Rebekah L McLaughlin¹, David C Newitt², Catherine Park³, Dorota Wisner², Lisa J Wilmes², Evelyn Proctor², and Nola Hylton^1,2
1The UC Berkeley - UCSF Graduate Program in Bioengineering, UC Berkeley & UCSF, San Francisco, CA, United States, ²Department of Radiology and Biomedical Imaging, UCSF, San Francisco, CA, United States, ³Department of Radiation Oncology, UCSF Helen Diller Comprehensive Cancer Center, San Francisco, CA, United States

Using a high resolution DWI sequence for the breast, we investigated the changes in diffusion measurements relative to distance from the tumor boundary in patients receiving neoadjuvant treatment. In this study, we found that tumors and surrounding tissue had variable ADC patterns before and during neoadjuvant chemotherapy, suggesting that water mobility properties vary among tumors. Our early results suggest that a decrease in the change of tumor to stroma mean ADC values may identify tumors that will respond.

Jeon-Hor Chen^1,2, Wei-Fan Pan², Chih-Chen Kuo², Julian Kao¹, Li-Kuang Chen¹, Jocelyn Lu¹, Shadfar Bahri¹, Rita S Mehta³, Daniel H-E Chang¹, Orhan Nalcioglu¹, and Min-Ying Lydia Su^1
1Center for Functional Onco-Imaging, Department of Radiological Science, University of California, Irvine, California, United States, ²Department of Radiology, China Medical University Hospital, Taichung, Taiwan, ³Department of Medicine, University of California Irvine, Orange, California, United States

The results from our previous MRI monitoring study had shown that the reduction of breast density in patients receiving neoadjuvant chemotherapy (NAC) was majorly related to Doxorubicin and Cyclophosphamide (AC). The effect of taxane on the change of breast tissue is still not clear. 49 women receiving a new NAC protocol starting with 12 weeks of taxane with optional AC were studied for their 3D MR density change. The results show that NAC using Abraxane and Carboplatin also had similar effect on the reduction of FV and PD as did AC. The effect is age-related and duration-related. Since quantitative measures of FV and PD can be obtained on MRI, the reduction of these two parameters after NAC may potentially be used an imaging biomarker to correlate with future cancer risk occurring in the contralateral normal breast.

Xia Li¹, Lori R. Arlinghaus¹, A. Bapsi chakravarthy¹, E. Brian Welch¹, Jaime Farley¹, Ingrid A. Mayer¹, Vandana G. Abramson¹, Richard G. Abramson¹, Mark C. Kelley¹, Ingrid M. Meszoely¹, Julie A. Means-Powell¹, Ana M. Grau¹, Sandeep Bhave¹, and Thomas E. Yankeelov^1
1Vanderbilt University Institute of Imaging Science, Nashville, Tennessee, United States

Tumor response to neoadjuvant chemotherapy is currently monitored by frank changes in tumor size, which often do not correlate with pathologic findings at surgery. DCE-MRI offers information related to tumor perfusion and permeability, vascular volume, and extravascular extracellular volume fraction. In this study, we attempted to determine the optimal analysis method for assessing if changes in these parameters after one cycle of therapy could separate pathologic complete responders from non-responders. The results show that tumor perfusion and permeability and vascular volume yield significant differences between responders and non-responders, but that changes in tumor volume do not.

Xia Li¹, Lori R. Arlinghaus¹, A. Bapsi chakravarthy¹, E. Brian Welch¹, Jaime Farley¹, Ingrid A. Mayer¹, Vandana G. Abramson¹, Mark C. Kelley¹, Ingrid M. Meszoely¹, Julie A. Means-Powell¹, Ana M. Grau¹, Sandeep Bhave¹, and Thomas E. Yankeelov^1
1Vanderbilt University Institute of Imaging Science, Nashville, Tennessee, United States

To monitor tumor response to neoadjuvant chemotherapy, investigators have begun to employ the quantitative physiological parameters available from dynamic contrast enhanced MRI (DCE-MRI). However, most studies track the changes in average parameter values obtained from the whole tumor region of interest, thereby discarding all spatial information on tumor heterogeneity. In this study, we applied a novel registration algorithm to longitudinal DCE-MRI data and performed a voxel-by-voxel analysis. The results indicate that voxel-based analysis, after longitudinal registration, may improve the ability of DCE-MRI to separate pathologic complete responders from non-responders after one cycle of therapy when using the fast exchange regime model.

