Deepa Darshini Gunashekar¹, Lars Bielak^1,2, Benedict Oerther ³, Matthias Benndorf ³, Anca Grosu ^2,3, Constantinos Zamboglou^2,3, and Michael Bock ^1,2
1Dept.of Radiology, Medical Physics, Medical Center University of Freiburg, Faculty of Medicine, University of Freiburg, Freiburg, Germany, ²German Cancer Consortium (DKTK), Partner Site Freiburg, Freiburg, Germany, Freiburg im Breisgau, Germany, ³Dept.of Radiology, Medical Center University of Freiburg, Faculty of Medicine, University of Freiburg, Freiburg, Germany

A convolutional neural network was implemented to automatically segment tumors in multi-parametric MRI data. The influence of the variability in the ground truth data was evaluated for automated prostate tumor segmentation. Therefore, the agreement between the predictions of the CNN was measured with co-registered whole mount histopathology images and the tumor contours drawn by an expert radio-oncologist. The results indicate that the network can discriminate tumor from healthy tissue rather than mimicking the radiologist.

Snigdha Sen¹, Vanya Valindria¹, Hayley Pye², Alistair Grey³, Aiman Haider⁴, Alex Freeman⁴, Caroline Moore³, Hayley Whitaker², Shonit Punwani⁵, Saurabh Singh⁵, and Eleftheria Panagiotaki^1
1Centre for Medical Image Computing, University College London, London, United Kingdom, ²Molecular Diagnostics and Therapeutics Group, University College London, London, United Kingdom, ³Department of Urology, University College London Hospitals NHS Foundation Trust, London, United Kingdom, ⁴Department of Pathology, University College London Hospitals NHS Foundation Trust, London, United Kingdom, ⁵Centre for Medical Imaging, University College London, London, United Kingdom

Mp-MRI has moderate specificity for the diagnosis of prostate cancer, resulting in many false positive results and unnecessary biopsies for patients with other prostatic diseases. This work presents an approach using VERDICT-MRI to investigate differences in microstructural parameter estimates between normal prostate tissue, false positive lesions and clinically-significant cancer.  The VERDICT intracellular volume fraction and intrinsic diffusivity in the extracellular-extravascular space can discriminate false positive lesions with greater accuracy than the clinical ADC from mp-MRI. These results show VERDICT has the potential to identify false positives and reduce unnecessary biopsies in prostate patients.

Hsin-Yu Chen¹, Hao G. Nguyen², Zhen J. Wang¹, Michael A. Ohliger¹, Jeffry P. Simko², Matthew R. Cooperberg², Daniel Gebrezgiabhier¹, Lucas Carvajal¹, Jeremy W. Gordon¹, John Kurhanewicz¹, Rahul Aggarwal², Daniel B. Vigneron¹, and Robert A. Bok^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

Serial hyperpolarized ¹³C+mpMRI and paired metabolic image-guided biopsy were employed in a new integrative active surveillance (AS) paradigm, exemplified by a patient who initially enrolled in AS for 5 years, but later underwent radical prostatectomy upon progression. Metabolic biomarker k_PL images detected early progression preceding radiologic and clinical parameters. Final whole-gland histopathology also linked k_PL and PI-RADS targets to multifocal tumors, where k_PL spatially reflected intratumoral heterogeneity, and high k_PL regions correlated with adverse pathologic features such as G4 cribriform pattern, extracapsular extension, and lymphovascular invasion. Results indicated HP ¹³C may help optimize AS by improving clinical risk calculation.

Ayumu Kido¹, Akira Yamamoto^1,2, Mitsuru Takeuchi³, Yu Ueda⁴, Jaladhar Neelavalli⁵, and Tsutomu Tamada^1
1Radiology, Kawasaki Medical School, Kurashiki, Japan, ²Radiology, Kawasaki Medical School, Kurashiki-shi, Japan, ³Radiology, Radiolonet Tokai, Nagoya, Japan, ⁴Philips Japan, Tokyo, Japan, ⁵Philips India, Bangalore, India

Single-shot EPI (sshEPI) DWI still suffers from distortion and blurring. Multi-shot EPI (mshEPI) DWI called IRIS enables to reduce image distortion and blurring. A total of 142 patients with suspected prostate cancer (PC) underwent mpMRI including sshEPI DWI and IRIS under the same scan time. The image quality and diagnostic performance of the lesions were compared between sshEPI DWI and IRIS. IRIS improved image distortion and blurring compared to sshEPI DWI, and also suggested a similar or higher diagnostic performance compared to sshEPI DWI.

