View Presentation Video

Manqian Huang¹, Kan Deng², Hui Xie¹, Wenjie Huang¹, Xiaoyi Wang³, Chao Luo¹, Shuqi Li¹, Chunyan Cui¹, Huali Ma¹, Lizhi Liu¹, and Haojiang Li^1
1Sun Yat-sen University Cancer Center, Guangzhou, China, ²Philips Healthcare, Guangzhou, China, ³Hainan General Hospital (Hainan Affiliated Hospital of Hainan Medical University), Haikou, China

Keywords: Radiomics, Cancer, Nasopharyngeal Carcinoma

     The purpose of this study was to assess the performance of radiomic models based on MR-determined metastatic lymph nodes phenotype in predicting the prognosis of nasopharyngeal carcinoma (NPC) with three different feature selection methods: all lymph nodes (ALN), the largest lymph node (LLN), and the largest slice of the largest lymph node (LSLN). The results showed that LSLN radiomic features showed better accuracy in predicting the overall survival (OS).

View Presentation Video

jing song¹, hui qian zong¹, ya zhang¹, jing wang¹, hong yang wei¹, and li zhi xie^2
1The Second Hospital of Hebei Medical University, Shijiazhuang, China, ²GE Healthcare, Beijing, China

Keywords: Radiomics, Brain, Magnetic resonance spectroscopy

Question: Among the studies using radiomics methods to predict glioma grading, most of them are based on conventional magnetic resonance imaging sequences, and functional magnetic resonance imaging is less studied.

Methods: This study predicted malignant glioma grading based on magnetic resonance structural images and magnetic resonance spectroscopy using an radiomics approach.

Results: The test set AUC of the model constructed based on T1-enhanced images and the ratio of three metabolites of MRS was 0.95.

Conclusion: Radiomics based on T1-CE and MRS has a good performance in identifying both grade III and grade IV gliomas.

View Presentation Video

Jingpu Wu^1,2, Yiqing Shen^1,3, Qianqi Huang³, Pengfei Guo^1,3, Jinyuan Zhou¹, and Shanshan Jiang^1
1Department of Radiology, School of Medicine, Johns Hopkins University, Baltimore, MD, United States, ²Department of Applied Mathematics and Statistics, Whiting School of Engineering, Johns Hopkins University, Baltimore, MD, United States, ³Department of Computer Science, Whiting School of Engineering, Johns Hopkins University, Baltimore, MD, United States

Keywords: Radiomics, Radiomics

Sensitivity encoding (SENSE) is a conventional practice for accelerating APTw image acquisition. To achieve an even higher acceleration, SENSE with compressed sensing (CS-SENSE) was introduced. However, its effect on the radiomic features extracted from the APTw images was yet studied. Here we extracted radiomic features from both SENSE- and CS-SENSE-APTw images and evaluated their reliability in different regions of interest (ROIs). Moreover, filters play an important part in emphasizing specific image characteristics in radiomics. The impact of filters on the radiomic features were also discussed. Our results provided a comprehensive reference for radiomic analyses where CS is implemented for acceleration.

View Presentation Video

Yihang Zhou¹, Jing Yuan¹, Cindy Xue¹, Bin Yang², Kin Yin Cheung², and Siu Ki Yu^2
1Research Department, Hong Kong Sanatorium and Hospital, Hong Kong, China, ²Medical Physics Department, Hong Kong Sanatorium and Hospital, Hong Kong, China

Keywords: Radiomics, Prostate, MRCAT, synthetic-CT, reproducibility

MR-guided radiotherapy and radiomics have gained considerable attention. Synthetic-CT (sCT) derived from MRI has been adopted to facilitate MRI-only radiotherapy planning. Researches have evaluated the sCT both qualitatively and quantitatively. However, few studies have looked into the potential that sCT offers at radiomics level. We hypothesize that sCT generated from MR can faithfully reproduce radiomics features compared to those extracted from true planning-CT. We aim to investigate the reproducibility of radiomics features derived from a commercially available sCT generation (MRCAT) acquired on a 1.5T MR-simulator to those obtained from a CT simulation scan in a cohort of prostate cancer patients.