Xia Li¹, Subramani Mani¹, Lori R. Arlinghaus¹, A. Bapsi chakravarthy¹, E. Brian Welch¹, and Thomas E. Yankeelov^1
1Vanderbilt University Institute of Imaging Science, Nashville, Tennessee, United States

Dynamic contrast enhanced MRI (DCE-MRI) can offer information related to tumor perfusion and permeability, vascular volume, extravascular extracellular volume fraction, and the intracellular water lifetime of a water molecule. There have been many efforts employing DCE-MRI as a surrogate biomarker for predicting the response of breast tumors to neoadjuvant chemotherapy. However, most studies perform univariate analysis on these parameters. In this study, we perform multivariate analysis using machine learning methods. The preliminary results demonstrate the feasibility of using DCE-MRI data and machine learning for predicting the response of breast tumors to a single cycle of neoadjuvant chemotherapy.

Lisa J Wilmes¹, Rebekah L McLaughlin¹, Sumedha Sinha¹, David C Newitt¹, Lisa Singer¹, Evelyn Proctor¹, Dorota Wisner¹, Emine U Saritas², Ajit Shankaranarayanan³, Suchandrima Banerjee³, Bonnie N Joe¹, and Nola M Hylton^1
1University of California San Francisco, San Francisco, CA, United States, ²University of California Berkeley, Berkeley, CA, United States, ³Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States

A high-resolution reduced field of view diffusion-weighted imaging sequence (HR-DWI) was optimized for breast imaging and evaluated against a standard (STD-DWI) sequence in patients with invasive breast cancer undergoing neoadjuvant treatment. Mean tumor apparent diffusion coefficients (ADCs) were not significantly different between sequences. However, the mean lower percentile ADCs were lower for HR-DWI, and this difference was significant at the early treatment time point. Of the ADC metrics evaluated, the pre-treatment 15th percentile HR-ADC was found to correlate most strongly with tumor volume change at the end of treatment. These results suggest HR-DWI may have value in characterizing treatment responses.

Sungheon Kim¹, Li Feng¹, Linda Moy¹, Melanie Moccaldi¹, Kai T. Block¹, Daniel K. Sodickson¹, and Ricardo Otazo^1
1Center for Biomedical Imaging, Radiology, NYU School of Medicine, New York, NY, United States

Both high temporal and high spatial resolutions of DCE-MRI are required for screening patients with high risk of the breast cancer using pharmacokinetic model analyses. The objective of our study was to investigate the feasibility of using a novel image reconstruction method that combines compressed sensing and parallel imaging for radial trajectories (k-t RASPS) for DCE-MRI study of the breast cancer. Unilateral breast MRI was performed with six patients using golden-angle radial VIBE pulse sequence. Our results demonstrate that high temporal resolutions images (up to 1.2 s/frame) can be reconstructed without noticeable temporal blurring.

Catherine J Moran¹, Brian A Hargreaves¹, Manojkumar Saranathan¹, and Bruce L Daniel^1
1Radiology, Stanford University, Stanford, CA, United States

T2-weighted images contribute to differential diagnosis in breast MRI. In this work we evaluate a clinically available high-resolution 3D extended echo train T2-weighted acquisition in comparison to a conventional Fast Spin Echo acquisition in 27 breast MRI patients. Depiction of lesion morphology, lesion signal intensity, and the discernment of benign from malignant lesions are assessed for each sequence. Potential diagnostic implications of the utilization of 3D T2-weighted imaging in the breast are also noted.

