M., Hermes, B., Radiology data from the cancer genome atlas low grade glioma [TCGA-LGG] collection (2016) The Cancer Imaging Archive; Scarpace, L., Flanders, A.E., Jain, R., Mikkelsen, T., Andrews, D.W., Data from rembrandt (2016) The Cancer Imaging Archive; Starck, J.-L., Fadili, J., Murtagh, F., The undecimated wavelet decomposition and its reconstruction (2007) IEEE Trans Image Process, 16, pp. 297-309; Ojala, T., Pietikainen, M., Maenpaa, T., (2000) Gray Scale and Rotation Invariant Texture Classification with Local Binary Patterns, , Berlin/Heidelberg, Germany: Springer; Rister, B., Horowitz, M.A., Rubin, D.L., Volumetric image registration from invariant keypoints (2017) IEEE Trans Image Process, 26, pp. 4900-4910; Cheung, W., Hamarneh, G., N-SIFT: N-dimensional scale invariant feature transform (2009) IEEE Trans Image Process, 18, pp. 2012-2021; Scholkopf, B., Smola, A.J., (2002) Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond, , Cambridge, MA: MIT press; Breiman, L., Random forests (2001) Machine Learn, 45, pp. 5-32; Ratsch, G., Onoda, T., Muller, K.-R., Soft margins for AdaBoost (2001) Machine Learn, 42, pp. 287-320; Seiffert, C., Khoshgoftaar, T.M., Van Hulse, J., RusBoost: A hybrid approach to alleviating class imbalance (2010) IEEE Transactions on Systems, 40, pp. 185-197; Guyon, I., Elisseeff, A., An introduction to variable and feature selection (2003) J Machine Learn Res, 3, pp. 1157-1182; Matthews, B.W., Comparison of the predicted and observed secondary structure of T4 phage lysozyme (1975) Biochim Biophys Acta, 405, pp. 442-451; Powers, D.M., Evaluation: From precision, recall and F-measure to ROC, informedness, markedness and correlation (2011) J Mach Learn Tech, 2, pp. 37-63; Mukaka, M.M., A guide to appropriate use of correlation coefficient in medical research (2012) Malawi Med J, 24, pp. 69-71; Hinkle, D.E., Wiersma, W., Jurs, S.G., (2003) Applied Statistics for The Behavioral Sciences, , Boston, MA: Houghton Mifflin College Division; Upadhyay, N., Waldman, A., Conventional MRI evaluation of gliomas (2011) Br J Radiol, 84, pp. S107-S111; Scott, J., Brasher, P.M., Sevick, R.J., Rewcastle, N.B., Forsyth, P.A., How often are nonenhancing supratentorial gliomas malignant? A population study (2002) Neurology, 59, pp. 947-949; Wiestler, B., Capper, D., Holland-Letz, T., Korshunov, A., von Deimling, A., Pfister, S.M., ATRX loss refines the classification of anaplastic gliomas and identifies a subgroup of IDH mutant astrocytic tumors with better prognosis (2013) Acta Neuropathol, 126, p. 443; Law, M., Young, R.J., Babb, J.S., Peccerelli, N., Chheang, S., Gruber, M.L., Gliomas: Predicting time to progression or survival with cerebral blood volume measurements at dynamic susceptibility-weighted contrast-enhanced perfusion MR imaging (2008) Radiology, 247, pp. 490-498; Law, M., Yang, S., Babb, J.S., Knopp, E.A., Golfinos, J.G., Zagzag, D., Comparison of cerebral blood volume and vascular permeability from dynamic susceptibility contrast-enhanced perfusion MR imaging with glioma grade (2004) Am J Neuroradiol, 25, pp. 746-755; Choi, C., Ganji, S.K., DeBerardinis, R.J., Hatanpaa, K.J., Rakheja, D., Kovacs, Z., 2-hydroxyglutarate detection by magnetic resonance spectroscopy in IDH-mutated patients with gliomas (2012) Nat Med, 18, pp. 624-629; Van Cauter, S., Veraart, J., Sijbers, J., Peeters, R.R., Himmelreich, U., De Keyzer, F., Gliomas: Diffusion kurtosis MR imaging in grading (2012) Radiology, 263, pp. 492-501; Raab, P., Hattingen, E., Franz, K., Zanella, F.E., Lanfermann, H., Cerebral gliomas: Diffusional kurtosis imaging analysis of microstructural differences 1 (2010) Radiology, 254, pp. 876-881; Pereira, S., Pinto, A., Alves, V., Silva, C.A., Deep convolutional neural networks for the segmentation of gliomas in multi-sequence MRI (2015) Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries. Lecture Notes in Computer Science, , Crimi A, Menze B, Maier O, Reyes M, Handels H, editors. Springer: Cham, Switzerland; Nie, D., Zhang, H., Adeli, E., Liu, L., Shen, D., 3D deep learning for multimodal imaging-guided survival time prediction of brain tumor patients (2016) Medical Image Computing and Computer-Assisted Intervention – MICCAI 2016. Lecture Notes in Computer Science, , Ourselin S, Joskowicz L, Sabuncu M, Unal G, Wells W, editors. Springer: Cham, Switzerland; Shin, H.-C., Roth, H.R., Gao, M., Lu, L., Xu, Z., Nogues, I., Deep convolutional neural networks for computer-aided detection: CNN architectures, dataset characteristics and transfer learning (2016) IEEE Trans Med Imaging, 35, pp. 1285-1298",
year = "2018",
month = "9",
day = "15",
doi = "10.1158/1078-0432.CCR-17-3445",
language = "English",
volume = "24",
pages = "4429--4436",
journal = "Clinical Cancer Research",
issn = "1078-0432",
publisher = "American Association for Cancer Research Inc.",
number = "18",

