Katja Pinker-Domenig

This presentation will explain the concept and methodology of AI-enhanced multiparametric MRI, summarize the current applications in breast cancer and address its challenges and limitations.

Silvia Perez

Eun Sook Ko

High-risk women can be identified by genetic testing, mathematical risk prediction models and risk factor assessment. Reasons supporting the use of MRI for breast cancer screening may be summarized in two ways: the performance of MRI and the biologic characteristics of MRI-detected cancer. In general, annual screening is recommended by the ACS for BRCA mutation, any untested first degree relative of a known mutation carrier, anyone with a > 20% lifetime risk of developing breast cancer, history of thoracic radiation, or a known syndrome (Li-Fraumeni or Cowden’s). Abbreviated MRI, ultrafast MRI, or DWI could be additional option for high-risk screening. 

 

Nola Hylton

Muge Karaman^1,2, Yangyang Bu^3,4, Guangyu Dan^1,2, Zheng Zhong^1,2, Qingfei Luo¹, Shiwei Wang^3,4, Changyu Zhou^3,4, Weihong Hu^3,4, X. Joe Zhou^1,2,5, and Maosheng Xu^3,4
1Center for Magnetic Resonance Research, University of Illinois at Chicago, Chicago, IL, United States, ²Department of Bioengineering, University of Illinois at Chicago, Chicago, IL, United States, ³The First Clinical Medical College, Zhejiang Chinese Medical University, Hangzhou, China, ⁴Department of Radiology, The First Affiliated Hospital of Zhejiang Chinese Medical University, Hangzhou, China, ⁵Departments of Radiology and Neurosurgery, University of Illinois at Chicago, Chicago, IL, United States

Breast cancer is the second most common cancer among women. Breast tissue has a variety of structural features leading to different tissue properties which are collectively linked to the pathologic state of the breast lesions. In this study we demonstrate a hybrid DWI approach for simultaneous assessment of tissue cellularity, vascularity, and heterogeneity (DISMANTLE) based on intravoxel incoherent motion (IVIM) and continuous-time random walk (CTRW) diffusion models. Our results have shown that DISMANTLE improved the differentiation among benign and malignant breast lesions compared to the classical implementation of either IVIM or CTRW model. 

Janis M. Yee¹, Daniel S. Hippe¹, Michael Hirano¹, Bonny Chau¹, Debosmita Biswas¹, Anum S. Kazerouni¹, Mary Lynn Bryant¹, Isabella Li¹, Jennifer Xiao¹, Wei Huang², Savannah C. Partridge¹, and Habib Rahbar^1
1Radiology, University of Washington, Seattle, WA, United States, ²Oregon Health & Science University, Portland, OR, United States

Mammographic calcifications remain a diagnostic dilemma with low positive predictive value for malignancy. Our study investigated the use of quantitative DCE and DW-MRI metrics to problem solve suspicious calcifications prior to biopsy. In patients with a suspicious enhancing MRI correlate, malignant lesions exhibited higher peak PE, peak SER, functional tumor volume, and K^trans. Basic and advanced 3D DW-MRI parameters did not yield statistically significant discriminatory values in this small pilot study. Use of quantitative MRI features shows potential to reduce the number of unnecessary biopsies for suspicious mammographic calcifications among the subset that demonstrate suspicious enhancement.

Sai Man Cheung¹, Vasiliki Mallikourti¹, Tanja Gagliardi^2,3, Ehab Husain^4,5, Yazan Masannat^5,6, Steven D Heys⁶, and Jiabao He^1
1Institute of Medical Sciences, University of Aberdeen, Aberdeen, United Kingdom, ²Clinical Radiology, Aberdeen Royal Infirmary, Aberdeen, United Kingdom, ³Radiology, Royal Marsden Hospital, London, United Kingdom, ⁴Pathology Department, Aberdeen Royal Infirmary, Aberdeen, United Kingdom, ⁵School of Medicine, University of Aberdeen, Aberdeen, United Kingdom, ⁶Breast Unit, Aberdeen Royal Infirmary, Aberdeen, United Kingdom

Deregulation of lipid metabolism has been shown in BRCA1/2 genetic mutation carriers. Mammary adipose tissues in postmenopausal women are the primary sites of oestrogen production linked to tumour initiation and progression. Therefore, lipid composition in postmenopausal breast plays a key role in breast cancer monitoring and subsequent development of prevention strategies. Previous studies focused on cell or animal models and invasive lipid extraction methods, while conventional MRS is inadequate in complete lipid composition measurement. We hypothesised that lipid composition in peri-tumoural breast adipose tissue is affected by the presence of tumour in postmenopausal women, using a non-invasive 2D MRS approach.