Min Jung Kim^1,2, Ji Soo Choi¹, Hon J Yu³, Jeon-Hor Chen³, Min-Ying Su³, Eun-Kyung Kim¹, and Hee Jung Moon^1
1Radiology, Yonsei University College of Medicine, Seoul, Korea, ²John Tu and Thomas Yuen Center for Functional Oncoimaging, UC Irvine, Irvine, CA, United States, ³John Tu and Thomas Yuen Center for Functional Oncoimaging, UC Irvine

 

Jeon-Hor Chen^1,2, Christine McLaren³, Wen-Pin Chen³, Siwa Chan⁴, Dah-Cherng Yeh⁵, Orhan Nalcioglu¹, and Min-Ying Lydia Su^1
1Center for Functional Onco-Imaging, Department of Radiological Science, University of California, Irvine, California, United States, ²Department of Radiology, China Medical University Hospital, Taichung, Taiwan, ³Department of Epidemiology, University of California, Irvine, California, United States, ⁴Department of Radiology, Taichung Veterans General Hospital, Taichung, Taiwan, ⁵Department of Surgery, Taichung Veterans General Hospital, Taichung, Taiwan

The correlation between levels of endogenous estrogen and progesterone and measures of fibroglandular tissue volume and percent breast density analyzed on 3D MRI was studied. Twenty-four healthy premenopausal Asian women were recruited. Each woman received 4 weekly breast MRIs during their MC. Blood sample was also collected at the same day. It was found that there was significant correlation of FV and PD with endogenous estradiol, and FV with progesterone in the third week after the starting of menstruation. Our study did not find strong evidence of a weekly lag effect of endogenous hormone on the measured FV and PD. The results from this study raised the possibility that the association between sex hormone and breast cancer may be mediated, in part, by increasing breast density.

Zhan Xu¹, Leah Henze Bancroft¹, Andrew L Alexander^2,3, and Frederick Kelcz^4
1Department of BioMedical Engineering, University of Wisconsin Madison, Madison, WI, United States, ²Department of Medical Physics and Psychiatry, University of Wisconsin Madison, Madison, WI, United States, ³Waisman Laboratory for Brain Imaging, University of Wisconsin Madison, Madison, WI, United States, ⁴Department of Radiology, University of Wisconsin Madison, Madison, WI, United States

Our work aims at restoring undistorted DWI for breast imaging which is suffering from imperfect fat suppression and low SNR. We acquire two identical DWIs with reversed Phase Encoding Direction then apply correction in a manner of averaging magnetic field inhomogeneity. Corrected DWIs were compared with identical T2W NoFatSuppression images. Our methods well reconstruct the undistorted DWI,both location and signal intensity of lesion and breast are well maintained.

Alana R. Amarosa¹, Sungheon Kim¹, Amy Melsaether¹, Jason McKellop¹, Melanie Moccaldi¹, and Linda Moy^1
1Radiology, NYU School of Medicine, New York, New York, United States

The purpose of this study is to investigate whether the amount of fibroglandular tissue on MRI and background parenchymal enhancement are associated with breast cancer risk. This retrospective study included 36 BRCA mutation carriers and 40 controls who underwent MRI-guided biopsy for an enhancing lesion. When comparing signal percent enhancement for pre- and post-menopausal BRCA carriers, we found a significant difference (p=0.04). Also, menopausal BRCA carriers had higher early enhancement rate than pre-menopausal BRCA carriers. No such difference was found between pre- and post-menopausal control groups. These findings substantiate our hypothesis that breast cancer risk is associated with breast density.

Bertine L. Stehouwer¹, Dennis W.J. Klomp¹, Peter R. Luijten¹, Karel A.F. Houwert², Paul J. van Diest³, Willem P.Th.M. Mali¹, Maurice A.A.J. van den Bosch¹, and Wouter B. Veldhuis^1
1Radiology, University Medical Center, Utrecht, Utrecht, Netherlands, ²Radiology, Zuwe Hofpoort Hospital, Woerden, Utrecht, Netherlands, ³Pathology, University Medical Center, Utrecht, Utrecht, Netherlands

Eighteen patients with 21 biopsied breast lesions were imaged at 7T. The protocol included a dynamic series consisting of 7 consecutive 3D T1-weighted turbo field echo sequences and ultra-high resolution imaging with a 0.45x0.57x0.45mm3 acquisition matrix. All 21 malignant lesions were scored according the BI-RADS-MRI criteria by two radiologists, independently, using the dynamic series. Subsequent ultra-high resolution imaging increased reader confidence in n=13 and=9 for the separate radiologists and changed interpretation in n=3 and n=5 cases for the radiologists. Overall, we showed the feasibility of clinical 7T CE breast MRI, and its amenability to BI-RADS-MRI conform analysis.

