PURPOSE: Imaging biomarker research focuses on discovering relationships between radiological features and histological findings. In glioblastoma patients, methylation of the O(6)-methylguanine methyltransferase (MGMT) gene promoter is positively correlated with an increased effectiveness of current standard of care. In this paper, the authors investigate texture features as potential imaging biomarkers for capturing the MGMT methylation status of glioblastoma multiforme (GBM) tumors when combined with supervised classification schemes. METHODS: A retrospective study of 155 GBM patients with known MGMT methylation status was conducted. Co-occurrence and run length texture features were calculated, and both support vector machines (SVMs) and random forest classifiers were used to predict MGMT methylation status. RESULTS: The best classification system (an SVM-based classifier) had a maximum area under the receiver-operating characteristic (ROC) curve of 0.85 (95% CI: 0.78-0.91) using four texture features (correlation, energy, entropy, and local intensity) originating from the T2-weighted images, yielding at the optimal threshold of the ROC curve, a sensitivity of 0.803 and a specificity of 0.813. CONCLUSIONS: Results show that supervised machine learning of MRI texture features can predict MGMT methylation status in preoperative GBM tumors, thus providing a new noninvasive imaging biomarker.

Department of Health Sciences Research Mayo Clinic 200 1st Street SW Rochester Minnesota 55905

Department of Medical Oncology Mayo Clinic 200 1st Street SW Rochester Minnesota 55905

Department of Neurologic Surgery Mayo Clinic 200 1st Street SW Rochester Minnesota 55905

Department of Neurology Mayo Clinic 200 1st Street SW Rochester Minnesota 55905

Department of Radiology Mayo Clinic 200 1st Street SW Rochester Minnesota 55905

Department of Radiology Mayo Clinic 200 1st Street SW Rochester Minnesota 55905; Department of Neurosurgery of 1st Faculty of Medicine Charles University Prague Military University Hospital Prague 128 21 Czech Republic; and International Clinical Research Center St Anne's University Hospital Brno Brno 656 91 Czech Republic

See more in PubMed

Johnson D. R. and O’Neill B. P., “Glioblastoma survival in the United States before and during the temozolomide era,” J. Neuro-Oncol. 107, 359–364 (2012).10.1007/s11060-011-0749-4 PubMed DOI

Stupp R., Mason W. P., van den Bent M. J., Weller M., Fisher B., Taphoorn M. J., Belanger K., Brandes A. A., Marosi C., Bogdahn U., Curschmann J., Janzer R. C., Ludwin S. K., Gorlia T., Allgeier A., Lacombe D., Cairncross J. G., Eisenhauer E., Mirimanoff R. O., and European Organisation for Research and Treatment of Cancer, Brain Tumor, and Radiotherapy Groups, and National Cancer Institute of Canada Clinical Trials Group, “Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma,” N. Engl. J. Med. 352, 987–996 (2005).10.1056/NEJMoa043330 PubMed DOI

Ellingson B. M., Wen P. Y., van den Bent M. J., and Cloughesy T. F., “Pros and cons of current brain tumor imaging,” Neuro Oncol. 16(Suppl. 7), vii2–vii11 (2014).10.1093/neuonc/nou224 PubMed DOI PMC

Rivera A. L., Pelloski C. E., Gilbert M. R., Colman H., De La Cruz C., Sulman E. P., Bekele B. N., and Aldape K. D., “MGMT promoter methylation is predictive of response to radiotherapy and prognostic in the absence of adjuvant alkylating chemotherapy for glioblastoma,” Neuro Oncol. 12, 116–121 (2010).10.1093/neuonc/nop020 PubMed DOI PMC

Zhang K., Wang X. Q., Zhou B., and Zhang L., “The prognostic value of MGMT promoter methylation in glioblastoma multiforme: A meta-analysis,” Fam. Cancer 12, 449–458 (2013).10.1007/s10689-013-9607-1 PubMed DOI

Ellingson B. M., “Radiogenomics and imaging phenotypes in glioblastoma: Novel observations and correlation with molecular characteristics,” Curr. Neurol. Neurosci. Rep. 15, 506–518 (2015).10.1007/s11910-014-0506-0 PubMed DOI

