Wesley Surento¹, Anum S. Kazerouni², Janis Yee², Debosmita Biswas², Daniel S. Hippe³, Habib Rahbar², and Savannah C. Partridge^2
1Department of Biomedical Informatics and Medical Education, University of Washington, Seattle, WA, United States, ²Department of Radiology, University of Washington, Seattle, WA, United States, ³Clinical Research Division, Fred Hutchinson Cancer Research Center, Seattle, WA, United States

In this study, we investigated the association of fibroglandular tissue (FGT) apparent diffusion coefficient (ADC) measures with risk of breast cancer. Whole breast FGT regions were segmented on diffusion-weighted MRI (DWI) using a semi-automated fuzzy c-means-based algorithm. In high-risk screening cohort, ADC measures of FGT were compared in subjects with subsequent cancer diagnosis versus matched negative controls. Our findings indicate a modest negative association between FGT ADC and subsequent cancer diagnosis, with greater predictive value than mammographic density in the same cohort, pointing to the possible utility of DWI in the assessment of breast cancer risk.

Jessica Gibbs¹, Nu Le¹, Todor Karaulanov², Lisa Wilmes¹, David Newitt¹, Kathryn Keenan³, Bonnie Joe¹, and Nola Hylton^1
1University of California, San Francisco, San Francisco, CA, United States, ²CaliberMRI, Boulder, CO, United States, ³National Institute of Standards and Technology, Boulder, CO, United States

Quantitative breast MRI data was acquired at seven clinical sites using the CaliberMRI phantom. T1 and DWI data were acquired, and the qCal software was used for analysis. Protocol adherence was assessed, quantitative measures were automatically derived, and variability between measurements was evaluated. Going forward, the phantom will be used for calibration as part of a program to assess quantitative accuracy in a large multi-site clinical trial. The phantom program will be expanded to other sites in the trial, with standardized reporting used to understand variability across sites.

Jieying Zhang¹, Xiaoduo Yu¹, Shuang Chen¹, Lizhi Xie², Yan Chen¹, and Han Ouyang^1
1Diagnostic Radiology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China, ²MR Research, GE Healthcare, Beijing, China

Making an accurate preoperative diagnosis of endometrial lesions is critical to avoid unnecessary surgical procedures and protect the patients’ fertility. Noninvasive pre-surgical evaluation of abnormalities in the uterine cavity remains to be challenging. The volumetric ADC histogram analysis provides a mechanism for quantifying the overall heterogeneity and comprehensive assessment of a given abnormality. We demonstrated that whole-lesion ADC histogram-based texture analysis could serve as a quantitative technique of differentiating benign endometrial lesions (BELs) from stage IA endometrial cancer (EC) preoperatively. The combined ADC histogram model showed high classification ability for both premenopausal and postmenopausal patients.

Wei Wang¹, Yunxi Li¹, Mengchao Zhang ¹, and Yueluan Jiang^2
1Department of Radiology, China Japan Union Hospital, Changchun, China, ²MR Scientific Marketing, Siemens Healthineers, Beijing, China

This study explored the feasibility of ECV based on T1 mapping evaluate early cervical cancer pathological characters, such as deep stromal invasion (DSI) and lymphovascular space invasion (LVSI). The ECV in the group with exceeding deep 1/2 of stromal invasion was significantly higher than that of the non-exceeding 1/2 group. DSI group was significantly higher than that in the non-DSI group. Our research shows that ECV based on T1 mapping could represent an emerging preoperative imaging biomarker for the classification of DSI of early cervical cancer.

Xin Li¹, Ryan P Kopp^2,3, Eric M Baker¹, Brendan Moloney¹, William D. Rooney¹, Charles S Springer¹, Fergus V Coakley⁴, and Mark G. Gartotto^2,3
1Advanced Imaging Research Center, Oregon Health & Science University, Portland, OR, United States, ²Portland VA Medical Center, Portland, OR, United States, ³Urology, Oregon Health & Science University, Portland, OR, United States, ⁴Diagnostic Radiology, Oregon Health & Science University, Portland, OR, United States

Transcytolemmal water exchange effects are quantified with SSM-DCE MRI and DWI kurtosis imaging with b-values up to 2000 s/mm².  The two-dimensional scatter plot of the unidirectional cellular water efflux rate constant (k_io) and the diffusion kurtosis (K) shows good separation between lesion and normal appearing prostate tissue. 

