View Presentation Video

Guohua Zhao¹, Yizhou Su², Yusong Lin², Jie Dong³, Eryuan Gao¹, Xiaoyue Ma¹, Jie Bai¹, Huiting Zhang⁴, Xu Yan⁴, Guang Yang⁵, and Jingliang Cheng^1
1The Department of Magnetic Resonance Imaging, The First Affiliated Hospital of Zhengzhou University, Zhengzhou, China, ²Collaborative Innovation Center for Internet Healthcare, Zhengzhou, China, ³School of Information Engineering, North China University of Water Resources and Electric Power, Zhengzhou, China, ⁴MR Scientific Marketing, Siemens Healthineers, Shanghai, China, ⁵Shanghai Key Laboratory of Magnetic Resonance, East China Normal University, Shanghai, China

Keywords: Tumors, Diffusion/other diffusion imaging techniques

Preoperative differentiation between glioblastomas (GBM) and solitary brain metastases (SBM) would aid in appropriate treatment planning and follow-up. Mean apparent propagator (MAP) MRI is effective in evaluating the inhomogeneity of brain microstructure. Texture analysis can be used to extract and quantify these inhomogeneities. In this study, we extracted the texture features from MAP-MRI, and validated the candidate features in discriminating between GBM and SBM. The RTAP model achieved the best discriminative power for the single metric model, and performed similarly to the MAP-MRI combined model. Texture analysis based on MAP-MRI has great potential to distinguish between the two entities.

View Presentation Video

Yan-Lin Liu¹, Ching-Chung Ko^2,3, Yang Zhang¹, Lee-Ren Yeh⁴, Jeon-Hor Chen¹, and Min-Ying Su^1
1Department of Radiological Sciences, University of California, Irvine, CA, United States, ²Department of Medical Imaging, Chi Mei Medical Center, Tainan, Taiwan, ³Department of Health and Nutrition, Chia Nan University of Pharmacy and Science, Tainan, Taiwan, ⁴Department of Radiology, E-DA Hospital, I-Shou University, Kaohsiung, Taiwan

Keywords: Tumors, Machine Learning/Artificial Intelligence

A subset of primary central nervous system lymphoma (PCNSL) may show early relapsed/refractory (R/R) disease after treatments. This study investigated the role of radiomics and machine learning for the prediction of R/R in PCNSL after treatments. Total 46 patients with pathologically confirmed PCNSL were included. Total 321 radiomic features were extracted from various pre-treatment MR sequences in each patient to build prediction models. Among various machine learning algorithms, the best predictive performance with accuracy of 82.6%, precision of 80%, and AUC of 0.85 were obtained in support vector machine.

View Presentation Video

Zong-Ze Chen¹, Yong-Han Lai¹, Ching-An Liao¹, Teng-Yi Huang², Ping-Hong Lai^1,3,4, and Tzu-Chao Chuang^1
1National Sun Yat-Sen University, Kaohsiung, Taiwan, ²National Taiwan University of Science and Technology, Taipei, Taiwan, ³Kaohsiung Veterans General Hospital, Kaohsiung, Taiwan, ⁴National Yang-Ming University, Taipei, Taiwan

Keywords: Tumors, Susceptibility

Susceptibility-weighted imaging (SWI) has shown its potential to discriminate between high-grade and low-grade astrocytoma.  In this study, we developed a fully automatic diagnosis system for astrocytoma grading by using convolutional neural network with contrast-enhanced T1-weighted images and SWI, separately or jointly, as input data.  The results show that the model with both imaging modalities as input data provides high accuracy in astrocytoma grading and is potentially helpful for clinical diagnosis.

View Presentation Video

Jinlong He¹, Yang Gao¹, Shaoyu Wang², and Huapeng Zhang^2
1Department of Imaging Diagnosis, The Affiliated Hospital of Inner Mongolia Medical University, Hohhot, China, ²MR Scientific Marketing, Siemens Healthineers, Shanghai, China

Keywords: Tumors, Radiomics

This study aimed to explore the value of MR multi-sequence radiomics combined with clinical features and genomics in predicting the survival of patients with glioma. Results showed that the clinical and imaging characteristics, radiomics features, and genotype status were important risk factors for glioma survival. The combination of multiple factors can better predict and evaluate the prognosis of glioma. Multi sequence based radiomics combined with clinical and imaging features and genotype status can better reflect the heterogeneity and prognosis of glioma.