Rundle-Thiele D., Day B., Stringer B., Fay M., Martin J., Jeffree R. L., Thomas P., Bell C., Salvado O., Gal Y., Coulthard A., Crozier S., and Rose S., “Using the apparent diffusion coefficient to identifying MGMT promoter methylation status early in glioblastoma: Importance of analytical method,” J. Med. Radiat. Sci. 62, 92–98 (2015).10.1002/jmrs.103 PubMed DOI PMC

Drabycz S., Roldan G., de Robles P., Adler D., McIntyre J. B., Magliocco A. M., Cairncross J. G., and Mitchell J. R., “An analysis of image texture, tumor location, and MGMT promoter methylation in glioblastoma using magnetic resonance imaging,” Neuroimage 49, 1398–1405 (2010).10.1016/j.neuroimage.2009.09.049 PubMed DOI

Levner I., Drabycz S., Roldan G., De Robles P., Cairncross J. G., and Mitchell R., “Predicting MGMT methylation status of glioblastomas from MRI texture,” Med. Image Comput. Comput. Assist. Interv. 12, 522–530 (2009).10.1007/978-3-642-04271-3_64 PubMed DOI

Moon W. J., Choi J. W., Roh H. G., Lim S. D., and Koh Y. C., “Imaging parameters of high grade gliomas in relation to the MGMT promoter methylation status: The CT, diffusion tensor imaging, and perfusion MR imaging,” Neuroradiology 54, 555–563 (2012).10.1007/s00234-011-0947-y PubMed DOI

Ahn S. S., Shin N. Y., Chang J. H., Kim S. H., Kim E. H., Kim D. W., and Lee S. K., “Prediction of methylguanine methyltransferase promoter methylation in glioblastoma using dynamic contrast-enhanced magnetic resonance and diffusion tensor imaging,” J. Neurosurg. 121, 367–373 (2014).10.3171/2014.5.JNS132279 PubMed DOI

Gupta A., Omuro A. M., Shah A. D., Graber J. J., Shi W., Zhang Z., and Young R. J., “Continuing the search for MR imaging biomarkers for MGMT promoter methylation status: Conventional and perfusion MRI revisited,” Neuroradiology 54, 641–643 (2012).10.1007/s00234-011-0970-z PubMed DOI PMC

Korfiatis P. D., Kline T. L., Blezek D. J., Langer S. G., Ryan W. J., and Erickson B. J., “MIRMAID: A content management system for medical image analysis research,” Radiographics 35, 1461–1468 (2015).10.1148/rg.2015140031 PubMed DOI PMC

Yushkevich P. A., Piven J., Hazlett H. C., Smith R. G., Ho S., Gee J. C., and Gerig G., “User-guided 3D active contour segmentation of anatomical structures: Significantly improved efficiency and reliability,” Neuroimage 31, 1116–1128 (2006).10.1016/j.neuroimage.2006.01.015 PubMed DOI

Fonov V., Evans A. C., Botteron K., Almli C. R., McKinstry R. C., and Collins D. L., “Unbiased average age-appropriate atlases for pediatric studies,” Neuroimage 54, 313–327 (2011).10.1016/j.neuroimage.2010.07.033 PubMed DOI PMC

Avants B. B., Tustison N. J., Song G., Cook P. A., Klein A., and Gee J. C., “A reproducible evaluation of ANTs similarity metric performance in brain image registration,” Neuroimage 54, 2033–2044 (2011).10.1016/j.neuroimage.2010.09.025 PubMed DOI PMC

Galloway M. M., “Texture analysis using gray level run lengths,” Comput. Graphics Image Process. 4, 172–179 (1975).10.1016/S0146-664X(75)80008-6 DOI

Yoo T. S., Ackerman M. J., Lorensen W. E., Schroeder W., Chalana V., Aylward S., Metaxas D., and Whitaker R., “Engineering and Algorithm Design for an Image Processing API: A Technical Report on ITK - The Insight Toolkit,” in Proc. of Medicine Meets Virtual Reality, edited byWestwood J. (IOS Press Amsterdam, 2002), pp. 586–592. PubMed

Tustison N. and Gee J., “Run-Length Matrices For Texture Analysis,” Insight J. 1, 1–6 (2008).