Maya Honda¹, Masako Kataoka¹, Rie Ota¹, Mami Iima^1,2, Akane Ohashi³, Kanae Kawai Miyake¹, Marcel Dominik Nickel⁴, Yosuke Yamada⁵, Masakazu Toi⁶, and Yuji Nakamoto^1
1Graduate School of Medicine, Kyoto University, Kyoto, Japan, ²Institute for Advancement of Clinical and Translational Science (iACT), Kyoto University Hospital, Kyoto, Japan, ³Department of Radiology, National Hospital Organization Kyoto Medical Center, Kyoto, Japan, ⁴MR Application Predevelopment, Siemens Healthcare GmbH, Erlangen, Germany, ⁵Department of Diagnostic Pathology, Kyoto University Hospital, Kyoto, Japan, ⁶Department of breast surgery, Graduate School of Medicine, Kyoto University, Kyoto, Japan

The study evaluated the accuracy for predicting pathologic complete response (pCR) after neo-adjuvant systemic therapy (NST) using ultrafast dynamic contrast-enhanced (UF-DCE) MRI. The receiver operating characteristics (ROC) analysis for the presence of residual lesion revealed higher diagnostic performance of UF-DCE MRI compared with conventional dynamic contrast-enhanced (DCE) MRI overall and in the group of triple negative subtype. The deviation from pathology was smaller for UF-DCE MRI derived sizes compared to conventional DCE MRI overall and in luminal group. UF-DCE MRI potentially assesses the post-NAC status in breast cancer patients accurately in a shorter acquisition time.

Deep K Hathi¹, Rohan Nadkarni¹, Natsuko Onishi¹, Alex Anh-Tu Nguyen¹, Wen Li¹, Efstathios D Gennatas², Bonnie N Joe¹, Elissa R Price¹, I-SPY 2 Consortium³, David C Newitt¹, Ella F Jones¹, and Nola M Hylton^1
1Radiology and 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

This study explores the prediction of pathologic complete response (pCR) using tumor-derived textural features in breast cancer patients receiving neoadjuvant chemotherapy. Textural features were generated from increasingly restricted tumor masks applied on DCE-MRI signal enhancement ratio maps. Elastic net and random forests models were trained on features from baseline and early treatment timepoints, resulting in minimal differences in AUC between percent enhancement segmentation thresholds and a mean AUC of 0.68 (range 0.60-0.75). Our analysis suggests that, for the prediction of pCR, textural features derived from strongly enhancing regions dominate over those from regions of lower enhancement.

Tien Anh Nguyen¹, Hyo Jae Lee², Luu-Ngoc Do¹, Hyo-Soon Lim^2,3, and Ilwoo Park^3,4,5
1Radiology, Chonnam National University, GWANGJU, Korea, Republic of, ²Radiology, Chonnam National University Hwasun Hospital, Hwasun, Korea, Republic of, ³Radiology, Chonnam National University, Gwangju, Korea, Republic of, ⁴Radiology, Chonnam National University Hospital, GWANGJU, Korea, Republic of, ⁵Artificial Intelligence Convergence, Chonnam National University, Gwangju, Korea, Republic of

The purpose of this study was to investigate the feasibility of machine learning classifiers combining radiomics and clinical imaging interpretation for predicting malignancy in additional MR-detected enhancing lesions on multiparametric breast MRI. Machine learning algorithms trained with the combination of radiomics features extracted from breast MRI and clinical imaging interpretation what was obtained by an experienced breast radiologist demonstrated the maximal accuracy and AUC of 86.2% and 92.6%, respectively. The results from this study suggest that our approach may provide a noninvasive assisting tool to guide proper management that can reduce the use of unnecessary US or biopsy.