Xinying Ren^1,2, Diaohan Xiong^1,2, Yujing Li^1,2, Kai Ai³, and Jing Zhang^1
1Lanzhou University Second Hospital, Lanzhou, China, ²Second Clinical School, Lanzhou University, Lanzhou, China, ³Philips Healthcare, Xi'an, China

Keywords: Tumors, Machine Learning/Artificial Intelligence, Radiomics

The study aimed to predict glioma genotypes combined with radiomic features and deep learning networks by using amide proton transfer (APT) imaging. The genetic subtypes of gliomas can be predicted by radiomics and deep learning networks using conventional MRI, however there are still problems with low accuracy and insufficient generalization. This study puts the screened APT radiomics features into a neural network and compares it with traditional radiomic. The results demonstrated that the proposed model had better performance. Therefore, APTw-derived radiomic features have good ability to predict 3-class molecular typing, providing novel classification tool for non-invasive evaluation for glioma genotypes.

View Presentation Video

Shao ru zhang¹, Zhi qiang Chen², Xiao hua Chen¹, Zhuo Wang¹, Shi li liu¹, Yun shu Zhou¹, Ruo di Zhang¹, Yu hui Xiong³, and Bing Chen^4
1Clinical medicine school of Ningxia Medical University, Yinchuan, China, ²Department of Radiology ,the First Hospital Affiliated to Hainan Medical College, Haikou, China, ³GE Healthcare, Beijing, China, ⁴Department of Radiology, General Hospital of Ningxia Medical University, Yinchuan, China

Keywords: Tumors, Brain

In this study, we aim to explore the radiomics model constructed by a single sequence or a combined sequences whether has the same or better performance, and to explore the effect of different choices of region of interest to the performance of radionmics model.

LIU YANLING¹, CUI YUELONG¹, NIU ZHEN¹, GUO JINXIA², ZHANG SHASHA¹, ZHANG JUNLI¹, and WANG YONG^1
1Anyang District Hospital, Anyang City, China, ²GE Healthcare, Beijing, China

Keywords: Tumors, Cancer

This study is to utilize the relaxometry generated by unenhanced and Gd-enhanced synthetic MRI to identify the recurrence and necrosis in patient with high grade glioma after radiotherapy. The results indicated enhanced T1 was significantly shortened in tumor and peripheral edema region for recurrent tissue in compare with necrosis and showed well ability of differentiation

View Presentation Video

Gulnur Semahat Ungan^1,2, Albert Pons-Escoda³, Daniel Ulinic², Carles Arus^1,2, Alfredo Vellido^1,4, and Margarida Julia-Sape^1,2
1Centro de Investigacion Biomedica en Red (CIBER), Cerdanyola del Valles, Spain, ²Universitat Autonoma de Barcelona, Cerdanyola del Valles, Spain, ³Hospital de Bellvitge, L'Hospitalet del Llobregat, Spain, ⁴Universitat Politecnica de Catalunya, Barcelona, Spain

Keywords: Tumors, Spectroscopy, visualization, data analysis, MRS, brain tumours

We used SV-MRS 1.5T data of patients with brain tumors to create colored-classification images of MV-MRS 3T grids of an independent cohort of patients. In 10 out of 15 MV cases the solid tumor region corresponded to the correct class. In the remaining 5 cases, the reasons for (partial) misclassification included heterogeneity and bad spectral quality.

Diaohan Xiong¹, Xinying Ren¹, Yujing Li¹, Rui Wang¹, Kai Ai², and Jing Zhang^1
1Lanzhou University Second Hospital, Lanzhou, China, ²Philips Healthcare, Xi'an, China

Keywords: Tumors, Machine Learning/Artificial Intelligence

The aim of this study was to test deep learning classification models of glioma subtypes using the generated images. GANs were created based on the two frameworks, pix2pix and cycleGAN. The source domain was T2 and the target domain was T1c, T2-FLAIR or ADC. The results demonstrated that the T2 to T1c pix2pix model has the highest PSNR and SSI. When only the T2-flair or T1c sequence is replaced with the generated image, the classification accuracy is same as the original image. Therefore, depending solely on T2 sequences, GANs networks could generate other sequences for Use in  gliomas classification Model.