Zhaohuan Zhang¹, Sohrab Afshari Mirak¹, Melina Hosseiny¹, Afshin Azadikhah¹, Amir Bajgiran¹, Alan Priester², Kyunghyun Sung¹, Anthony E Sisk³, Robert E Reiter², Steven Raman¹, Dieter R Enzmann¹, and Holden H Wu^1
1Department of Radiology, UCLA, Los Angeles, CA, United States, ²Department of Urology, UCLA, Los Angeles, CA, United States, ³Department of Pathology, UCLA, Los Angeles, CA, United States

Diffusion-Relaxation Correlation Spectrum Imaging (DR-CSI) can provide unique microstructural information for prostate cancer characterization, but requires longer scan times for two-dimensional encoding of TE and b-values. This study developed a data-driven sequential backward selection framework that determined subsampled encoding schemes for DR-CSI, achieving 70% reduction of scan time to 6min while maintaining accurate ex vivo prostate microstructure mapping.

Yiming Wang¹, Christina L. Herrera², David M. Owen², Quyen N. Do¹, Yin Xi^1,3, Matthew A. Lewis¹, Baowei Fei^1,4,5, Catherine Y. Spong², Diane M. Twickler^1,2, and Ananth J. Madhuranthakam^1,4
1Department of Radiology, UT Southwestern Medical Center, Dallas, TX, United States, ²Department of Obstetrics and Gynecology, UT Southwestern Medical Center, Dallas, TX, United States, ³Department of Population and Data Sciences, UT Southwestern Medical Center, Dallas, TX, United States, ⁴Advanced Imaging Research Center, UT Southwestern Medical Center, Dallas, TX, United States, ⁵Department of Bioengineering, University of Texas at Dallas, Richardson, TX, United States

Measuring placental perfusion can provide important information about its function. Arterial Spin Labeled (ASL) MRI is a non-contrast perfusion imaging method particularly suitable in pregnancy as it does not require an exogenous contrast. We performed a preliminary assessment of placental perfusion in pregnant subjects, both normal and with CHTN, at 16-20 week and 24-28 week gestational ages. Perfusion reduction was observed in all 5 normal subjects and in 75% (12 of 16) of the subjects with CHTN. This could provide important information about perfusion changes during pregnancy and can serve as a precursor for a larger-scale longitudinal study in CHTN. 

Milica Medved¹, Keiko Tsuchiya², Xiaobing Fan¹, Gregory S Karczmar¹, and Hiroyuki Abe^1
1Department of Radiology, The University of Chicago, Chicago, IL, United States, ²Department of Radiology, Shiga University of Medical Science, Otsu, Japan

Breast DCEMRI signal is standardized for variable imaging protocols and the linear rate of background parenchymal enhancement (BPE) is calculated, to quantitatively and objectively describe changes in BPE rates following preventative tamoxifen treatment.  This is in contrast with the current practice of using 4 subjective categories to describe BPE.  Decreased BPE rates post-treatment agree with earlier results showing that BPE, like breast density, correlates with breast cancer risk.  Standardization for imaging parameters and contrast agent relaxivity increases the observed effect size, pointing to increased sensitivity to treatment-induced changes and the potential as a tool for individual breast cancer risk management.

Zhen Ren¹, Federico D. Pineda¹, Frederick M. Howard², Elle Hill¹, Teodora Szasz³, Rabia Safi¹, Milica Medved¹, Rita Nanda², Thomas E. Yankeelov^4,5,6,7,8,9, Hiroyuki Abe¹, and Gregory S. Karczmar^1
1Radiology, The University of Chicago, Chicago, IL, United States, ²Medicine, The University of Chicago, Chicago, IL, United States, ³Research Computing Center, The University of Chicago, Chicago, IL, United States, ⁴Biomedical Engineering, The University of Texas at Austin,, Austin, TX, United States, ⁵Diagnostic Medicine, The University of Texas at Austin, Austin, TX, United States, ⁶Oncology, The University of Texas at Austin, Austin, TX, United States, ⁷Institute for Computational and Engineering Sciences, The University of Texas at Austin, Austin, TX, United States, ⁸Livestrong Cancer Institutes, The University of Texas at Austin, Austin, TX, United States, ⁹Imaging Physics, MD Anderson Cancer Center, Houston, TX, United States

We retrospectively reviewed data from 23 patients who received neoadjuvant therapy (NAT) and were scanned with a protocol that included ultrafast DCE-MRI (temporal resolution = 3-7 seconds) for the first minute after contrast injection prior to NAT. We measured parenchymal kinetics from ipsi- and contra-lateral normal parenchyma separately, so that new parameters related to bilateral parenchymal enhancement asymmetry could be calculated. The results showed that patients with similar pre-NAT parenchymal enhancement kinetics in ipsi- and contralateral normal parenchyma were more likely to achieve pCR post NAT ($$$p < 0.02$$$).