View Presentation Video

Yan Tan¹, Dawei Tian¹, Wenqiao Zheng¹, Xiaochun Wang¹, and Hui Zhang^1
1Department of Radiology, First Hospital of Shanxi Medical University, Taiyuan, China

Keywords: Radiomics, Diffusion/other diffusion imaging techniques

For investigating the usefulness of radiomics signature based on diffusion kurtosis imaging in overall survival prediction for astrocytoma, radiomics features extracted from whole-tumor on MK and MD images were selected to construct radiomics signature. Then the radiomics signature was integrated with clinical-genetic risk factors to develop a combined model, which was represented as nomogram. The results showed that the radiomics signature have offered clinically relevant prognostic information for astrocytoma and further stratified patients into different risk groups. The combined model achieved the highest predictive performance and facilitated clinical decision-making by nomogram, demonstrating the incremental value of radiomics signature.

Xiaoqian Bian¹, Zhibin Yue², Siyao Du¹, Yan Xu³, Yang Song³, Min Zhao⁴, and Lina Zhang^1
1The First Hospital of China Medical University, Shenyang, China, ²Department of Biomedical Engineering, School of Intelligent Medicine, China Medical University, Shenyang, China, ³MR Scientific Marketing, Siemens Healthineers, Shanghai, China, ⁴Pharmaceutical Diagnostics, GE Healthcare, Beijing, China

Keywords: Radiomics, Breast

Recently, HER2-low expression tumor was proposed as a new entity in the field of breast cancer and radiomics studies related to HER2-low are rare. This study investigated whether multiparametric MRI-based radiomics can distinguish HER2- low from HER2-positive and HER2-zero in breast cancer. Results showed that clinico-radiomics nomograms including hormone receptor status and radiomics signatures combined contrast-enhanced T1-weighted and apparent diffusion coefficient map performed best in both prediction tasks: HER2-positive vs. HER2-negative and HER2-low vs. HER2-zero. This suggests that multiparametric MRI radiomics achieved effective prediction of HER2-low, which might be further guidance of the anti-HER2 targeted therapy in breast cancer.

Ai Shuangquan^1,2, Peng Wei¹, He Yaoyao¹, Zhang Huiting³, Grimm Robert⁴, Zhang Zhaoxi¹, Peng Lin⁵, Liu Yulin¹, and Yuan Zilong^1
1Department of Radiology, Hubei Cancer Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China, ²School of Biomedical Engineering, South-Central Minzu University, Wuhan, China, ³MR Scientific Marketing, Siemens Healthcare, Wuhan, China, Wuhan, China, ⁴MR Application Predevelopment, Siemens Healthcare GmbH, Erlangen, Germany, Erlangen, Germany, ⁵Department of Gynecology, Hubei Cancer Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China

Keywords: Radiomics, Cancer, Cervical Cancer

The aim of this study was to explore the effects of SMS on radiomics features in quantitative parametric maps based on IVIM and DKI models. Meanwhile, the effects of whole-tumor and maximal-tumor layer delineation on radiomics features were compared, whether stable features are associated with clinical staging of cervical cancer was analyzed. The results showed that SMS had the greatest effect on features extracted from D* and f map from IVIM model and the least effect on features extracted from ADC map; SMS had a greater effect on the radiomics features extracted by maximum-tumor layer delineation than by whole-tumor delineation.

View Presentation Video

Jinfa Ren¹, Xiaoyang Zhai¹, Dongming Han¹, Huijia Yin¹, Ruifang Yan¹, and Kaiyu Wang^2
1Department of MR, The First Affiliated Hospital of Xinxiang Medical University, Weihui, China, ²GE Healthcare, MR Research China, Beijing, China

Keywords: Radiomics, Radiomics

Early diagnosis of postoperative glioma recurrence is difficult. But emerging measurements of radiomics could provide a powerful tool for this dilemma. We used the least absolute shrinkage and selection operator to select features and generate radiomics scores based on multiple modalities of conventional MRI to discriminate recurrence from treatment-related effects. We found that tumor recurrence could be independently identified by features from both the postoperative enhanced region and edematous region with a best performance of the combined one.

