Zhen Ren¹, Federico D Pineda¹, Elle Hill¹, Teodora Szasz¹, Rabia Safi¹, Chengyue Wu², Thomas E Yankeelov², Kirti Kulkarni¹, Hiroyuki Abe¹, Rita Nanda¹, and Gregory S Karczmar^1
1University of Chicago, Chicago, IL, United States, ²The University of Texas at Austin, Austin, TX, United States

We retrospectively reviewed data from 11 patients who received NAC and were scanned with a protocol that included ultrafast DCE-MRI for the first minute after contrast injection, followed by high spatial-resolution post-contrast imaging prior to therapy and post-treatment. The results showed a significant correlation between response to therapy and the contrast media bolus arrival time (BAT) in tumor, and normal parenchyma in each breast separately. In addition, BAT is significantly associated with background parenchymal enhancement (BPE) while BPE rate is independent of BPE. BPE rate and tumor enhancement rate may be independent predictors for malignancy. 

Alex Anh-Tu Nguyen¹, Natsuko Onishi¹, Wen Li¹, Deep K Hathi¹, Rohan Nadkarni¹, Efstathios D Gennatas², Ella F Jones¹, I-SPY2 Consortium³, David C Newitt¹, and Nola M Hylton^1
1Radiology & Biomedical Imaging, University of California, San Francisco, San Francisco, CA, United States, ²Epidemiology & Biostatistics, University of California, San Francisco, San Francisco, CA, United States, ³Quantum Leap Healthcare Collaborative, San Francisco, CA, United States

Background parenchymal enhancement (BPE) observed in normal fibroglandular tissue in breast dynamic contrast-enhanced MRI shows an association with breast cancer risk and has been investigated to predict treatment response. Contralateral BPE radiomic features and mean BPE values were calculated from DCE-MRI and their predictive performance of pathologic complete response after neoadjuvant chemotherapy was tested. The best mean AUCs were found using BPE radiomics based random forest model with minimal-redundancy-maximal-relevance feature selection in the HR- group and HER2- group. Our results show that contralateral BPE radiomic features have potential as imaging biomarkers for treatment response.

Nan Zhang¹, Lina Zhang¹, Qingwei Song¹, Ailian Liu¹, Keshuo Tang², Lei Dong², Jiazheng Wang³, and Zhiwei Shen^3
1The First Affiliated Hospital of Dalian Medical University, Dalian, China, ²Dalian Medical University, Dalian, China, ³Philips Healthcare, Beijing, China, Beijing, China

Amide proton transfer (APT) imaging is based on the chemical exchange between free bulk water protons and the amide protons (-NH) of endogenous mobile proteins and peptides in tissue[1]. Nan Meng et al suggested that APTw imaging can be used for the differential diagnosis of benign and malignant breast lesions[2]. A simple sample study has shown that APTw MRI provides a possible biomarker for assessing chemotherapy response in human breast cancer patients on 3T MRI[3]. This study aims to explore the feasibility of APTw-MRI in early evaluation of pathological complete response to NAC in molecular subtyping of breast cancer

Maren M. Sjaastad Andreassen¹, Michelle W. Tong², Ana E. Rodríguez-Soto², Somaye Zare³, Tyler M. Seibert^2,4,5, Michael Hahn², Haydee Ojeda-Fournier ², Neil P. Jerome^1,6, Tone F. Bathen^1,7, Anders M. Dale^2,8, and Rebecca Rakow-Penner^2
1Department of Circulation and Medical Imaging, Norwegian University of Science and Technology, Trondheim, Norway, ²Department of Radiology, University of California San Diego, La Jolla, CA, United States, ³Department of Pathology, University of California San Diego, La Jolla, CA, United States, ⁴Department of Bioengineering, University of California San Diego, La Jolla, CA, United States, ⁵Department of Radiation Medicine and Applied Sciences, University of California San Diego, La Jolla, CA, United States, ⁶Department of Physics, Norwegian University of Science and Technology, Trondheim, Norway, ⁷Department of Radiology and Nuclear Medicine, Norwegian University of Science and Technology, Trondheim, Norway, ⁸Department of Neuroscience, University of California San Diego, La Jolla, CA, United States

The purpose of this work was to develop a method to automatically monitor response to neoadjuvant treatment in breast cancer using longitudinally registered multi-b diffusion MRI acquisitions. The two slowest diffusion signal contributions from a three-component model were used to generate cancer probability maps (C₁C₂ classifier) that estimated tumor volume at each timepoint. Our results demonstrated that changes in C₁C₂ classifier agreed with standard dynamic-contrast-enhanced MRI criteria (RECIST) in 23 out of 24 cases. In 35% of treatment responders, the C₁C₂ classifier captured response at an earlier timepoint than RECIST. 