}

TY - JOUR

T1 - Machine learning–based radiomics for molecular subtyping of gliomas

AU - Lu, C.-F.

AU - Hsu, F.-T.

AU - Hsieh, K.L.-C.

AU - Kao, Y.-C.J.

AU - Cheng, S.-J.

AU - Hsu, J.B.-K.

AU - Tsai, P.-H.

AU - Chen, R.-J.

AU - Huang, C.-C.

AU - Yen, Y.

AU - Chen, C.-Y.

N1 - Export Date: 27 October 2018 CODEN: CCREF Correspondence Address: Chen, C.-Y.; Department of Radiology, School of Medicine, College of Medicine, Taipei Medical University, 250 Wuxing Street, Taiwan; email: sandy0932@gmail.com References: Eckel-Passow, J.E., Lachance, D.H., Molinaro, A.M., Walsh, K.M., Decker, P.A., Sicotte, H., Glioma groups based on 1p/19q, IDH, and TERT promoter mutations in tumors (2015) N Engl J Med, 372, pp. 2499-2508; Brat, D.J., Verhaak, R.G., Aldape, K.D., Yung, W.K., Salama, S.R., Cooper, L.A., Comprehensive, integrative genomic analysis of diffuse lower-grade gliomas (2015) N Engl J Med, 372, pp. 2481-2498; Ceccarelli, M., Barthel, F.P., Malta, T.M., Sabedot, T.S., Salama, S.R., Murray, B.A., Molecular profiling reveals biologically discrete subsets and pathways of progression in diffuse glioma (2016) Cell, 164, pp. 550-563; Louis, D.N., Ohgaki, H., Wiestler, O.D., Cavenee, W.K., (2016) World Health Organization Histological Classification of Tumours of The Central Nervous System, , Lyon, France: International Agency for Research on Cancer; Louis, D.N., Perry, A., Reifenberger, G., von Deimling, A., Figarella-Branger, D., Cavenee, W.K., The 2016 World Health Organization classification of tumors of the central nervous system: A summary (2016) Acta Neuropathol, 131, pp. 803-820; Diehn, M., Nardini, C., Wang, D.S., McGovern, S., Jayaraman, M., Liang, Y., Identification of noninvasive imaging surrogates for brain tumor gene-expression modules (2008) Proc Natl Acad Sci U S A, 105, pp. 5213-5218; Lambin, P., Rios-Velazquez, E., Leijenaar, R., Carvalho, S., Van Stiphout, R.G., Granton, P., Radiomics: Extracting more information from medical images using advanced feature analysis (2012) Eur J Cancer, 48, pp. 441-446; Aerts, H.J., Velazquez, E.R., Leijenaar, R.T., Parmar, C., Grossmann, P., Carvalho, S., Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach (2014) Nat Commun, 5, p. 4006; Hsieh, K.L.-C., Chen, C.-Y., Lo, C.-M., Quantitative glioma grading using transformed gray-scale invariant textures of MRI (2017) Computers Biol Med, 83, pp. 102-108; Hsieh, K.L.-C., Lo, C.-M., Hsiao, C.-J., Computer-aided grading of gliomas based on local and global MRI features (2017) Computer Methods Prog Biomed, 139, pp. 31-38; Zhang, B., Chang, K., Ramkissoon, S., Tanguturi, S., Bi, W.L., Reardon, D.A., Multimodal MRI features predict isocitrate dehydrogenase genotype in high-grade gliomas (2017) Neuro-Oncol, 19, pp. 109-117; Hsieh, K., Chen, C., Lo, C., Radiomic model for predicting mutations in the isocitrate dehydrogenase gene in glioblastomas (2017) Oncotarget, 8, pp. 45888-45897; Kickingereder, P., Burth, S., Wick, A., Gotz, M., Eidel, O., Schlemmer, H.P., Radiomic profiling of glioblastoma: Identifying an imaging predictor of patient survival with improved performance over established clinical and radiologic risk models (2016) Radiology, 280, pp. 880-889; Clark, K., Vendt, B., Smith, K., Freymann, J., Kirby, J., Koppel, P., The Cancer Imaging Archive (TCIA): Maintaining and operating a public information repository (2013) J Digit Imaging, 26, pp. 1045-1057; Scarpace, L., Mikkelsen, T., Cha, S., Rao, S., Tekchandani, S., Gutman, D., Radiology data from the cancer genome atlas glioblastoma multiforme [TCGA-GBM] collection (2016) The Cancer Imaging Archive; Pedano, N., Flanders, A.E., Scarpace, L., Mikkelsen, T., Eschbacher, J.M., Hermes, B., Radiology data from the cancer genome atlas low grade glioma [TCGA-LGG] collection (2016) The Cancer Imaging Archive; Scarpace, L., Flanders, A.E., Jain, R., Mikkelsen, T., Andrews, D.W., Data from rembrandt (2016) The Cancer Imaging Archive; Starck, J.