cancers Article Fully Automated MR Based Virtual Biopsy of Cerebral Gliomas Johannes Haubold 1, *, René Hosch 1 , Vicky Parmar 1 , Martin Glas 2 , Nika Guberina 3 , Onofrio Antonio Catalano 4 , Daniela Pierscianek 5 , Karsten Wrede 5 , Cornelius Deuschl 1 , Michael Forsting 1 , Felix Nensa 1 , Nils Flaschel 1,† and Lale Umutlu 1,†   Citation: Haubold, J.; Hosch, R.; Parmar, V.; Glas, M.; Guberina, N.; Catalano, O.A.; Pierscianek, D.; Wrede, K.; Deuschl, C.; Forsting, M.; et al. Fully Automated MR Based Virtual Biopsy of Cerebral Gliomas. Cancers 2021, 13, 6186. https:// doi.org/10.3390/cancers13246186 Academic Editor: Ahmed Idbaih Received: 10 November 2021 Accepted: 28 November 2021 Published: 8 December 2021 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional afﬁl- iations. Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). 1 Department of Diagnostic and Interventional Radiology and Neuroradiology, University Hospital Essen, Hufelandstr. 55, D-45147 Essen, Germany; rene.hosch@uk-essen.de (R.H.); vicky.parmar@uk-essen.de (V.P.); Cornelius.Deuschl@uk-essen.de (C.D.); Michael.Forsting@uk-essen.de (M.F.); Felix.Nensa@uk-essen.de (F.N.); Nils.Flaschel@uk-essen.de (N.F.); lale.umutlu@uk-essen.de (L.U.) 2 Department of Neurology, Division of Clinical Neurooncology, University Hospital Essen, D-45147 Essen, Germany; Martin.Glas@uk-essen.de 3 Department of Radiotherapy, University Hospital Essen, D-45147 Essen, Germany; nika.guberina@uk-essen.de 4 Department of Radiology, Division of Abdominal Imaging, A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Harvard University Medical School, Boston, MA 02114, USA; ocatalano@mgh.harvard.edu 5 Department of Neurosurgery, University Hospital Essen, D-45147 Essen, Germany; Daniela.Pierscianek@uk-essen.de (D.P.); Karsten.Wrede@uk-essen.de (K.W.) * Correspondence: johannes.haubold@uk-essen.de; Tel.: +49-201-723-84528; Fax: +49-201-723-1548 † Contributed equally. Simple Summary: Over the past few years, radiomics-based tissue characterization has demon- strated its potential for non-invasive prediction of the genetic proﬁle and grading in cerebral gliomas using multiparametric MRI. The aim of our study was to investigate the feasibility and diagnostic accuracy of a fully automated radiomics analysis based on a simpliﬁed MR protocol derived from various scanner systems to prospectively ease the transition of radiomics-based non-invasive tissue sampling into clinical practice. Using an MRI with non-contrast and post-contrast T1-weighted sequences and FLAIR, our workﬂow automatically predicts the IDH1/2 mutation, the ATRX expres- sion loss, the 1p19q co-deletion and the MGMT methylation status. It also effectively differentiates low-grade from high-grade gliomas. In summary, the present study demonstrated that a fully auto- mated prediction of grading and the genetic proﬁle of cerebral gliomas could be performed with our proposed method using a simpliﬁed MRI protocol that is robust to variations in scanner systems, imaging parameters and ﬁeld strength. Abstract: Objective: The aim of this study was to investigate the diagnostic accuracy of a radiomics analysis based on a fully automated segmentation and a simpliﬁed and robust MR imaging pro- tocol to provide a comprehensive analysis of the genetic proﬁle and grading of cerebral gliomas for everyday clinical use. Methods: MRI examinations of 217 therapy-naïve patients with cerebral gliomas, each comprising a non-contrast T1-weighted, FLAIR and contrast-enhanced T1-weighted sequence, were included in the study. In addition, clinical and laboratory parameters were incor- porated into the analysis. The BraTS 2019 pretrained DeepMedic network was used for automated segmentation. The segmentations generated by DeepMedic were evaluated with 200 manual seg- mentations with a DICE score of 0.8082 ± 0.1321. Subsequently, the radiomics signatures were utilized to predict the genetic proﬁle of ATRX, IDH1/2, MGMT and 1p19q co-deletion, as well as differentiating low-grade glioma from high-grade glioma. Results: The network provided an AUC (validation/test) for the differentiation between low-grade gliomas vs. high-grade gliomas of 0.981 ± 0.015/0.885 ± 0.02. The best results were achieved for the prediction of the ATRX expression loss with AUCs of 0.979 ± 0.028/0.923 ± 0.045, followed by 0.929 ± 0.042/0.861 ± 0.023 for the prediction of IDH1/2. The prediction of 1p19q and MGMT achieved moderate results, with AUCs of 0.999 ± 0.005/0.711 ± 0.128 for 1p19q and 0.854 ± 0.046/0.742 ± 0.050 for MGMT. Conclusion: This fully automated approach utilizing simpliﬁed MR protocols to predict the genetic proﬁle and grading of cerebral gliomas provides an easy and efﬁcient method for non-invasive tumor decoding. Cancers 2021, 13, 6186. https://doi.org/10.3390/cancers13246186 https://www.mdpi.com/journal/cancers