Seung-Kyun Lee¹, and Ileana Hancu^1
1GE Global Research, Niskayuna, NY, United States

Susceptibility-induced B0 field in bilateral breast was calculated in thirteen healthy volunteers based on segmented anatomical images at 3 T. For each volunteer, nine axial B0 maps were fitted with spherical harmonics up to the third order, and population variability was investigated. Common features in the susceptibility-induced B0 gradient were identified and the effectiveness of higher order shim was evaluated.

Ritse M Mann¹, Roel DM Mus¹, Christian Geppert², Cindy PM Frentz¹, Nico Karssemeijer¹, Henkjan Huisman¹, and Bram Platel^1,3
1Radiology, Radboud University Nijmegen Medical Centre, Nijmegen, Gld, Netherlands, ²Oncology, Siemens Healthcare, Erlangen, Germany, ³Fraunhofer MEVIS, Bremen, Germany

Initial maximum slope of enhancement as derived from ultrafast TWIST images acquired during contrast inflow provides a strong discriminator between benign and malignant disease that outperforms classic 3-timepoint analysis. By using a fixed grid of relative enhancement versus time curve types can be defined that can be easily implemented in current radiological practice.

Kristin Granlund^1,2, Jafi Lipson¹, Jennifer Kao¹, Debra Ikeda¹, Brian Hargreaves¹, and Bruce Daniel^1
1Radiology, Stanford University, Stanford, California, United States, ²Electrical Engineering, Stanford University, Stanford, California, United States

The DESS sequence has been proposed to acquire 3D high-resolution, low-distortion diffusion-weighted images in reasonable scan times. In this abstract, we have three radiologists evaluate images acquired with DCE, DWI, and DESS sequences. Radiologists evaluated the visibility of lesions, the image sharpness, the lesion margin, rim signal, and internal setpa. We found that DESS highlighted more lesions than DWI, had better image quality, which made morphology easier to assess.

Dennis Lai Hong Cheong^1,2, Bingwen Zheng¹, Bo Zhang^1,3, Soo Chin Lee^4,5, and Thian Chor Ng^1,6
1Clinical Imaging Research Center, A*STAR & National University of Singapore, 117456, Singapore, ²Neuroradiology Department, National Neuroscience Institute, 308433, Singapore, ³Quantitative Image Processing Group, SBIC/A*STAR, 138671, Singapore, ⁴Department of Haematology-Oncology, National University Health System, 119074, Singapore, ⁵Cancer Science Institute, 117456, Singapore, ⁶Department of Radiology, National University of Singapore, 119074, Singapore

Individually measured arterial input functions (AIF) are critical for quantitative DCE MRI studies. Typical field of view (FOV) for breast DCE MRI do not cover major artery supplying the breast. We explored the feasibility of including the descending aorta in the FOV by using an additional surface coil placed on the back of patients for better proximity in positioning for direct measurements of AIF from the aorta. However, there are serious blood inflow effects at the descending aorta where blood flow is fast. Nevertheless, inflow effect is avoidable by extracting AIF at a more distal end along the aorta.

Martin D Pickles¹, Peter Gibbs¹, and Lindsay W Turnbull^1
1Centre for MR Investigations, University of Hull, Hull, East Yorkshire, United Kingdom

Re-operation rates post breast conserving surgery for DCIS pose a significant healthcare burden. Accurate estimation of DCIS extent should minimize re-operation rates. The aim of this study was to compare the accuracy of DCIS extent measurements from both x-ray mammography and MRI. Thirty-one participants underwent both x-ray mammography and MR assessment of DCIS. Measurements from both imaging modalities were compared to histopathology findings via Bland Altman plot methodology. MR estimates of DCIS extent were demonstrated to be more accurate than x-ray mammography and were less affected by certain histopathological features such as grade and necrosis.

Daniel M Krainak¹, Brain Garra², and Sunder S Rajan^1
1CDRH/OSEL/DP, U.S. Food & Drug Administration, Silver Spring, MD, United States, ²CDRH/OSEL/DIAM, U.S. Food & Drug Administration, Silver Spring, MD, United States

We measured T1 and T2 relaxation times of silicone breast implants and phantom materials that may mimic breast tissue at 3 Tesla. The characterization of the relaxivities of silicone breast implants provides the opportunity for improved pulse sequence design for the diagnosis of implant rupture, leak, deformity, or degradation. To the authors’ knowledge, this is the first report of MR relaxations times (T1 and T2) for silicone breast implants at 3 Tesla.