Anum S. Kazerouni¹, Laura C. Kennedy², Michael Hirano¹, Daniel S. Hippe¹, Bonny Chau¹, Debosmita Biswas¹, Wesley Surento¹, Shaveta Vinayak¹, Matthew Nyflot¹, Habib Rahbar¹, Suzanne Dintzis¹, and Savannah C. Partridge^1
1University of Washington, Seattle, WA, United States, ²Vanderbilt University, Nashville, TN, United States

We investigated the relationship between immune infiltration and imaging metrics derived from breast MRI in patients with triple negative breast cancer (TNBC) treated with neoadjuvant chemotherapy (NAC). Thirty-six patients with localized TNBC were imaged with diffusion-weighted and dynamic contrast-enhanced MRI prior to initiation of NAC. Baseline tumor infiltrating lymphocytes (TILs) and functional tumor volume were found to significantly differ between patients who achieved pathological complete response (pCR) compared to non-pCR patients. Baseline ADC showed a weak negative correlation with TIL counts.

Patriek Jurrius^1,2, Omar Darwish^1,3,4, Belul Shifa², Joanna Bell², John Spence², Daniel Fovargue¹, Giacomo Annio³, Renee Miller¹, Marian Troelstra¹, Ashutosh Kothari², Hisham Hamed², Diana Stavrou², Ali Sever², Sultana Hasso², Katerina Ntailiani², Keshthra Satchithananda⁵, Sarah Willson², Sarah Pinder^1,2, Anne Rigg², David Nordsletten¹, Radhouene Neji^1,4, Arnie Purushotham^1,2, and Ralph Sinkus^1,3
1King's College London, London, United Kingdom, ²Guy's and St. Thomas' NHS Foundation Trust, London, United Kingdom, ³University Paris Diderot, Paris, France, ⁴Siemens Healthcare Ltd., Frimley, United Kingdom, ⁵King's College Hospital NHS Foundation Trust, London, United Kingdom

Breast cancer is the most prevalent cancer among women. Over the years the overall survival has increased, with metastatic dissemination being one the most important prognostic factors. Currently, the presence of lymph node involvement or distant metastasis is diagnosed on imaging and core biopsy results. Yet, no technique to predict a cancer’s metastatic potential a priori is available. By incorporating magnetic resonance elastography (MRE) in a patient’s routine MRI scan the tumour interstitial fluid pressure, a factor associated with metastatic propensity, can be non-invasively measured. Preliminary results indicate a correlation between MRE measured tumour pressure and the cancer’s metastatic potential.

Jie Ding¹, Ruoshui Ha¹, Weiwei Deng², and Xiuzheng Yue^3
1Medical Imaging Center, People's Hospital of Ningxia Hui Autonomous Region, Yinchuan, China, ²Philips Healthcare, Shanghai, China, ³Philips Healthcare, Beijing, China

The accuracy of early efficacy evaluation of neoadjuvant chemotherapy for breast cancer is important for clinicians to choose chemotherapy regimen.

75 lesions were assessed according to the ELV method in comparison with RECIST1.1 and TLV. In terms of predicting pathological response, the diagnostic accuracy of the three methods were analyzed and compared by receiver operating characteristic (ROC) curve analysis.

It is feasible for ELV in evaluating the responsiveness of breast cancer patients to early NAC on DCE-MRI images. Referring to the prediction the pathological response, ELV had the best diagnostic performance with highest sensitivity, accuracy, as well as AUC value.

Josiah D Blaisdell¹, Wei Huang², and Yue Zhang^1
1EECS, Oregon State University, Corvallis, OR, United States, ²Advanced Imaging Research Center, Oregon Health & Science University, Portland, OR, United States

We applied Topological Data Analysis (TDA) to pre-neoadjuvant chemotherapy (pre-NAC) DCE-MRI datasets from the ISPY-1 trial for NAC treatment of breast cancer (BC). The pairwise topological distance between tumors’ signal enhancement ratio (SER) maps was computed. Hierarchical Agglomerative Clustering (HAC) was applied to cluster topologically similar patients.   In combination with clinical and histopathological data using logistic regression models, the predictive performance of MRI topology for pathologic complete response (pCR) was compared to longest diameter (LD) and functional tumor volume (FTV). The preliminary results show that MRI topology may be a more accurate early predictor of BC response to NAC.

