Manabu Kinoshita^1,2, Masato Uchikoshi³, Souichiro Tateishi², Shohei Miyazaki², Mio Sakai², Tomohiko Ozaki², Katsunori Aasai², Yuya Fujita², Takahiro Matsuhashi², Yonehiro Kanemura⁴, Eku Shimosegawa¹, Jun Hatazawa¹, Shin-ichi Nakatsuka², Haruhiko Kishima¹, and Katsuyuki Nakanishi^2
1Osaka University Graduate School of Medicine, Suita, Japan, ²Osaka International Cancer Institute, Osaka, Japan, ³Canon Medical Systems Corporation, Tochigi, Japan, ⁴National Hospital Organization Osaka National Hospital, Osaka, Japan

The authors attempted to elucidate the correlation of tumor cell density within the brain in glioma patients and T1-, T2-relaxometry and ADC. First, an exploratory study compared T1-, T2-relaxometry, and ADC with 11C-methionine PET followed by a validation study using intraoperative stereo-tactically obtained tissues. A range of T1 values indicative of high cell density was identified, which finding was confirmed by stereotactic tissue sampling. For T2 values and ADC, however, no statistically significant correlation was confirmed regarding tumor cell density. The proposed technique was further able to create predictive tumor cell density map by T1 supplemented by T2 values.

Jason Glenn Parker¹, Emily E Diller², Sha Cao³, Jeremy T Nelson⁴, Kristen Yeom⁵, Chang Ho¹, and Robert Lober^6
1Radiology & Imaging Sciences, Indiana University School of Medicine, Indianapolis, IN, United States, ²School of Health Sciences, Purdue University, West Lafayette, IN, United States, ³Biostatistics, Indiana University School of Medicine, Indianapolis, IN, United States, ⁴Military Health Institute, University of Texas Health San Antonio, San Antonio, TX, United States, ⁵Neuroradiology, Lucile Salter Packard Children’s Hospital and Stanford University Medical Center, Palo Alto, CA, United States, ⁶Neurosurgery, Dayton Children's Hospital, Dayton, OH, United States

We demonstrate statistical relationships between routine multiparametric imaging signatures and underlying cellular and molecular properties of brain tumors. We apply advanced statistical methods to correct for the family-wise error rate problem associated with whole-brain statistical parametric mapping, and show that the results have strong agreement with surgical biopsy. These results imply that cellular and molecular mapping of tumor heterogeneity from minimally-invasive images may be possible in the near future.

Prateek Kalra¹, Varun Varadarajan ², Michael S Harris³, Ashonti Harper³, Omar Mohamed³, Oliver Adunka², Daniel M. Prevedello⁴, and Arunark Kolipaka^1
1Radiology, Ohio State University Wexner Medical Center, Columbus, OH, United States, ²Otolaryngology, Ohio State University Wexner Medical Center, Columbus, OH, United States, ³Ohio State University Wexner Medical Center, Columbus, OH, United States, ⁴Neurological Surgery, Ohio State University Wexner Medical Center, Columbus, OH, United States

Microsurgery in brain tumor patients aim to complete tumor resection without compromising neurological functionality. Inadequate preoperative knowledge of tumor may prolong surgical time and increase risk of postoperative complications which may depends upon tumor-brain adhesion and tumor stiffness. Previous studies have only looked into adhesion and stiffness separately using Magnetic Resonance Elastography (MRE). Previously, we proposed both adhesion and stiffness in vestibular schwannoma patients. Aim of this study is to also include meningioma cases in the analysis in order to broaden the tumor types and complexity. Preliminary results show good correlation between MRE-derived preoperative assessment of tumor and surgical findings.

Ziying Yin¹, Xin Lu¹, Salomon Cohen Cohen², Yi Sui¹, Armando Manduca³, Jamie J Van Gompel², Richard L. Ehman¹, and John III Huston^1
1Radiology, Mayo Clinic, Rochester, MN, United States, ²Neurosurgery, Mayo Clinic, Rochester, MN, United States, ³Physiology and Biomedical Engineering, Mayo Clinic, Rochester, MN, United States

Brain tumor adherence has been long recognized to impact surgical resection difficulty. Recently-developed slip interface imaging (SII) can preoperatively predict tumor-brain adhesion. In previous studies, subjectively-determined SII assessment of tumor adhesion has been shown to agree well with intraoperative findings. The purpose of this work was to develop an objective quantitative method for analyzing SII data for adherence, thereby minimizing inter- and intraobserver variability. We developed a radiomics-based metric (termed “adhesion degree”) based on SII to quantify the degree of tumor adhesion. In 46 meningiomas, the adhesion degree showed excellent accuracy in predicting completely adherent tumors (AUROC=0.96) from non-adherent tumors.

