Candice JX Ip^1,2, Silvia D Chang^3,4,5, Shirin Sabouri¹, Andrew Yung¹, Stefan Reinsberg², and Piotr Kozlowski^1,2,3,4,5
1UBC MRI Research Center, University of British Columbia, Vancouver, BC, Canada, ²Department of Physics and Astronomy, University of British Columbia, Vancouver, BC, Canada, ³Department of Urologic Sciences, University of British Columbia, Vancouver, BC, Canada, ⁴Vancouver Prostate Centre, Vancouver, BC, Canada, ⁵Department of Radiology, The University of British Columbia, Vancouver, BC, Canada

Clinical prostate carcinoma (PCa) detection with MRI has focused on qualitative assessment. However, novel Luminal Water Imaging (LWI) provides quantitative information with promising results for detecting PCa. To overcome shortcomings of clinical prostate MRI protocol, we propose using LWI to augment T2-weighted (T2W) images to improve image contrast for PCa detection while maintaining anatomical details needed for radiological diagnosis. Here, we investigate automatic segmentation models and various weighting functions of LWI parameter maps to generate semi-quantitative T2W images that also preserves anatomical detail. Our results show that a combined T2W and LWI parameter image provides enhanced detection of PCa.

Eric A. Borisch¹, Adam T. Froemming¹, Roger C. Grimm¹, Akira Kawashima², Joshua D. Trzasko¹, and Stephen J. Riederer^1
1Radiology, Mayo Clinic, Rochester, MN, United States, ²Radiology, Mayo Clinic, Phoenix, AZ, United States

Results of a radiological review of a previously described Model Based Iterative Reconstruction (MBIR) compensating for slice profile arising in a thick and overlapping-slice acquisition of the prostate are detailed. Blinded review by two radiologists indicated improvements in SNR, contrast fidelity, and diagnostic quality.

Christopher C Conlin¹, Roshan A Karunamuni², Christine H Feng², Ana E Rodriguez-Soto¹, Joshua M Kuperman¹, Dominic Holland³, Rebecca Rakow-Penner¹, Michael E Hahn¹, Anders M Dale^1,3,4, and Tyler M Seibert^2,5
1Department of Radiology, University of California San Diego, La Jolla, CA, United States, ²Department of Radiation Medicine and Applied Sciences, University of California San Diego, La Jolla, CA, United States, ³Department of Neurosciences, University of California San Diego, La Jolla, CA, United States, ⁴Halıcıoğlu Data Science Institute, University of California San Diego, La Jolla, CA, United States, ⁵Department of Bioengineering, University of California San Diego, La Jolla, CA, United States

Restriction Spectrum Imaging (RSI) examines diffusion in discrete tissue compartments to better detect and characterize prostate cancer. T₂ information from these compartments may further improve prostate cancer evaluation. In this study, RSI data was acquired using multiple echo times to measure both compartmental T₂ and diffusion in patients with suspected prostate cancer. A multivariable model was then developed to identify cancer from compartmental T₂ and diffusion measurements. Significant differences in compartmental T₂ were observed between normal and cancerous prostatic tissue. However, the multivariable model did not significantly improve cancer detection performance over diffusion measurements alone.

Angeliki Stamatelatou¹, Carlo Giuseppe Bertinetto², Jeroen Jansen², Geert Postma², Arend Heerschap¹, and Tom Scheenen^1
1Radiology and Nuclear Medicine, Radboud University Medical Center, Nijmegen, Netherlands, ²Analytical Chemistry & Chemometrics, Institute for Molecules and Materials, Nijmegen, Netherlands

Three-dimensional MRSI data of the prostate was analyzed with a Multivariate Curve Resolution-Alternating (MCR) approach for rapid automated localization and classification of cancer and healthy tissue. This data-driven method was used to extract common spectroscopic components without a need of prior knowledge, and compared to fitting a linear combination of prior knowledge models (LCModel). The MCR method identified components with known prostate metabolites and residual lipid and water signals Altogether, our approach can be considered as a step towards the development of an automated tool for classification of prostate MRSI spectra, avoiding subjective human intervention.