View Presentation Video

Jixin Luan¹, Chuanchen Zhang², Bin Liu¹, Aocai Yang¹, Kuan Lv³, Pianpian Hu³, and Guolin Ma^1
1China-Japan Friendship Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing, China, ²Department of Radiology, Liaocheng People's Hospital, Shandong First Medical University & Shandong Academy of Medical Sciences, Liaocheng, China, ³China-Japan Friendship Hospital, Beijing, China

Keywords: Radiomics, Cancer

In this study, we aimed to compare the performance of radiomics-based machine learning survival models in predicting the prognosis of glioblastoma multiforme (GBM) patients. The Cox proportional-hazards model (Cox-PH) and SurvivalTree, Random survival forest (RSF), DeepHit, DeepSurv four machine learning models were constructed, and the performance of the models was evaluated using C-index. We found that deep learning algorithms based on radiomics in predicting the overall survival of GBM patients, and the DeepSurv model showed the best predictive ability.

View Presentation Video

Eros Montin^1,2 and Riccardo Lattanzi^1,2,3
1Center for Advanced Imaging Innovation and Research (CAI2R) Department of Radiology, Radiology Department, New York University Grossman School of Medicine, New York, New York, USA, New York, NY, United States, ²Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology, New York University Grossman School of Medicine, New York, New York, USA, New York, NY, United States, ³Vilcek Institute of Graduate Biomedical Sciences, New York University Grossman School of Medicine, New York, New York, USA, New York, NY, United States

Keywords: Radiomics, Software Tools

Rado is a web-based application designed to fully automatically perform a radiomic analysis. Rado allows users to extract features from medical images, apply feature selections, and mine the dataset using the most common classifiers and regressors. It also offers the option to augment the dataset by means of rigid transformations.The current version of Rado allows users to interact with the dataset by means of restful APIs or using a standardized web GUI. The application will be distributed via the Cloud MR portal, which allows running the feature extraction on sparse servers as well as on local computers.

View Presentation Video

Eros Montin^1,2, Richard Kijowski³, and Riccardo Lattanzi^1,2,4
1Center for Advanced Imaging Innovation and Research (CAI2R) Department of Radiology, Radiology Department, New York University Grossman School of Medicine, New York, New York, USA, New York, NY, United States, ²Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology, New York University Grossman School of Medicine, New York, New York, USA, New York, NY, United States, ³Department of radiology, New York University Grossman School of Medicine, New York, NY, United States, New York, NY, United States, ⁴Vilcek Institute of Graduate Biomedical Sciences, New York University Grossman School of Medicine, New York, New York, USA, New York, NY, United States

Keywords: Radiomics, MSK

The results of this study showed that radiomic can automatically distinguish a healthy joint from one with impingement using water-only Dixon MRI. To our knowledge, this is the first application of radiomic for FAI diagnosis. Our radiomic analysis achieved an accuracy greater than 97%, which is higher than the 90% accuracy for detecting FAI reported for standard diagnostic tests (90%). Combining our proposed radiomic analysis with methods for automated joint segmentation could be used to rapidly identify patients with FAI, avoiding time-consuming radiological measurements of bone morphology.

View Presentation Video

Gabriel Addio Nketiah^1,2, Mohammed RS Sunoqrot ^1,3, Elise Sandsmark³, Sverre Langørgen ³, Kirsten M Selnæs ^1,3, Helena Bertilsson ^1,4, Mattijs Elschot ^1,3, and Tone F Bathen^1,3
1Department of Circulation and Medical Imaging, Norwegian University of Science and Technology, Trondheim, Norway, ²Department of Radiology and Nuclear Medicine, St. Olavs Hospital, Trondheim University Hospital,, Trondheim, Norway, ³Department of Radiology and Nuclear Medicine, St. Olavs Hospital, Trondheim University Hospital, Trondheim, Norway, ⁴Department of Urology, St. Olavs Hospital, Trondheim University Hospital, Trondheim, Norway

Keywords: Machine Learning/Artificial Intelligence, Prostate

Synopsis: Recently, predictive machine learning models have shown promise for prostate cancer diagnosis. The utility of MRI radiomic features for prostate cancer detection and classification has been shown several studies, but mostly using relatively small and single centre cohort. In this study, we showed that radiomics-based machine learning can perform relatively well compared to clinical practice, especially in large multicentre settings. On the patient-level analysis, the areas under the receiver-operating curves for PI-RADS reading by a radiologist and machine learning model were 90% and 89%, respectively.