Junyu Guo¹, Keith Husley¹, Yin Xi¹, and Ivan Pedrosa^1
1Radiology, UT southwestern medical center, Dallas, TX, United States

Clear cell renal carcinoma cancer (ccRCC) is the most common subtype among renal masses. ccRCC identification helps in decision making between active surveillance and definitive intervention. A clear cell likelihood score (ccLS) using subjective interpretation of multiparametric MRI by radiologists was proposed recently. In this study, we investigate whether deep learning (DL) using the three main MR sequences for ccLS can facilitate the diagnosis of ccRCC. We compared the results of twelve trained DL models with the reported ccLS performance. Our results demonstrate that DL may achieve a performance comparable to radiologists and provide useful information for identification of ccRCC.

Ignacio Dominguez¹, Paola Caprile², Odette Rios², Ignacio San-Francisco³, and Cecilia Besa ^1
1Radiology, Pontificia Universidad Catolica de Chile, Santiago, Chile, ²Physics, Pontificia Universidad Catolica de Chile, Santiago, Chile, ³Urology, Pontificia Universidad Catolica de Chile, Santiago, Chile

We developed a non-invasive tool to predict the GS classification of PCa based on mpMRI information using ML. This retrospective study included 86 male patients with positive PCa fusion (mpMRI-ultrasound) biopsy. A radiomic analysis was performed considering first order, textural, shape, and clinical information. The best model found included image (T2w - ADC) and clinical information. The mean AUC was 0.91 [0.75−0.99] (p <0.05), with a validation AUC of 0.91 for a classification of high-lower aggressiveness (GS≥7 vs GS=6). Combining MRI-based radiomic and clinical information can significantly improve the model performance to classify PCa aggressiveness.

Ke Zhang^1,2, Simon M.F. Triphan¹, Felix T. Kurz², Chrstian H. Ziener², Hans-Ulrich Kauczor¹, Heinz-Peter Schlemmer², and Oliver Sedlaczek^1,2
1Department of Diagnostic and Interventional Radiology, Heidelberg University Hospital, Heidelberg, Germany, ²Department of Radiology, German Cancer Research Center, Heidelberg, Germany

Renal perfusion is an important physiological parameter in health and disease (1). To measure kidney perfusion using arterial spin labelling (ASL) the respiratory motion is a major problem. In this study, respiratory motion information is acquired from a projection signal and used to adjust the position of the excited slice in real time. The feasibility of free-breathing multi-slice kidney perfusion imaging using EPI based pseudocontinuous ASL (pCASL) with navigator-based slice tracking method is investigated.
 

Armando Tartaro^1,2, Roberto Renzetti³, Michele De Cristofaro Aulisa⁴, Federica Morrone⁵, Andrea De Nicola^6,7, Ekaterina Bliakharskaia⁷, and Paul E Summers^7
1Department of Clinical, Oral and Biotechnology Sciences, University "G. d'Annunzio" Chieti-Pescara, Chieti-Pescara, Italy, ²Magnetic Resonance Service, Ospedale di Popoli, AUSL, Pescara, Italy, ³UOC of Urology, Ospedale di Pescara, AUSL, Pescara, Italy, ⁴Faculty of Medicine, University "G. d'Annunzio“, Chieti-Pescara, Chieti-Pescara, Italy, ⁵Radiology Department, Centro Morrone, Caserta, Italy, ⁶Ospedale di Chieti, Chieti, Italy, ⁷QMRI Tech, Pescara, Italy

Prostate dynamic contrast-enhanced (DCE) MRI is often criticized for its weaknesses in terms of discriminating transition zone lesions and tumors in general. Using high-temporal resolution DCE MRI in patients who went on to prostatectomy, we evaluated time to peak (TTP), as well as the enhancement rate factor, and initial enhancement slope of a simple empirical mathematical model of early contrast uptake. Differences between tumors and surrounding healthy tissues were seen for the mean values of all three parameters. The AUCs were good (>0.8 in all cases except for TTP between tumor and transition zone) but case numbers limited.