Virendra Kumar¹, Pradeep Kumar¹, Rajeev Kumar², S. Datta Gupta³, Sanjay Thulkar⁴, N. R. Jagannathan⁵, and Rama Jayasundar^1
1Department of NMR & MRI Facility, All India Institute of Medical Sciences, New Delhi, India, ²Department of Urology, All India Institute of Medical Sciences, New Delhi, India, ³Department of Pathology, All India Institute of Medical Sciences, New Delhi, India, ⁴Department of Radio-diagnosis, IRCH, All India Institute of Medical Sciences, New Delhi, India, ⁵Dept. of Radiology, Chettinad Academy of Research, Kelambakkam, Tamil Nadu, India

Gleason score (GS) is currently clinical predictor of prostate cancer (PCa) mortality. The present study investigates the metabolic profile of blood plasma for distinguishing aggressive PCa patients from non aggressive using targeted ¹H-NMR metabolomics. A significantly higher concentration of lactate, pyruvate, choline, dimethylamine, glucose, betaine and lower concentration of leucine, isoleucine, valine, histidine, phenylalanine, and tyrosine were observed in patients with GS 8-10 as compared to patients with GS 7 and 6. Our results suggested that pathway alterations, amino acids, phospholipids and glucose related to PCa progression. Identifying biomarker/s associated with GS can be improved diagnosis and treatment.

Michael Mills¹, Malou Van Zanten¹, Julian Pearce², Bernard Ho², Kristiana Gordon^1,2, Peter S Mortimer^1,2, Pia Ostergaard¹, and Franklyn A Howe^1
1St George's University of London, London, United Kingdom, ²St George's University Hospitals NHS Foundation Trust, London, United Kingdom

Despite its prevalence, methods for investigating lymphoedema (chronic tissue swelling due to abnormal lymphatic functioning) remain limited. Dynamic Contrast-Enhanced Magnetic Resonance Lymphangiography (DCE-MRL) allows identification of lymph vessels in both healthy and lymphoedematous limbs, however quantitative analyses are lacking. DCE-MRL data was acquired for 30 minutes in lymphoedematous (n=7) and healthy limbs (n=4), with estimates of lymphatic vessel signal characteristics compared between groups. Maximum enhancement compared to baseline was significantly increased in healthy volunteers, while other metrics were similar. Larger groups and longer imaging sessions are likely required to detect additional differences. 

Hsin-Yu Chen¹, Michael A. Ohliger¹, Zhen J. Wang¹, Mary E. Frost¹, Lucas Carvajal¹, Philip M. Lee¹, Jeremy W. Gordon¹, John Kurhanewicz¹, Peder E.Z. Larson¹, Robert A. Bok¹, Rahul Aggarwal², 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

Harnessing unique metabolic insight from hyperpolarized (HP) ¹³C MRI, this study describes an integrated approach toward routine cancer staging and restaging using combined HP ¹³C plus standard of care (SOC) MRI with commercially-available ¹³C and ¹H array coils and a time-efficient adaptive protocol. A series of patients with osseous, liver, nodal metastases, and locally advanced primary tumors suggested HP ¹³C not only detected high metabolic activity in progressive tumors, but it also revealed inter- and intratumoral metabolic heterogeneity. Results indicated HP ¹³C-MRI can greatly complement SOC-MRI in assessing therapeutic response and guiding clinical decisions.

Aidan Tollefson¹, Nate Roberts¹, Ali Pirasteh^1,2, and Diego Hernando^1,2
1Medical Physics, University of Wisconsin, Madison, WI, United States, ²Radiology, University of Wisconsin, Madison, WI, United States

B0 inhomogeneities often lead to failures in chemical shift-based fat suppression, particularly in complex susceptibility environments such as the abdomen. To optimize the application of fat suppression, characterization of B0 inhomogeneities in fat regions is needed. However, such characterization has not been reported for abdominal fat. Therefore, the purpose of this work was to measure the magnitude and variability of B0 inhomogeneity at 3T in abdominal subcutaneous fat. This study shows that several subcutaneous fat regions exhibit B0 inhomogeneities large enough to induce failures in common fat suppression methods. These measurements may guide future optimization of fat suppression methods.