View Presentation Video

Arda Arpak¹, Esra Sümer¹, Asım Samlı¹, Pınar S. Özbay¹, Uluç Pamuk^2,3, Kübra Çağlar², Elif M. Demir², Mehmet Yumak², Can A. Yucesoy¹, and Esin Ozturk-Isik^1
1Institute of Biomedical Engineering, Boğaziçi University, İstanbul, Turkey, ²Targeted Therapy Technologies Experimental Animal Imaging Laboratory, Boğaziçi University, İstanbul, Turkey, ³Department of Biomedical Engineering, İstinye University, İstanbul, Turkey

Keywords: Image Reconstruction, Preclinical, Radiomics

This study aims to investigate the effect of compressed sensing reconstruction on radiomics parameters of rat brain MRI at 7T. T2 weighted MRI of rat brain were randomly under-sampled with R=2.2 and reconstructed using compressed sensing. Brains were segmented after registration to a Wistar rat brain atlas. The radiomics of original and accelerated MRI data were compared in seven scans of rat brains using a Wilcoxon signed-rank test. Stable radiomics features were identified using intraclass correlation coefficients. The findings of this study indicated that not all radiomics features of rat brain were robust to compressed sensing acceleration. 

Jingjing Lu¹, Songqi Cai¹, Fang Wang², Pu-Yeh Wu³, Xianpan Pan², Jinwei Qiang⁴, Haiming Li⁵, and Mengsu Zeng^1
1Department of Radiology, Zhongshan Hospital, Fudan University, Shanghai, China, ²Department of Research and Development, Shanghai United Imaging Intelligence Co., Ltd, Shanghai, China, ³GE Healthcare, Beijing, China, ⁴Department of Radiology, Jinshan Hospital, Fudan University, Shanghai, China, ⁵Department of Radiology, Fudan University Shanghai Cancer Center, Shanghai, China

Keywords: Machine Learning/Artificial Intelligence, Radiomics, Ovarian carcinoma, Residual tumor prediction

Residual tumor (RT) status is associated with the prognosis and survival rate of patients with high-grade serous ovarian carcinoma (HGSOC). However, current RT status prediction approach through laparoscopy has disadvantages of invasiveness, high cost and incidence of tumor metastases. In this study, we proposed a radiomic-clinical nomogram, based on multiple-sequence MRI combined with score of abdominal metastases and clinical markers, for preoperative prediction of RT status. We demonstrated that the radiomic-clinical nomogram had satisfactory prediction performance in all cohorts (AUC = 0.900-0.936). The clinical application value of the nomogram was further confirmed by decision curves.

Qiu Bi¹, Kunhua Wu¹, and Yunzhu Wu^2
1the First People’s Hospital of Yunnan Province, Kunming, China, ²Siemens Healthcare, Shanghai, China

Keywords: Artifacts, Uterus

In the study, age and irregular vaginal bleeding were the valid predictive parameters in clinical model. On the basis of several common machine learning algorithms, the diverse multiparametric MRI-based radiomics models were developed to differentiate stage IA EC from benign endometrial lesions, and LR algorithm model were selected as the optimal radiomics model with the highest AUC and accuracy. Compared with clinical model and radiologist, the optimal radiomics model and the compositive models combining clinical parameters with radiomics features, like the nomogram, stacking model, and ensemble model showed better diagnostic performance and achieved good clinical net benefits. The nomogram had a higher AUC than that of the optimal radiomics model, and revealed more stable discrimination efficiency and better generalization ability than stacking and ensemble modals.