-L., Fadili, J., Murtagh, F., The undecimated wavelet decomposition and its reconstruction (2007) IEEE Trans Image Process, 16, pp. 297-309; Ojala, T., Pietikainen, M., Maenpaa, T., (2000) Gray Scale and Rotation Invariant Texture Classification with Local Binary Patterns, , Berlin/Heidelberg, Germany: Springer; Rister, B., Horowitz, M.A., Rubin, D.L., Volumetric image registration from invariant keypoints (2017) IEEE Trans Image Process, 26, pp. 4900-4910; Cheung, W., Hamarneh, G., N-SIFT: N-dimensional scale invariant feature transform (2009) IEEE Trans Image Process, 18, pp. 2012-2021; Scholkopf, B., Smola, A.J., (2002) Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond, , Cambridge, MA: MIT press; Breiman, L., Random forests (2001) Machine Learn, 45, pp. 5-32; Ratsch, G., Onoda, T., Muller, K.-R., Soft margins for AdaBoost (2001) Machine Learn, 42, pp. 287-320; Seiffert, C., Khoshgoftaar, T.M., Van Hulse, J., RusBoost: A hybrid approach to alleviating class imbalance (2010) IEEE Transactions on Systems, 40, pp. 185-197; Guyon, I., Elisseeff, A., An introduction to variable and feature selection (2003) J Machine Learn Res, 3, pp. 1157-1182; Matthews, B.W., Comparison of the predicted and observed secondary structure of T4 phage lysozyme (1975) Biochim Biophys Acta, 405, pp. 442-451; Powers, D.M., Evaluation: From precision, recall and F-measure to ROC, informedness, markedness and correlation (2011) J Mach Learn Tech, 2, pp. 37-63; Mukaka, M.M., A guide to appropriate use of correlation coefficient in medical research (2012) Malawi Med J, 24, pp. 69-71; Hinkle, D.E., Wiersma, W., Jurs, S.G., (2003) Applied Statistics for The Behavioral Sciences, , Boston, MA: Houghton Mifflin College Division; Upadhyay, N., Waldman, A., Conventional MRI evaluation of gliomas (2011) Br J Radiol, 84, pp. S107-S111; Scott, J., Brasher, P.M., Sevick, R.J., Rewcastle, N.B., Forsyth, P.A., How often are nonenhancing supratentorial gliomas malignant? A population study (2002) Neurology, 59, pp. 947-949; Wiestler, B., Capper, D., Holland-Letz, T., Korshunov, A., von Deimling, A., Pfister, S.M., ATRX loss refines the classification of anaplastic gliomas and identifies a subgroup of IDH mutant astrocytic tumors with better prognosis (2013) Acta Neuropathol, 126, p. 443; Law, M., Young, R.J., Babb, J.S., Peccerelli, N., Chheang, S., Gruber, M.L., Gliomas: Predicting time to progression or survival with cerebral blood volume measurements at dynamic susceptibility-weighted contrast-enhanced perfusion MR imaging (2008) Radiology, 247, pp. 490-498; Law, M., Yang, S., Babb, J.S., Knopp, E.A., Golfinos, J.G., Zagzag, D., Comparison of cerebral blood volume and vascular permeability from dynamic susceptibility contrast-enhanced perfusion MR imaging with glioma grade (2004) Am J Neuroradiol, 25, pp. 746-755; Choi, C., Ganji, S.K., DeBerardinis, R.J., Hatanpaa, K.J., Rakheja, D., Kovacs, Z., 2-hydroxyglutarate detection by magnetic resonance spectroscopy in IDH-mutated patients with gliomas (2012) Nat Med, 18, pp. 624-629; Van Cauter, S., Veraart, J., Sijbers, J., Peeters, R.R., Himmelreich, U., De Keyzer, F., Gliomas: Diffusion kurtosis MR imaging in grading (2012) Radiology, 263, pp. 492-501; Raab, P., Hattingen, E., Franz, K., Zanella, F.E., Lanfermann, H., Cerebral gliomas: Diffusional kurtosis imaging analysis of microstructural differences 1 (2010) Radiology, 254, pp. 876-881; Pereira, S., Pinto, A., Alves, V., Silva, C.A., Deep convolutional neural networks for the segmentation of gliomas in multi-sequence MRI (2015) Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries. Lecture Notes in Computer Science, , Crimi A, Menze B, Maier O, Reyes M, Handels H, editors. Springer: Cham, Switzerland; Nie, D., Zhang, H., Adeli, E., Liu, L., Shen, D., 3D deep learning for multimodal imaging-guided survival time prediction of brain tumor patients (2016) Medical Image Computing and Computer-Assisted Intervention – MICCAI 2016. Lecture Notes in Computer Science, , Ourselin S, Joskowicz L, Sabuncu M, Unal G, Wells W, editors. Springer: Cham, Switzerland; Shin, H.-C., Roth, H.R., Gao, M., Lu, L., Xu, Z., Nogues, I., Deep convolutional neural networks for computer-aided detection: CNN architectures, dataset characteristics and transfer learning (2016) IEEE Trans Med Imaging, 35, pp. 1285-1298