Gorane Santamaría^1,2, Natali Naude^2,3, Peter Malycha^3,4, and Carolyn Mountford^2,3
1Radiology, Princess Alexandra Hospital, Brisbane, Australia, ²Queensland University of Technology, Brisbane, Australia, ³Translational Research Institute, Brisbane, Australia, ⁴Jones and Partners Radiology, St Andrew’s Hospital, Adelaide, Australia

Methylmalonic acid (MMA) has been linked to ageing and a systemic environment that favours the progression and tumours aggressiveness. We aim to assign MMA in breast tissue using in vivo 2D COSY in a series of 172 participants including women at average and elevated risk. A second purpose is to explore if MMA is associated with breast cancer risk factors. Two resonances were recorded at 3.15 and 3.19ppm in breast tissue, consistent with the presence of MMA1 and MMA2 molecules. The level of MMA correlates positively with breast density/menopausal status and IBIS risk score and, negatively with age and BMI. 

Nan Zhang¹, Zhuo Wang², Liguo Tan², Weizheng Chen², Lina Zhang¹, Ailian Liu¹, Qingwei Song¹, Jiazheng Wang³, and Zhiwei Shen^3
1The First Affiliated Hospital of Dalian Medical University, Dalian, China, ²Dalian Medical University, Dalian, China, ³Philips Healthcare, Beijing, China, Beijing, China

Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is an important tool for the diagnosis of breast cancer. It is of great importance to develop new MRI acquisition and analysis methods that can yield improved biomarkers with higher specificity for the diagnosis of breast lesions. The purpose of the present study is to explore visualization and quantitative information of tumor-associated vessels to improve diagnostic accuracy of breast tumors (MR BI-RADS 4). The results show that combination of quantitative breast vascular information may provide an accurate means for the diagnosis of breast cancer.

Masako Y Kataoka¹, Maya Honda¹, Mami Iima¹, Akane Ohashi², Rie Ota¹, Yosuke Yamada³, Masakazu Toi⁴, and Yuji Nakamoto^1
1Diagnostic Imaging and Nuclear Medicine, Kyoto University Graduate School of Medicine, Kyoto, Japan, ²Kyoto Medical Center, Kyoto, Japan, ³Department of Diagnostic Pathology, Kyoto University Hospital, Kyoto, Japan, ⁴Department of Breast Surgery, Kyoto University Hospital, Kyoto, Japan

Using ultrafast DCE (UF-DCE) MRI, size and morphology of DCIS on very early phase of the post contrast injection were compared to those on early phase of conventional DCE (C-DCE) MRI and on High resolution post contrast enhancement (HR-CE). Lesion size tended to be smaller on UF-DCE MRI. Clustered ring is not frequently observed on UF-DCE MRI compared to HR-CE. These data suggest evaluating DCIS on UF-DCE may be dealt with caution.

Ping N Wang¹, Sagar Mandava², Xinzeng Wang³, Ty A Cashen⁴, Frederick Felcz⁵, and James H Holmes^5
1Department of Medical Physics, University of Wisconsin-Madison, Madison, WI, United States, ²Global MR Applications and Workflow, GE Healthcare, Atlanta, GA, United States, ³Global MR Applications and Workflow, GE Healthcare, Houston, TX, United States, ⁴Global MR Applications and Workflow, GE Healthcare, Madison, WI, United States, ⁵Department of Radiology, University of Wisconsin-Madison, Madison, WI, United States

DCE imaging is the primary technique for MR evaluation of breast cancer.  A key problem is ghosting due to cardiac motion obscuring axillary breast tissue. Addressing this challenge, motion-insensitive technologies such as stack-of-stars acquisition have been proposed, which introduces its own problem of streaking. Based on success with a DL reconstruction to reduce noise, blurring, and ringing, this work investigated re-purposing this deep CNN to also mitigate streaking. Phantom imaging demonstrated improved CNR and more accurate line profiles. 15 patients undergoing a clinical MR exam were scanned with the additional motion-robust method, and images showed reduced structured/unstructured noise and blurring.

