Guangyu Dan^1,2, Cui Feng^1,3, Zheng Zhong^1,2, Kaibao Sun¹, Muge Karaman^1,2, Daoyu Hu³, and Xiaohong Joe Zhou^1,2,4
1Center for MR Research, University of Illinois at Chicago, Chicago, IL, United States, ²Department of Bioengineering, University of Illinois at Chicago, Chicago, IL, United States, ³Department of Radiology, Tongji Hospital, Wuhan, China, ⁴Departments of Radiology and Neurosurgery, University of Illinois at Chicago, Chicago, IL, United States

Characterization of diffusion-weighted imaging signal is typically performed by modeling the data based on biophysical, mathematical, and/or statistical models to estimate quantitative biomarkers. However, conventional nonlinear fitting, which is required for the estimation of model parameters, often suffers from instability and degeneracy. In this study, we propose a Model-free Diffusion-wEighted MRI technique (MODEM) with machine learning to detect cervical carcinomas by using diffusion signal intensities and the first-order statistical features extracted from the signal attenuation as the input. By using MODEM, superior diagnostic performance and stability can be achieved even with limited number of b-values in cervical cancer detection.

Dahua Cui¹, Ailian Liu¹, Shifeng Tian¹, and Qingwei Song^1
1The First Affiliated Hospital of Dalian Medical University, Dalian, China

The aim of this study was to explore the value of multiple quantitative parameters of enhanced T2 star-weighted angiography (ESWAN) in differentiating cervical adenocarcinoma (CA) from Squamous cell carcinoma (SCC). The areas under the ROC curve of amplitude, R2* and T2* in US group were 0.729, 0.851 and 0.843, respectively, promising to be a valuable diagnostic method for differentiation CA from SCC. 

Akiyo Takada¹, Hajime Yokota², Miho Watanabe Nemoto², Takuro Horikoshi¹, Koji Matsumoto¹, Yuji Habu³, Hirokazu Usui³, Katsuhiro Nasu¹, Makio Shozu³, and Takashi Uno^2
1Radiology, Department of Radiology, Chiba University Hospital, Chiba, Japan, ²Radiology, Department of Diagnostic Radiology and Radiation Oncology, Graduate School of Medicine, Chiba University, Chiba, Japan, ³Obstetrics and gynecology, Department of Reproductive Medicine, Graduate School of Medicine, Chiba University, Chiba, Japan

To realize precision medicine for cervical cancer, it is essential to predict the prognosis. We focused on changes in ADC histogram during the clinical course. Sixty-eight patients were included and they underwent MRI evaluation 4 times during the clinical course. Imaging features were extracted from the ADC histogram in the whole tumor VOI and their change rates were calculated. The change rates of kurtosis between the first and third, and between the first and second imaging could predict recurrence with high AUCs (0.90 and 0.83). By predicting the prognosis early during clinical course, a more personalized treatment can be provided.

Liuhong Zhu¹, Pu-Yeh Wu², Hao Liu¹, and Jianjun Zhou^1
1Radiology, Xiamen Branch, Zhongshan Hospital, Fudan University, Xiamen, China, ²GE Healthcare, Beijing, China

MRI is the best imaging tool for evaluation of uterine tumors, but conventional MRI diagnosis results are subjective. T2 mapping is an objective quantification technique under certain magnetic field. Here we compared T2 values of common benign and malignant tumors under different field strengths. We found that T2 value of benign lesions was significantly lower than those of malignant lesions under both 1.5T and 3.0T, while diagnostic performance of T2 value under 3.0T were higher than 1.5T. We concluded that T2 mapping can be an effective quantitative tool in distinguishing between benign and malignant tumors, especially under 3.0T MR.

Changjun Ma¹, Ailian Liu¹, Shifeng Tian¹, and Jiazheng Wang^2
1Radiology Department, The First Affiliated Hospital of Dalian Medical University, Dalian, China, Dalian,China, China, ²Philips Healthcare,Beijing,China, Beijing,China, China

DTI is a novel MRI sequence that enables quantitative assessment of various diseases. However, its potential for diagnosis of uterus has not been explored,and it may also help differential diagnosis of myoma of uterus and uterine sarcoma.In this study, the DTI was used for quantitative analyses of myoma of uterus and uterine sarcoma. 

Yida Wang¹, YinQiao Yi¹, Haijie Wang¹, Changan Chen², Yingfang Wang², Guofu Zhang², He Zhang², and Guang Yang^1
1East China Normal University, Shanghai Key Laboratory of Magnetic Resonance, Shanghai, China, ²Department of Radiology, Obstetrics and Gynecology Hospital, Fudan University, Shanghai, China

We proposed a deep learning (DL) approach to segment ovarian lesion and differentiate ovarian malignant from borderline tumors in MR Imaging. Firstly, we used U-net++ with deep supervision to automatically define lesion region on conventional MRI; secondly, the segmented ovarian masses regions were classified with an SE-ResNet model. We compared the performance of classification model with those of radiologist’. The results showed the trained DL network model could help to identify and categorize ovarian masses with a high accuracy from MR images.