PY - 2018/9/15

Y1 - 2018/9/15

N2 - Purpose: The new classification announced by the World Health Organization in 2016 recognized five molecular subtypes of diffuse gliomas based on isocitrate dehydrogenase (IDH) and 1p/19q genotypes in addition to histologic phenotypes. We aim to determine whether clinical MRI can stratify these molecular subtypes to benefit the diagnosis and monitoring of gliomas. Experimental Design: The data from 456 subjects with gliomas were obtained from The Cancer Imaging Archive. Overall, 214 subjects, including 106 cases of glioblastomas and 108 cases of lower grade gliomas with preoperative MRI, survival data, histology, IDH, and 1p/19q status were included. We proposed a three-level machine-learning model based on multimodal MR radiomics to classify glioma subtypes. An independent dataset with 70 glioma subjects was further collected to verify the model performance. Results: The IDH and 1p/19q status of gliomas can be classified by radiomics and machine-learning approaches, with areas under ROC curves between 0.922 and 0.975 and accuracies between 87.7% and 96.1% estimated on the training dataset. The test on the validation dataset showed a comparable model performance with that on the training dataset, suggesting the efficacy of the trained classifiers. The classification of 5 molecular subtypes solely based on the MR phenotypes achieved an 81.8% accuracy, and a higher accuracy of 89.2% could be achieved if the histology diagnosis is available. Conclusions: The MR radiomics-based method provides a reliable alternative to determine the histology and molecular subtypes of gliomas.

AB - Purpose: The new classification announced by the World Health Organization in 2016 recognized five molecular subtypes of diffuse gliomas based on isocitrate dehydrogenase (IDH) and 1p/19q genotypes in addition to histologic phenotypes. We aim to determine whether clinical MRI can stratify these molecular subtypes to benefit the diagnosis and monitoring of gliomas. Experimental Design: The data from 456 subjects with gliomas were obtained from The Cancer Imaging Archive. Overall, 214 subjects, including 106 cases of glioblastomas and 108 cases of lower grade gliomas with preoperative MRI, survival data, histology, IDH, and 1p/19q status were included. We proposed a three-level machine-learning model based on multimodal MR radiomics to classify glioma subtypes. An independent dataset with 70 glioma subjects was further collected to verify the model performance. Results: The IDH and 1p/19q status of gliomas can be classified by radiomics and machine-learning approaches, with areas under ROC curves between 0.922 and 0.975 and accuracies between 87.7% and 96.1% estimated on the training dataset. The test on the validation dataset showed a comparable model performance with that on the training dataset, suggesting the efficacy of the trained classifiers. The classification of 5 molecular subtypes solely based on the MR phenotypes achieved an 81.8% accuracy, and a higher accuracy of 89.2% could be achieved if the histology diagnosis is available. Conclusions: The MR radiomics-based method provides a reliable alternative to determine the histology and molecular subtypes of gliomas.

UR - http://www.scopus.com/inward/record.url?scp=85053181904&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85053181904&partnerID=8YFLogxK

U2 - 10.1158/1078-0432.CCR-17-3445

DO - 10.1158/1078-0432.CCR-17-3445

M3 - Article

VL - 24

SP - 4429

EP - 4436

JO - Clinical Cancer Research

JF - Clinical Cancer Research

SN - 1078-0432

IS - 18

ER -