Anderson Nnewihe¹, Bruce L Daniel¹, Jafi Lipson¹, Catherine Moran¹, and Brian A Hargreaves^1
1Radiology, Stanford University, Stanford, California, United States

The purpose of this study was to assess the potential diagnostic impact of very high-resolution breast MRI scans with a newly designed, fitted coil array versus lower resolution scans with a standard commercially available array. Preliminary results show better depiction of lesion morphology with the high resolution, but because the cases were clearly BIRADS 5, there was no difference in the diagnosis between the two resolutions. A wider range of cases is necessary to assess the diagnostic benefits of high-resolution breast MRI and further studies are underway.

Martin D Pickles¹, Peter Gibbs¹, Ersin Bayram², and Lindsay W Turnbull^1
1Centre for MR Investigations, University of Hull, Hull, East Yorkshire, United Kingdom, ²GE Healthcare, Waukesha, WI, United States

High temporal and high spatial resolution images are usually mutually exclusive. In this work, we demonstrate proof of principle, that both high spatial and temporal resolution DCE-MRI images can be acquired by utilising time resolved imaging of contrast kinetics (TRICKS). A pulse sequence was developed that combined TRICKS with a breast optimised 3D T1W gradient echo sequence. The resulting sequence was successfully acquired in multiple patients on a 3.0T scanner resulting in high spatial (1.0x1.0x1.4mm) and high temporal (10 second) resolution images. This work has demonstrated the ability to acquire both high spatial and high temporal resolution breast DCE-MRI datasets.

Shelley Waugh^1,2, Richard Lerski^1,2, Luc Bidaut², and Alastair Thompson^2,3
1Medical Physics, Ninewells Hospital, Dundee, Angus, United Kingdom, ²University of Dundee, Dundee, United Kingdom, ³Department of Surgery, Ninewells Hospital, Dundee, United Kingdom

This study considers the use of texture analysis in the grading of the two most common types of breast cancer- invasive ductal and infiltrative lobular cancers. It was believed that the underlying histological nature of cancer growth patterns used for staging would also result in textural features on MR imaging that could potentially allow classification between grades using computer-based texture analysis methods. Our preliminary results show classification between grades 1, 2 and 3 of invasive ductal cancer is possible and reliable, as is classification between grades 2 and 3 of lobular cancer. The co-occurrence matrix produced the best classification accuracy.

Jeon-Hor Chen^1,2, Siwa Chan³, Dah-Cherng Yeh⁴, Chih-Chen Kuo², Jocelyn Lu¹, Peter Fu¹, MuQing Lin¹, Orhan Nalcioglu¹, and Min-Ying Lydia Su^1
1Center for Functional Onco-Imaging, Department of Radiological Science, University of California, Irvine, California, United States, ²Department of Radiology, China Medical University Hospital, Taichung, Taiwan, ³Department of Radiology, Taichung Veterans General Hospital, Taichung, Taiwan, ⁴Department of Surgery, Taichung Veterans General Hospital, Taichung, Taiwan

Functional breast symmetry is less known. In this study we used 3D MRI to investigate how the bilateral breast tissues of healthy women respond to the hormonal fluctuations during a menstrual cycle. Twenty-four healthy premenopausal Asian women were studied. The obtained CV and maximal percent difference of FV in the left and the right breast of all subjects were compared. The results demonstrate that during the MC, the CV of the fibroglandular tissue volume measured from the left and the right breasts shows a strong correlation (r > 0.7), so as the maximal changes of FV. The fluctuations of the morphological parameters between the left and the right breasts were not significantly different. Since most of breast cancers occur unilaterally, it would be interesting to investigate whether the functional response of tissues in bilateral breasts to hormonal stimulation is associated with different risks in developing cancer.