Kiyohisa Kamimura¹, Masanori Nakajo¹, Tomohide Yoneyama¹, Manisha Bohara¹, Yoshihiko Fukukura¹, Shingo Fujio², Takashi Iwanaga³, Hiroshi Imai⁴, Marcel Dominik Nickel⁵, and Takashi Yoshiura^1
1Radiology, Kagoshima University Graduate School of Medical and Dental Sciences, Kagoshima, Japan, ²Neurosurgery, Kagoshima University Graduate School of Medical and Dental Sciences, Kagoshima, Japan, ³Clinical Engineering Department Radiation Sectio, Kagoshima University Hospital, Kagoshima, Japan, ⁴Siemens Healthcare K.K., Tokyo, Japan, ⁵Siemens Healthcare, Erlangen, Germany

Preoperative information of tumor consistency is important in patients with pituitary adenoma. Our aim was to evaluate the possible role of high-temporal and spatial resolution dynamic contrast enhanced MR imaging (DCE-MRI) and quantitative pharmacokinetic analysis in differentiation of a hard adenoma from soft adenoma. The hard adenoma showed significantly higher extravascular extracellular space per unit volume of tissue (v_e) than soft adenoma and the v_e was significantly correlated to collagen IV content of pituitary adenomas. This quantitative pharmacokinetic parameter v_e may be useful for differentiation of the hard adenoma from soft adenoma.

Laura C. Bell¹, Natenael B. Semmineh¹, Leland S. Hu², Yuxiang Zhou², Melissa Prah³, Kathleen M. Schmainda³, Jerrold L. Boxerman⁴, Hongyu An⁵, Cihat Eldeniz⁵, Richard Wahl⁵, Bradley Erickson⁶, Panagiotis Korfiatis⁶, Chengyue Wu⁷, Thomas Yankeelov⁷, Anna Sorace⁷, Neal Rutledge⁷, Thomas Chenevert⁸, Dariya Malyarenko⁸, Yichu Liu⁹, Andrew Brenner⁹, Yi-Fen Yen¹⁰, Jayashree Kalpathy-Cramer¹⁰, Andrew Beers¹⁰, Mark Muzi¹¹, Ananth J. Madhuranthakam¹², Marco Pinho¹², Brian Johnson¹², and C. Chad Quarles^1
1Barrow Neurological Institute, Phoenix, AZ, United States, ²Mayo Clinic, Scottsdale, AZ, United States, ³Medical College of Wisconsin53226, Milwaukee, WI, United States, ⁴Rhode Island Hospital and Alpert Medical School of Brown University, Providence, RI, United States, ⁵Washington University in St. Louis, St. Louis, MO, United States, ⁶Mayo Clinic, Rochester, MN, United States, ⁷The University of Texas at Austin, Austin, TX, United States, ⁸University of Michigan, Ann Arbor, MI, United States, ⁹University of Texas Health Science Center at San Antonio, San Antonio, TX, United States, ¹⁰Massachusetts General Hospital, Boston, MA, United States, ¹¹University of Washington, Seattle, WA, United States, ¹²The University of Texas Southwestern, Dallas, TX, United States

Using a dynamic susceptibility contrast (DSC) MRI DRO we previously characterized brain tumor relative cerebral blood volume (rCBV) reproducibility across 12 sites employing a range of imaging protocols and software platforms. Our goal in this study is to determine the impact of rCBV reproducibility for tumor grade classification. We found that varying software platforms produced a range of optimal thresholds, but the performance of these thresholds were similar. These results indicate that different software platforms are able to classify tumor grades, but the site-specific thresholds underscore the importance of standardizing acquisition and analysis protocols across sites and software benchmarking.