View Presentation Video

Sena Azamat^1,2, Buse Buz-Yaluğ¹, Alpay Ozcan³, Ayça Ersen Danyeli^4,5,6, Necmettin Pamir^5,7, Alp Dinçer^4,8, Koray Ozduman^4,7, and Esin Ozturk-Isik^1,4
1Institute of Biomedical Engineering, Bogazici University, Istanbul, Turkey, ²Basaksehir Cam and Sakura City Hospital, Istanbul, Turkey, ³Electric and Electronic Engineering Department, Bogazici University, Istanbul, Turkey, ⁴Brain Tumor Research Group, Acibadem University, Istanbul, Turkey, ⁵Center for Neuroradiological Applications and Reseach, Acibadem University, Istanbul, Turkey, ⁶Department of Medical Pathology, Acibadem University, Istanbul, Turkey, ⁷Department of Neurosurgery, Acibadem University, Istanbul, Turkey, ⁸Department of Radiology, Acibadem University, Istanbul, Turkey

Keywords: Tumors, Machine Learning/Artificial Intelligence

Gliomas with IDH mutations tend to have a better prognosis regardless of the histopathological grade. The main aim of this study was to identify isocitrate dehydrogenase (IDH) mutations in glioma patients using deep learning models based on SWI, FLAIR and CE-T1W images separately and together. As a result, a 2D CNN model based on multiparametric MRI resulted in an accuracy of 92.8%, while CNN based on just SWI had 72.2% and CNN based on just FLAIR had 61.1% accuracies for predicting IDH mutation in gliomas.

View Presentation Video

Anshika Kesari¹, Virendra Kumar Yadav¹, Raufiya Jafari¹, Rakesh Kumar Gupta², Rana Patir³, Sandeep Vaishya³, Sunita Ahlawat⁴, and Anup Singh^1,5,6
1Centre for Biomedical Engineering, Indian Institute of Technology Delhi, New Delhi, India, ²Department of Radiology, Fortis Memorial Research Institute, Gurugram, India, ³Department of Neurosurgery, Fortis Memorial Research Institute, Gurugram, India, ⁴SRL Diagnostics, Fortis Memorial Research Institute, Gurugram, India, ⁵Yardi School for Artificial Intelligence, Indian Institute of Technology Delhi, New Delhi, India, ⁶Department of Biomedical Engineering, All India Institute of Medical Sciences, New Delhi, India

Keywords: Tumors, Brain

Quantitative DCE-MRI parameters are sensitive to tumor microvasculature and have shown potential in tumor grading. Generally, value of some of these parameters increases with tumor grades. During analysis of tumor region, large-blood-vessels (LBV) can influence the actual changes due to tumor microvasculature and hence provide erroneous results. Objectives of this study were to develop an automatic framework for segmentation of the LBV present within and outside of the tumor region and to reduce errors in tumor grading. The statistical analysis of parameters in tumor region showed significant differences between with and without LBV. Moreover, LBV removal improved tumor classification accuracy.

View Presentation Video

Yuankui Wu¹, Zhousan Huang¹, Yingxia Huang¹, Mingxi Li¹, Haimei Cao¹, Guanglong Huang², Hao Zhang¹, and Yikai Xu^1
1Department of Medical Imaging, Nanfang Hospital, Southern Medical University, Guangzhou, China, ²Department of Neurosurgery, Nanfang Hospital, Southern Medical University, Guangzhou, China

Keywords: Tumors, Perfusion, Inflow-based vascular space occupancy (iVASO)

Early detection of disease progression is of important relevance for the management of high-grade glioma (HGG) patients. Related studies showed that perfusion-weighted MR imaging (PWI) has power to detect early recurrence of glioblastoma. Inflow-based vascular space occupancy (iVASO) is a noninvasive perfusion technology that can provide absolute blood volume of precapillary arterioles (arteriolar blood volume, BVa). In this preliminary study, the potential value of BVa in detecting subclinical recurrence of HGG was investigated. The results showed that the histogram features of BVa might have the potential to detect subclinical recurrence.

View Presentation Video

George Russell Glenn¹, Sneha Sai Venkata Ka Mannam², Chibueze Nwagwu³, Subir Goyal⁴, Gustavo Pradilla², Edjah Kweku-Ebura Nduom², Jeffery James Olson ², David Painton Bray ², and Hoang Bojanowski Kimberly^2
1Diagnostic Radiology, Emory University, Lilburn, GA, United States, ²Neurosurgery, Emory University, Atlanta, GA, United States, ³College of Medicine, Emory University, Atlanta, GA, United States, ⁴Biostatistics and Bioinformatics, Emory University, Alanta, GA, United States

Keywords: Tumors, Surgery

Preoperative diffusion tensor imaging (DTI) data was analyzed using Automated Fiber Quantification (AFQ) along the corticospinal tract (CST) for 58 patients undergoing surgical resection of primary or metastatic brain lesions.  For each patient, DTI parameters were analyzed along the ipsilateral CST, the side of the pathological lesion, and the contralateral CST, the side opposite the pathological lesion. Patients were then separated based on the presence or absence of postoperative motor weakness. Patients with post operative motor weakness were found to have significantly different diffusion parameters along their ipsilateral CST compared to the ipsilateral CST of patients without postoperative motor weakness. 