Bertine L. Stehouwer¹, Dennis W.J. Klomp¹, Peter R. Luijten¹, Willem P.Th.M. Mali¹, Maurice A.A.J. van den Bosch¹, and Wouter B. Veldhuis^1
1Radiology, University Medical Center, Utrecht, Utrecht, Netherlands

The purpose was to intra-individually compare BI-RADS-MRI assessment on 7T and 3T dynamic contrast-enhanced Breast MRI. Six patients with breast cancer were imaged on both field strengths. Two radiologists scored all exams. Image quality was scored to be at least sufficient for both field strengths. An inter-observer variability in the application of BI-RADS descriptors was observed. However, this did however not affect final assessment category, which was scored as at least suspicious in all cases.

Lori R. Arlinghaus^1,2, Richard D. Dortch^1,2, Jennifer G. Whisenant^1,2, Gregory D. Ayers³, Adrienne N. Dula^1,2, Seth A. Smith^1,2, and Thomas E. Yankeelov^1,2
1Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States, ²Department of Radiology and Radiological Sciences, Vanderbilt University, Nashville, TN, United States, ³Department of Biostatistics, Vanderbilt University, Nashville, TN, United States

Magnetization transfer (MT) imaging is sensitive to changes in the macromolecular content of tissue and is, therefore, gaining increased attention as a noninvasive approach to probe the complex tumor environment in cancer. The amount of magnetization transfer between macromolecules and the surrounding free water in tissue can be quantified by the MT ratio (MTR). In this study, we explore the repeatability of MTR measurements in the breast fibroglandular tissue of healthy controls at 3T to serve as a benchmark for future longitudinal studies of breast cancer treatment.

David A. Broadbent¹, Daniel J. Wilson¹, Nisha Sharma², Barbara J. Dall², Sarah E. Bacon¹, Steven P. Sourbron³, and David L. Buckley^3
1Medical Physics & Engineering, Leeds Teaching Hospitals NHS Trust, Leeds, West Yorkshire, United Kingdom, ²Radiology, Leeds Teaching Hospitals NHS Trust, Leeds, West Yorkshire, United Kingdom, ³Division of Medical Physics, University of Leeds, Leeds, West Yorkshire, United Kingdom

The introduction of quantitative MRI into clinical routine has often been hampered by the need to change and compromise clinical protocols. Here we show how quantitative data can be obtained with minimal interference in the case of DCE-MRI of the breast following neoadjuvant chemotherapy (NAC). Twenty patients were studied before and after 2 cycles of NAC. All studies were reported in the normal way with the addition of estimates of tumor perfusion, blood volume, capillary permeability and interstitial volume. Tumor perfusion decreased in responders more than in non-responders with no significant changes in the other parameters.

Martin D Pickles¹, Peter Gibbs¹, Ersin Bayram², Jessica Buzek², Zhenghui Zhang², and Lindsay W Turnbull^1
1Centre for MR Investigations, University of Hull, Hull, East Yorkshire, United Kingdom, ²GE Healthcare, Waukesha, WI, United States

CHESS and spectrally selective inversion recovery fat nulling techniques are sensitive to B0 and B1 inhomogeneities which results in inconsistent non uniform fat suppression. The aim of this work is to develop a more robust fat suppression. By combining a Shinnar-Le Roux spectral spatial pulse with an Adiabatic SPectral-selective Inversion Recovery (ASPIR) pulse B1 insensitive water-only excitation was achieved. Dual shim volumes were employed to reduce B0 sensitivity. Traditional and spectral spatial ASPIR fat suppression techniques were acquired in multiple patients. In all cases, spectral spatial ASPIR fat suppression was deemed superior to the traditional method by a breast radiologist.

Jie Huang¹, Sarah Schafer¹, Lori Hoisington¹, and Gerald Aben^1
1Department of Radiology, Michigan State University, East Lansing, MI, United States

This study validated lesion washout volume fraction as an MRI biomarker for improving the characterization of suspicious contrast-enhancing breast lesions. Using a sample consisting of 94 contrast-enhanced lesions with histopathology reports, the study found that the washout volume fraction was significantly larger for the malignant tumors than the benign lesions, reflecting the hypervascularity associated with tumor angiogenesis. It provides an additional MRI biomarker for improving the characterization of suspicious contrast-enhancing breast lesions, and the biomarker has the potential to improve the computer-based assessment in breast MRI.