Melany B Atkins¹, Arnaud Guidon², Michael Vinski³, Thomas Schrack⁴, Heidi Harris⁵, and Ersin Bayram^6
1Radiology, Fairfax Radiological Consultants, Arlington, VA, United States, ²GE Healthcare, Boston, MA, United States, ³GE Healthcare, Lynchburg, VA, United States, ⁴Fairfax Radiological Consultants, Fairfax, VA, United States, ⁵GE Healthcare, Waukesha, WI, United States, ⁶GE Healthcare, Houston, TX, United States

While 3T offers SNR benefits for prostate imaging, patients with hip prostheses are problematic due to susceptibility artifacts and implant heating concerns. These issues are better managed at 1.5T but 1.5T lacks the SNR versus 3T.  Deep learning reconstruction (DLR) may provide the SNR boost to clear this hurdle. Feasibility of 1.5T prostate MRI with DLR in patients with hip prosthesis is explored in up to 20 subjects focusing on two key contrasts in the PI-RADS guideline: T2Wand DWI. Preliminary results suggest that 1.5T with DLR might be the preferred option for patients with hip implants.

David Atkinson¹, Giorgio Brembilla¹, Fiona Gong¹, and Shonit Punwani^1
1Centre for Medical Imaging, University College London, London, United Kingdom

The use of a fast, direct, Luminal Water Fraction fitting technique is compared with a more commonly used iterative method on data from 76 prostate cancer patients with biopsy Gleason grade available. The fast method shows a similar Receiver Operating Characteristic Area Under the Curve (ROC-AUC)  when used to classify an image region as clinically significant or non-significant cancer.  

Huan Minh Luu¹, Gyu-sang Yoo², Won Park², and Sung-Hong Park^1
1Bio and Brain Engineering, Korea Advanced Institute of Science and Technology, Daejeon, Korea, Republic of, ²Department of Radiation Oncology, Samsung Medical Center, Seoul, Korea, Republic of

Radiotherapy treatment typically requires both CT and MRI as well as labor intensive contouring for effective planning and treatment. Deep learning can enable an MR-only workflow by generating synthetic CT (sCT) and performing automatic segmentation on the MR data. However, MR and CT data are usually unpaired and limited contours are available for MR data. In this study, we proposed CycleSeg-v2 that extends the previously proposed CycleSeg to work with unpaired data. To ensure robust training, we employed LPIPS loss in addition to pseudo label. Experiments with data from prostate cancer patients showed that CycleSeg-v2 improved upon previous approaches.

Ingerid Skjei Knudtsen¹, Bendik Skarre Abrahamsen¹, Kirsten Margrete Selnæs², Mattijs Elschot¹, Sverre Langorgen³, Thomas Morten Keil³, Håkon Johansen³, Helena Bertilsson^4,5, Dag Linthoe Halvorsen⁴, Torgrim Tandstad^5,6, and Tone Frost Bathen^1
1Department of Circulation and Medical Imaging, Norwegian University of Science and Technology, Trondheim, Norway, ²Department of Nuclear Medicine and Medical Physics, 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, ⁵Department of Clinical and Molecular Medicine, Norwegian University of Science and Technology, Trondheim, Norway, ⁶Cancer Clinic, St. Olavs Hospital, Trondheim University Hospital, Trondheim, Norway

The PSMA prostate recurrence study is a multisenter prospective trial where the overall aim is to evaluate the diagnostic accuracy of 18F/68Ga-PSMA PET for detection of disease in patients with biochemical recurrence. In this preliminary analysis, we have compared the clinical imaging report of PET/CT/mpMRI to a separate evaluation of mpMRI where the radiologist was blinded to the PET results. Of 94 patients included so far, recurrent disease and/or distant metastasis was detected in 44 % (median PSA 0.4 ng/ml), according to the imaging protocol while the detection rate was 17 % for the mpMRI study scoring.