Akane Ohashi¹, Masako Kataoka², Mami Iima², Maya Honda², Rie Ota², Yuta Urushibata ³, Nickel Marcel Dominic ⁴, Masakazu Toi⁵, Yuusuke Hirokawa¹, and Yuji Nakamoto^2
1Radiology, National Hospital Organisation, Kyoto Medical Center, Kyoto, Japan, ²Diagnostic Imaging and Nuclear Medicine, Kyoto University Graduate School of Medicine, Kyoto, Japan, ³Siemens Healthcare K.K., Tokyo, Japan, ⁴Siemens Healthcare GmbH, Erlangen, Germany, ⁵Breast Surgery, Kyoto University Graduate School of Medicine, Kyoto, Japan

We tried to construct MRI-based multiparametric model to predict triple negative (TN) subtype among invasive cancers presenting as masses using 165 lesions (including 26 TN subtype). Maximum slope (MS) and time to enhancement (TTE) from ultrafast (UF)-DCE MRI, apparent diffusion coefficient (ADC), signal to noise ratio on T2-WI, rim enhancement on different phases of the DCE MRI were examined with univariate and multivariate logistic regression analysis. The model using MS from UF-DCE MRI and rim enhancement from early phase of DCE MRI demonstrated the AUC of 0.74 in identifying TN subtype, indicating the MRI’s potential to identify TN subtype noninvasively.

Shi Yun Sun¹, Zhuo Lin Li¹, Ying Ying Ding¹, Li Sha Nie², Cheng De Liao¹, Yi Fan Liu¹, Rui Wang¹, Jia Zhang³, and Dong Xue Zhang^1
1The Third Affiliated Hospital of Kunming Medical University，Yunnan Cancer Hospital, Kunming, China, ²GE Healthcare, MR Research China, Beijing, China, Beijing, China, ³The Third People's Hospital of Yunnan Province, China, Kunming, China

Synthetic MRI (syMRI) can quantify multiple relaxation parameters at the same time, which might have potential application value in the BI-RADS 4 lesions. 77 breast disease patients who were defined as BI-RADS 4 in the preoperative MRI examination were prospectively enrolled in this study. Before and after contrast injection, all patients underwent routine MRI and syMRI examinations. The result show that relaxation time and ADC values provided by syMRI and DWI are useful in distinguishing breast BI-RADS 4 lesions. The multi parameter model combined with clinical and imaging features can significantly improve the diagnostic ability of BI-RADS 4 breast lesions.

Isaac Daimiel Naranjo¹, Roberto Lo Gullo², Carolina Sacarelli², Almir Bitencourt², Peter Gibbs², Elisabeth Morris², Caleb Sooknanan², Jeff Reiner², Maxine S Jochelson², Sunitha Thakur², and Katja Pinker-Domenig^2
1Radiology, Memorial Sloan Kettering Cancer Center, NEW YORK, NY, United States, ²Memorial Sloan Kettering Cancer Center, New york, NY, United States

Radiomics coupled with machine learning is based on the extraction of signatures from medical images that are invisible to the human eye to create models which would improve breast cancer diagnosis. Radiomics features extracted from dynamic contrast-enhanced MRI and diffusion-weighted imaging can be combined in multiparametric MRI. We hypothesize that radiomics features extracted from multiparametric MRI would allow for an improved model affording a more accurate breast cancer diagnosis. We developed a multiparametric model that achieved the best accuracy for breast cancer diagnosis compared to models based on dynamic contrast-enhanced MRI or diffusion-weighted imaging.

Sai Man Cheung¹, Ehab Husain^2,3, Iain D Miller², Yazan Masannat^3,4, Klaus Wahle⁵, Steven D Heys⁴, and Jiabao He^1
1Institute of Medical Sciences, University of Aberdeen, Aberdeen, United Kingdom, ²Pathology Department, Aberdeen Royal Infirmary, Aberdeen, United Kingdom, ³School of Medicine, University of Aberdeen, Aberdeen, United Kingdom, ⁴Breast Unit, Aberdeen Royal Infirmary, Aberdeen, United Kingdom, ⁵Strathclyde Institute of Pharmacy and Biological Sciences, University of Strathclyde, Glasgow, United Kingdom

Upregulation of aerobic glycolysis and an elevated lactate accumulation have been linked to tumour aggressiveness. However, current evidence drawn from cell culture and small animal models remains controversial. Since lactate and lipid share the same spectral frequency, conventional MRS is inadequate in quantification of lactate under overwhelming lipid signal. Double quantum filtered (DQF) MRS allows excellent suppression of lipid signal from adipose breast tissues. We conducted a cross sectional study to examine the prognostic role of lactate concentration in grade II and III whole breast tumours using DQF MRS for the quantification of lactate concentration.