Shubhangi Agarwal¹, Jinny Sun¹, Emilie Decavel-Bueff¹, Robert A Bok¹, Romelyn Delos Santos¹, Mark Van Criekinge¹, Rahul Aggarwal², Daniel B Vigneron¹, Donna Peehl¹, John Kurhanewicz¹, and Renuka Sriram^1
1Department of Radiology and Biomedical Imaging, University of California, San Francisco, San Francisco, CA, United States, ²2Division of Hematology/Oncology, Department of Medicine, University of California, San Francisco, San Francisco, CA, United States

This study demonstrates for the first time, a successful study of three small cell neuroendocrine prostate cancer patient derived xenografts (PDX) models in liver and bone in pre-clinical setting.  Kidney served as an optimum site for propagation with its rich blood supply and high take rate. Hyperpolarized ¹³C MRI was able to identify metabolic differences between kidney, bone and liver tumors. The same PDXs had different metabolism when implanted in bone and liver. Hyperpolarized [1-¹³C] pyruvate conversion to lactate may be used to indicate early response to carboplatin in small cell neuroendocrine prostate cancer models in the bone.

Hsin-Yu Chen¹, Robert A. Bok¹, Hao G. Nguyen², Katsuto Shinohara², Antonio C. Westphalen¹, Zhen J. Wang¹, Michael A. Ohliger¹, Lucas Carvajal¹, Jeremy W. Gordon¹, Peder E.Z. Larson¹, Rahul Aggarwal², John Kurhanewicz¹, and Daniel B. Vigneron^1
1Radiology and Biomedical Imaging, University of California, San Francisco, San Francisco, CA, United States, ²Helen Diller Family Comprehensive Cancer Center, University of California, San Francisco, San Francisco, CA, United States

This project developed a new approach of integrating research biopsy targets, defined by metabolic abnormalities on hyperpolarized (HP) ¹³C pyruvate MRI, into clinical multiparametric MRI workflow to guide software-fusion transrectal ultrasound (TRUS) biopsy for improved prostate cancer risk stratification. Feasibility was investigated in 3 prostate cancer patient studies, for which research biopsy targets corresponding to high pyruvate-to-lactate conversion rates (k_PL) were identified and outlined by experienced abdominal radiologists. Fusion biopsies identified histologically-confirmed cancer at two out of three “C13” targets, consistent with clinically low- to intermediate-risk disease, supporting continued active surveillance management for these patients.

Zhaohuan Zhang¹, Jiayun Li¹, Wenyuan Li¹, Haoxin Zheng¹, Sohrab Afshari Mirak¹, Sepideh Shakeri¹, Alan Priester¹, Clara Magyar², Anthony Sisk², Robert Reiter³, Kyunghyun Sung¹, Steven Raman¹, Dieter R. Enzmann¹, Corey Arnold¹, and Holden Wu^1
1Department of Radiological Sciences, UCLA, Los Angeles, CA, United States, ²Department of Pathology, UCLA, Los Angeles, CA, United States, ³Department of Urology, UCLA, Los Angeles, CA, United States

Tumor heterogeneity is considered a key factor in determining the aggressiveness of prostate cancer (PCa). Quantitative MRI and digital whole-mount histopathology (WMHP) have potential to characterize heterogeneity in PCa in terms of MRI properties and tissue composition, respectively. Therefore, the purpose of this study is to develop a data acquisition and processing framework that produces high-resolution (1x1mm2) co-registered quantitative MRI and digital histopathology maps to enable texture features based quantification of PCa heterogeneity

Joachim Snellings¹, Kader Avan¹, Marcus Markowski², Bernd Hamm¹, Patrick Asbach¹, Carsten Warmuth¹, Mehrgan Shahryari¹, Heiko Tzschätzsch¹, Ingolf Sack¹, and Jürgen Braun^3
1Institute of Radiology, Charité Üniversitätsmedizin, Berlin, Germany, ²School of Medicine & Klinikum Rechts der Isar, Technical University of Munich, Munich (TUM), München, Germany, ³Institut für Medizinische Informatik, Charité Üniversitätsmedizin, Berlin, Germany

Prostate cancer (PCa) is the second leading cause of cancer-death in men in the western world. Advanced techniques of clinical MR-elastography (MRE), allow the characterization of PCa, based on the viscoelastic tissue-properties, which provide rich biophysical signatures of tumor progression. Using multifrequency MRE, we investigated PCa introduced LNCaP cell-lines, in a immunodeficient murine model, in-vivo, in a 3-Tesla MRI-scanner and, ex-vivo, by a 0.5-Tesla compact MRE-device. In-vivo and ex-vivo MRE values of LNCaP were in good agreement given the viscoelastic frequency-dispersion typical for soft-tissues. Compared with patient data in literature, LNCaP in mice are softer than PCa in humans.