Hongbo Zhang¹, Hanwen Zhang², Beibei Zhou³, Yuze Zhang¹, Lei Wu¹, Yi Lei², and Biao Huang^1
1Guangdong Provincial Key Laboratory of Artificial Intelligence in Medical Image Analysis and Application, Guangdong Provincial People's Hospital, Guangdong Academy of Medical Sciences, Guangzhou, China, ²Department of Radiology, The First Affiliated Hospital of Shenzhen University, Health Science Center, Shenzhen Second People's Hospital, Shenzhen, China, ³Department of Radiology, Department of Radiology,The Seventh Affiliated Hospital, Sun Yat-sen University, Shenzhen, China

Keywords: Tumors, Radiomics

Radiomics uses computer software to mine massive quantitative image features from medical imaging images and then screens the most valuable radiomics features using statistical and/or machine learning methods. Furthermore, it is used to parse clinical information for disease characterisation, tumour grading and staging, efficacy evaluation, and prognosis prediction. In our study, we demonstrated that multiparametric MRI-based fusion radiomics model is an effective preoperative non-invasive method to predict telomerase reverse transcriptase promoter mutations and progression-free survival in glioblastoma  patients.

View Presentation Video

Dennis C. Thomas^1,2,3,4, Ralf Deichmann⁵, Elke Hattingen^1,2,3,4, Seyma Alcicek^1,2,3,4, Ulrich Pilatus^1,2,3,4, and Katharina J. Wenger^1,2,3,4
1Institute of Neuroradiology, University Hospital Frankfurt, Frankfurt, Germany, ²University Cancer Center Frankfurt (UCT), Frankfurt, Frankfurt, Germany, ³Frankfurt Cancer Institute (FCI), Frankfurt, Germany, ⁴German Cancer Research Center (DKFZ) and German Cancer Consortium (DKTK), Heidelberg, Germany, ⁵5 Brain Imaging Center, Goethe University, Frankfurt, Germany

Keywords: Tumors, Brain, Water content mapping, T1 mapping

Quantitative MRI was applied to assess T1 and water content (H₂O) in brain tumors. 1/T1 and 1/H₂O are linearly related in healthy WM and GM. Here, we explore deviations from this linear relationship in the tumor regions in 3 glioblastoma patients. Linear regression between 1/T1 and 1/H₂O was performed in the healthy brain in the contralateral hemisphere and the obtained values were used to calculate predicted 1/H₂O maps for the whole brain. Difference maps between predicted and true 1/H₂O were calculated . Further, the importance of robust bias field correction techniques is demonstrated here by comparing two bias field correction approaches.

View Presentation Video

Abdullah Bas¹, Banu Sacli-Bilmez¹, Ayca Ersen Danyeli^2,3, Ozge Can^2,4, Koray Ozduman^2,5, Alp Dincer^2,6, and Esin Ozturk-Isik^1,2
1Institute of Biomedical Engineering, Bogazici University, Istanbul, Turkey, ²Brain Tumor Research Group, Acibadem University, Istanbul, Turkey, ³Department of Medical Pathology, Acibadem University, Istanbul, Turkey, ⁴Department of Biomedical Engineering, Acibadem University, Istanbul, Turkey, ⁵Department of Neurosurgery, Acibadem University, Istanbul, Turkey, ⁶Department of Radiology, Acıbadem University, Istanbul, Turkey

Keywords: Tumors, Machine Learning/Artificial Intelligence, Deep Learning

Isocitrate dehydrogenase (IDH) and telomerase reverse transcriptase promoter (TERTp) mutations affect the clinical behavior and survival rate of diffuse gliomas. According to the latest WHO 2021 brain tumor classification, IDH mutation is an important factor for grouping adult-type diffuse gliomas. The preoperative detection of these mutations is very critical for treatment planning. In this study, we propose enhanced 1D-CNN models by adding an attention mechanism as a prior network to focus on relevant spectral frequencies of 1H-MRS to identify IDH-mutant (IDH-mut), TERTp-mutant (TERTp-mut), and IDH-wt, TERTp-mut (TERTp-only) gliomas using three binary models

View Presentation Video

Natasha Najam¹, Steven Tobochnik², Huijin Liao¹, Sanghoon Kim¹, Jong Woo Lee², and Alexander Lin^1
1Center for Clinical Spectroscopy, Department of Radiology, Brigham and Women's Hospital , Harvard Medical School, Boston, MA, United States, ²Department of Neurology, Brigham and Women's Hospital , Harvard Medical School,, Boston, MA, United States