View Presentation Video

Yang Song¹, Chengxiu Zhang², Jing Zhang², Shaoyu Wang¹, Xu Yan¹, Yefeng Yao², and Guang Yang^2
12. MR Scientific Marketing, Siemens Healthcare, Shanghai, China, ²1. Shanghai Key Laboratory of Magnetic Resonance, East China Normal University, Shanghai, China

Keywords: Machine Learning/Artificial Intelligence, Software Tools

The biological meaning, model robustness and the harmonization of the features are the focuses in the current radiomics development. We designed a software named nnFAE which extends the open-source FeatureExplorer (FAE) to extract habitats features using multi-parameter MR images, to extract robust features recommended by IBSI, and to harmonize features from multi-vendors, etc. nnFAE has a graphic user interface to process the images and feature matrix in batch and can be used readily in  radiomics studies.

Miaoran Guo¹, Hu Liu¹, and Guoguang Fan^1
1The First Hospital of China Medical University, Shenyang, China

Keywords: Machine Learning/Artificial Intelligence, fMRI (resting state), Parkinson’s disease, freezing of gait, feedforward neural network, receiver operating characteristic.

• In the present investigation, we built a non-invasive and automatic classification model, by extracting radiomic features based on whole-brain functional alterations of rs-fMRI indices (mALFF, mReHo, and DC) combined with clinical scales (MoCA, and HAMD) using feedforward neural network (FNN) models, which is a representative of supervised learning classification methods. We found that these models can effectively differentiate PD-FOG and PD-nFOG and find potential biomarkers of PD-FOG, which might facilitate the individual diagnosis of PD-FOG patients.

View Presentation Video

Pavan a Poojar¹, Priti Balchandani², Yasmin Hurd³, Shilpa Taufique³, and Sairam Geethanath^1
1Accessible Magnetic Resonance Laboratory, Biomedical Imaging and Engineering Institute, Department of Diagnostic, Molecular and Interventional Radiology, Icahn School of Medicine at Mount Sinai, New York, NY, United States, ²Biomedical Imaging and Engineering Institute, Department of Diagnostic, Molecular and Interventional Radiology, Icahn School of Medicine at Mount Sinai, New York, NY, United States, ³Department of Psychiatry, Addiction Institute at Mount Sinai, New York, NY, United States

Keywords: Data Analysis, Brain

Substance use disorder (SUD) affects the structure, function, and metabolism of the brain, and MR imaging helps to track and manage the therapeutic efficacy. We analyzed cortical white matter and Amygdala regions for 5 patients with cannabis use disorder using changes in texture and volume. Each patient was scanned 4 times(at 0,2,12 and 24 weeks) with 3D fast spin echo on a 3T scanner. We found increased textural changes across intervals for cortical white matter. However, for Amygdala, volumetric changes were greater in weeks 2 and 12.Both these changes help in the early monitoring of the effectiveness of the therapy.

View Presentation Video

Evangelia I Zacharaki¹, Mohammad Alhusseini ¹, Adrian L Breto¹, Isaac L Xu¹, Ahmad Algohary¹, Wendi Ma ¹, Sandra M Gaston ¹, Matthew C Abramowitz ¹, Alan Dal Pra ¹, Sanoj Punnen¹, Alan Pollack ¹, and Radka Stoyanova ^1
1University of Miami, Miami, FL, United States

Keywords: Radiomics, Prostate, multi-parametric MRI, prostate cancer radiosensitivity, genomic siganture, PORTOS

Genomic classifiers, such as PORTOS, have shown great promise in the prediction of prostate cancer radiosensitivity. However, the spatial heterogeneity of prostate cancer may confound genomic assessment due to tumor sampling error. We aimed to develop a model predictive of PORTOS genomic signature using multiparametric MRI (mpMRI) radiomics features and machine learning. Lesions were localized based on Habitat Risk Score maps. Eight radiomic features were selected (out of 167) including T2, ADC, high B-value intensity and texture variables and used to build logistic regression models through cross-validation. Our analysis shows association between the radiomics profile and prostate lesion radiosensitivity phenotype.