Luisa Charlotte Huck¹, Daniel Truhn², Ebba Dethlefsen², and Christiane Katharina Kuhl^2
1Radiology, RWTH Aachen University, Aachen, Germany, ²RWTH Aachen University, Aachen, Germany

According to the manufacturers intrauterine contraceptive devices (IUDs) should not be associated with systemic hormonal effects. However, women’s representatives claim that IUDs do exhibit side effects that are similar to oral or percutaneous hormonal medication. Background parenchymal enhancement (BPE) reduces the sensitivity and specificity of breast MRI and is influenced by exogenous hormonal exposures. We investigated whether IUD use is associated with BPE on contrast-enhanced breast MRI, as a surrogate marker for hormonal stimulation. We found that the application of an IUD is associated with increased BPE in breast MRI. This suggests that IUDs do have a systemic hormonal effect.

 

Kalina P Slavkova¹, Julie C DiCarlo^2,3, Viraj Wadhwa⁴, Jingfei Ma⁵, Gaiane M Rauch⁶, Zijian Zhou⁵, Thomas E Yankeelov^2,3,7,8,9, and Jonathan I Tamir^2,4,8
1Department of Physics, The University of Texas at Austin, Austin, TX, United States, ²Oden Institute for Computational Engineering and Sciences, The University of Texas at Austin, Austin, TX, United States, ³Livestrong Cancer Institutes, The University of Texas at Austin, Austin, TX, United States, ⁴Department of Electrical and Computer Engineering, The University of Texas at Austin, Austin, TX, United States, ⁵Department of Imaging Physics, MD Anderson Cancer Center, Houston, TX, United States, ⁶Department of Abdominal Imaging, MD Anderson Cancer Center, Houston, TX, United States, ⁷Department of Biomedical Engineering, The University of Texas at Austin, Austin, TX, United States, ⁸Department of Diagnostic Medicine, The University of Texas at Austin, Austin, TX, United States, ⁹Department of Oncology, The University of Texas at Austin, Austin, TX, United States

We evaluate the ability of the ConvDecoder architecture regularized using a physical model to reconstruct under-sampled dynamic MRI data, namely variable-flip angle data as a proof-of-principle. The performance of the reconstruction is evaluated by comparing the normalized error with results returned by compressed sensing and the non-regularized ConvDecoder. We hypothesize that ConvDecoder with physics-based regularization will enable significantly fewer k-space measurements, thereby allowing for expedited scan time while maintaining spatial resolution.

Collin J Buelo^1,2, Scott B Reeder^1,2,3,4,5, and Diego Hernando^1,2,5
1Medical Physics, University of Wisconsin-Madison, Madison, WI, United States, ²Radiology, University of Wisconsin-Madison, Madison, WI, United States, ³Emergency Medicine, University of Wisconsin-Madison, Madison, WI, United States, ⁴Medicine, University of Wisconsin-Madison, Madison, WI, United States, ⁵Biomedical Engineering, University of Wisconsin-Madison, Madison, WI, United States

Metallic clip soft tissue markers are a promising alternative to wire-guided localization in localizing breast cancer for surgery. However, the metallic core in the marker creates magnetic field distortions with MRI, which leads to various artifacts including signal voids and failed chemical shift-based fat suppression. This work characterizes the effect of a soft tissue marker at 1.5T and 3.0T using conventional breast imaging sequences as well as multi-spectral sequences. Several fat suppression methods were also evaluated. Field distortions and imaging artifacts are similar at 1.5T and 3.0T, and can be addressed using multi-spectral methods and T1 based fat suppression.

Tayeb Zaidi¹, John Mugler², Maureen Hood³, Brian Garra⁴, Erin Kelly⁵, Hung Do⁵, Mo Kadbi⁵, and Sunder Rajan^1
1Division of Biomedical Physics, Food and Drug Administration, Silver Spring, MD, United States, ²Department of Radiology and Medical Imaging Research, University of Virginia, Charlottesville, VA, United States, ³Department of Radiology & Radiological Sciences, Uniformed Services University, Bethesda, MD, United States, ⁴Division of Imaging, Diagnostics, and Software Reliability, Food and Drug Administration, Silver Spring, MD, United States, ⁵Canon Medical Systems USA, Tustin, CA, United States

Conventional silicone breast MR involves multi-contrast 2D scans. However, the use of 2D imaging risks the loss of fine structure visualization in the 3rd dimension. An isotropic 3D-FSE pulse sequence was first tested in a single-center study in 13 subjects with silicone breast implants (SBI). The images were evaluated by radiologists for quality and added utility of the bilateral, 3D scans. To allow for multi-site implementation of the 3D sequence, a bilateral breast phantom containing lipid, fibroglandular mimic and a silicone implant was tested across major MR vendors to ensure adequate signal-to-noise ratios and tissue contrast.