Keywords: Tumors, Spectroscopy, Brain Tumor-Related Epilepsy, Isocitrate Dehydrogenase Mutation

Brain Tumor Related Epilepsy (BTRE) is a multi-factorial condition with unclear pathophysiology. It has a robust prevalence in tumor patients with 20-40% experiencing onset seizures and a further 20-45% experiencing refractory seizures to treatment.¹ This work examines the tumor microenvironment using magnetic resonance spectroscopy (MRS) to learn more about the epileptogenic mechanisms through a comparison of metabolite composition in BTRE patients and non-seizure tumor patients. Increasing knowledge about the pathophysiology of BTRE might aid the management of anti-tumor and BTRE treatment in the future. 

View Presentation Video

Samuel A Bobholz¹, Allison K Lowman¹, Savannah R Duenweg², Aleksandra Winiarz², Margaret Stebbins², Fitzgerald Kyereme¹, Jennifer Connelly³, Dylan Coss⁴, Wade M Mueller⁵, Mohit Agarwal¹, Anjishnu Banerjee⁶, and Peter S LaViolette^1,7
1Radiology, Medical College of Wisconsin, Milwaukee, WI, United States, ²Biophysics, Medical College of Wisconsin, Milwaukee, WI, United States, ³Neurology, Medical College of Wisconsin, Milwaukee, WI, United States, ⁴Pathology, Medical College of Wisconsin, Milwaukee, WI, United States, ⁵Neurosurgery, Medical College of Wisconsin, Milwaukee, WI, United States, ⁶Biostatistics, Medical College of Wisconsin, Milwaukee, WI, United States, ⁷Biomedical Engineering, Medical College of Wisconsin, Milwaukee, WI, United States

Keywords: Tumors, Cancer, Machine learning, glioblastoma, radio-pathomics

This study applied autopsy-based radio-pathomic maps to the pre-surgical PENN-GBM dataset to test the hypothesis that the predicted tumor composition of the contrast-enhancing and FLAIR-hyperintense regions identify distinct pathological features of glioblastoma. We find that greater predicted tumor within the contrast-enhancing region is indicative of IDH1-wildtype mutation status, and show that larger tumors tend to have less predicted tumor within contrast-enhancement and more tumor within non-enhancing FLAIR hyperintensity. This technique could be used to non-invasively identify more aggressive tumors.

View Presentation Video

Bruno Madore¹, Jeffrey P. Guenette¹, Jonathan D. Schoenfeld¹, and Evangelia Kaza^1
1BWH, Harvard Medical School, Boston, MA, United States

Keywords: Head & Neck/ENT, Relaxometry, Radiotherapy, treatment, tumor

Radiation oncology patients often undergo longitudinal scans to evaluate the effect of treatments on lesions, and to monitor surrounding tissues. We are considering the inclusion of quantitative MRI in current imaging protocols. To this end, we implemented a simple quantitative imaging technique and tested it in a small cohort of five patients. Results captured differences in T1 and T2 between tissues on the treated side vs. the contralateral tissues. Little work has been done so far with MRI relaxometry for monitoring treatment in head and neck squamous cell carcinoma, and this work represents an early step in this direction.

View Presentation Video

Abdullah Bas¹, Ayca Ersen Danyeli^2,3, Ozge Can^2,4, Koray Ozduman^2,5, Alp Dincer^2,6, and Esin Ozturk-Isik^1,2
1Institute of Biomedical Engineering, Bogazici University, Istanbul, Turkey, ²Brain Tumor Research Group, Acibadem University, Istanbul, Turkey, ³Department of Medical Pathology, Acibadem University, Istanbul, Turkey, ⁴Department of Biomedical Engineering, Acibadem University, Istanbul, Turkey, ⁵Department of Neurosurgery, Acibadem University, Istanbul, Turkey, ⁶Department of Radiology, Acıbadem University, Istanbul, Turkey

Keywords: Tumors, Machine Learning/Artificial Intelligence, Deep Learning

NF2-L in meningiomas is a relevant indicator of prognosis. NF2-L is one of the most common genetic mutations in meningiomas. As a result of that situation, developing a non-invasive approach may assist the current clinical procedures. To our knowledge, some studies try to predict NF2-L using MRI modalities but either they use registration or performed using the modalities of MRI that are not in default MRI scanning procedures.  Hence, the solutions that they provide are not clinically feasible. In this study, we aim to develop new approaches that are capable to implement directly into clinical procedures.[1]