Haralick R. M., Shanmugam K., and Dinstein I. H., “Textural features for image classification,” IEEE Trans. Syst., Man, Cybern. 3, 610–621 (1973).10.1109/tsmc.1973.4309314 DOI

Hastie T., Tibshirani R., and Friedman J. H., The Elements of Statistical Learning: Data Mining, Inference, and Prediction, 2nd ed. (Springer, New York, NY, 2009).

Cortes C. and Vapnik V., “Support-vector networks,” Mach. Learn. 20, 273–297 (1995).10.1007/bf00994018 DOI

Breiman L., “Random forests,” Mach. Learn. 45, 5–32 (2001).10.1023/A:1010933404324 DOI

Bradley A. P., “The use of the area under the ROC curve in the evaluation of machine learning algorithms,” Pattern Recognit. 30, 1145–1159 (1997).10.1016/S0031-3203(96)00142-2 DOI

Robin X., Turck N., Hainard A., Tiberti N., Lisacek F., Sanchez J. C., and Muller M., “pROC: An open-source package for R and S plus to analyze and compare ROC curves,” BMC Bioinf. 12, 77–85 (2011).10.1186/1471-2105-12-77 PubMed DOI PMC

DeLong E. R., DeLong D. M., and Clarke-Pearson D. L., “Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach,” Biometrics 44, 837–845 (1988).10.2307/2531595 PubMed DOI

See supplementary material at http://dx.doi.org/10.1118/1.4948668 E-MPHYA6-43-012606 for imaging features that were qualitatively assessed for the two groups (MGMT methylated and unmethylated). DOI

Kassner A. and Thornhill R. E., “Texture analysis: A review of neurologic MR imaging applications,” Am. J. Neuroradiol. 31, 809–816 (2010).10.3174/ajnr.A2061 PubMed DOI PMC

Romano A., Calabria L. F., Tavanti F., Minniti G., Rossi-Espagnet M. C., Coppola V., Pugliese S., Guida D., Francione G., Colonnese C., Fantozzi L. M., and Bozzao A., “Apparent diffusion coefficient obtained by magnetic resonance imaging as a prognostic marker in glioblastomas: Correlation with MGMT promoter methylation status,” Eur. Radiol. 23, 513–520 (2013).10.1007/s00330-012-2601-4 PubMed DOI

Eckel-Passow J. E., Lachance D. H., Molinaro A. M., Walsh K. M., Decker P. A., Sicotte H., Pekmezci M., Rice T., Kosel M. L., Smirnov I. V., Sarkar G., Caron A. A., Kollmeyer T. M., Praska C. E., Chada A. R., Halder C., Hansen H. M., McCoy L. S., Bracci P. M., Marshall R., Zheng S., Reis G. F., Pico A. R., O’Neill B. P., Buckner J. C., Giannini C., Huse J. T., Perry A., Tihan T., Berger M. S., Chang S. M., Prados M. D., Wiemels J., Wiencke J. K., Wrensch M. R., and Jenkins R. B., “Glioma groups based on 1p/19q, IDH, and TERT promoter mutations in tumors,” N. Engl. J. Med. 372, 2499–2508 (2015).10.1056/NEJMoa1407279 PubMed DOI PMC

Ho A. J., Hua X., Lee S., Leow A. D., Yanovsky I., Gutman B., Dinov I. D., Lepore N., Stein J. L., Toga A. W., C. R. Jack, Jr., Bernstein M. A., Reiman E. M., Harvey D. J., Kornak J., Schuff N., Alexander G. E., Weiner M. W., Thompson P. M., and Alzheimer’s Disease Neuroimaging Initiative, “Comparing 3 T and 1.5 T MRI for tracking Alzheimer’s disease progression with tensor-based morphometry,” Hum. Brain Mapp. 31, 499–514 (2010).10.1002/hbm.20882 PubMed DOI PMC

Mayerhoefer M. E., Szomolanyi P., Jirak D., Materka A., and Trattnig S., “Effects of MRI acquisition parameter variations and protocol heterogeneity on the results of texture analysis and pattern discrimination: An application-oriented study,” Med. Phys. 36, 1236–1243 (2009).10.1118/1.3081408 PubMed DOI