Siqi Cai¹, Zhifeng Shi², Shihui Zhou^1,3, Chunxiang Jiang^1,3, and Lijuan Zhang^1
1Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China, ²Huashan Hospital of Fudan University, Shanghai, China, ³University of Chinese Academy of Science, Beijing, China

The systemic oscillation of the blood flow in cerebrum and cerebellum may vary with gliomas of different malignancy, resulting in neurovascular uncoupling and abnormal perfusion. In this study, we investigated the implication of glioma on the oscillation of cerebral blood flow based on time-shifted rs-fMRI. HGGs induces more widely alterations in the spontaneous fluctuations of cerebral and cerebellar blood flow at the global scale. Vascular oscillation changes derived from rs-fMRI may provide a novel insight for the assessment of the functional plasticity and its clinical relevance in the interpretation of the psychological and psychiatric symptoms in subjects with glioma.

Franklyn Howe¹, Timothy Jones², Philip Rich², Jordan Colman³, Guang Yang⁴, Felix Raschke⁵, Venus Liang¹, Alex Denley¹, and Thomas Barrick^1
1Neurosciences Research Centre, St George's, University of London, London, United Kingdom, ²St George's University Hospitals NHS Foundation Trust, London, United Kingdom, ³Ashford and St Peter's Hospitals NHS Foundation Trust, Surrey, United Kingdom, ⁴National Lund and Heart Institute, Imperial College, London, United Kingdom, ⁵OncoRay—National Center for Radiation Research in Oncology, Dresden, Germany

1H MRS and DTI measures were assessed for their ability to predict the future growth and malignant transformation of low-grade gliomas. The tumour core NAA concentration and the mean diffusivity (MD) within the MRS voxel, combined with the FLAIR tumour volume, provided a good predictor of tumours with higher growth rates. A ROC analysis gave an AUC of 0.86 to predict tumours likely to undergo malignant progression, and AUC of 0.98 when including those undergoing early debulking. The combined NAA, MD and volumetric parameter provided a single time-point assessment of future growth characteristics.

Yan Tan^1,2, Wei Mu³, Xiaochun Wang^1,2, Guoqiang Yang^1,2, Robert James Gillies³, and Hui Zhang^1,2
1Department of Radiology, First Hospital of Shanxi Medical University, Taiyuan, China, ²College of Medical Imaging, Shanxi Medical University, Taiyuan, China, ³Department of Cancer Physiology, H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL, United States

Assessment of glioma genotypes by quantifying MR diffusion imaging heterogeneity of whole tumour may serve as a powerful tool to instruct therapeutic decision-making. This study evaluated the role and incremental value of whole-tumor radiomics analysis based on DKI and DTI images in determining the IDH and MGMTmet genotypes of astrocytomas. A radiomics models based on whole-tumor MK and MD maps showed good diagnostic efficiency in predicting IDH and MGMTmet genotypes. Furthermore, the combined model constructed by radiomics score, edema degree and age further improved the performance of predicting IDH, while the combined model did not benefit for MGMTmet prediction.

Samuel Bobholz¹, Allison Lowman², Alexander Barrington³, Michael Brehler², Sean McGarry¹, Jennifer Connelly⁴, Elizabeth Cochran⁵, Anjishnu Banerjee⁶, and Peter LaViolette^2,3
1Biophysics, Medical College of Wisconsin, Wauwatosa, WI, United States, ²Radiology, Medical College of Wisconsin, Wauwatosa, WI, United States, ³Biomedical Engineering, Medical College of Wisconsin, Wauwatosa, WI, United States, ⁴Neurology, Medical College of Wisconsin, Wauwatosa, WI, United States, ⁵Pathology, Medical College of Wisconsin, Wauwatosa, WI, United States, ⁶Biostatistics, Medical College of Wisconsin, Wauwatosa, WI, United States

This study sought to provide a biological basis of radiomics-based analyses by assessing the relationship between MR features and analogous histomic features of the underlying tissue using coregistered histology samples taken at autopsy from brain cancer patients.  Several radiomic features demonstrated substantial mutual information with their histomic analogs, with first order features showing the strongest associations.  These histomic-preserving features were shown to be stable across potential confounds such as differences in scanner vender and acquisition field strength.  These findings suggest that MR radiomic features reflect information about the texture of the underlying tissue.