Catherine Judith Moran¹, Kristin L Granlund^1,2, Bruce L Daniel¹, Bragi Sveinsson^1,2, Ernesto Staroswiecki^1,2, Marcus T Alley¹, and Brian A Hargreaves^1
1Radiology, Stanford University, Stanford, CA, United States, ²Electrical Engineering, Stanford University, Stanford, CA, United States

Dual echo steady state (DESS) acquisitions can be utilized to acquire high-resolution, quantitative T2 maps in clinically feasible scan times, thus facilitating investigation of T2 properties of in vivo tissues. In this work we present the results of an initial investigation of DESS quantitative T2 mapping in the breast. DESS T2 maps are assessed in normal, benign and malignant in vivo breast tissue and measured T2 values are compared to those previously reported for tissues in the breast.

Martin D Pickles¹, Peter Gibbs¹, Ersin Bayram², Donglai Huo², and Lindsay W Turnbull^1
1Centre for MR Investigations, University of Hull, Hull, East Yorkshire, United Kingdom, ²GE Healthcare, Waukesha, WI, United States

While the sensitivity of breast MRI is high the specificity is somewhat lower. T2W sequences help to increase the specificity of breast MR. The aim of this work was to develop a 3D T2W fat nulled sequence. Additionally, by substituting the traditional CHESS fat nulling pulse with an IR technique improved fat nulling was anticipated. By combining a 3D CUBE FSE sequence with an IR pulse bilateral sagittal images of the breast (5min20sec with a 0.875x0.875x2.4mm voxel size) were obtained. The potential of the sequence was observed in the robust fat nulled high spatial resolution images obtained in 10 datasets.

Karl-Heinz Herrmann¹, Tim Sprenger¹, Werner A Kaiser², and Jürgen R Reichenbach^1
1IDIR I, Medical Physics Group, Jena University Hospital, Jena, Germany, ²IDIR I, Jena University Hospital, Jena, Germany

A fast multi-echo TSE sequence is proposed for Dixon based water, fat and silicone separation. The sequence acquires 5 echoes, using bipolar gradients, within each refocusing interval. To correct the phase errors introduced by the bipolar readouts, the central k-space line is acquired for each echo in two additional refocusing intervals after the imaging echo train. There the readout gradients in the second interval are inverted to provide phase calibration data. The spectral separation was performed with a region-growing stabilized VARPRO implementation. One in-vivo dataset of a patient is shown and the separation proved robust against the B0 inhomogeneities present in the patient.

Michel G.M. Italiaander¹, Peter J. Nijholt¹, Oliver Kraff², Alexander Raaijmakers¹, Bertine Stehouwer¹, Peter R. Luijten¹, and Dennis W.J. Klomp^1
1Imaging division, University Medical Center, Utrecht, Utrecht, Netherlands, ²Erwin L. Hahn Institute for MRI, University Duisburg-Essen, Essen, Germany

Detection and staging of breast cancer can benefit from increased SNR that is potentially available at high magnetic fields like 7T. However, as strong T1 weighting in CE-MRI is required, a high density of strong flip angles need to be applied, that generally will be limited due to SAR constrains. Therefore we propose the use of efficient quadrature surface arrays to maximize CE-MRI efficiency at 7T in bilateral breast imaging while remaining within SAR guidelines. The relatively uniform CE-MRI sensitivity in both breasts, and very strong T1 weighting is demonstrated in a volunteer and in a patient with breast cancer.

Alana R. Amarosa¹, Linda Moy¹, Amy Melsaether¹, Jason McKellop¹, Melanie Moccaldi¹, Jin Zhang¹, and Sungheon Kim^1
1Radiology, NYU School of Medicine, New York, New York, United States

The purpose of this study was to investigate the ability of quantitative kinetic analysis to identify malignant enhancing lesions on MRI in BRCA1 patients. In this retrospective study, we included 27 BRCA1 mutation carriers who underwent MRI guided biopsy. A linear principal component analysis (PCA) transformation was used to measure background parenchymal enhancement (BPE). Percent enhancement (PE) and early enhancement rate (S23) were measured in both lesion and BPE. PE and S23 for BPE were compared to those for malignant and benign lesions. Using our quantitative analysis, we found PE to be helpful in distinguishing between benign and malignant lesions.

Amy Melsaether¹, Nathaniel Margolis¹, Alana R. Amarosa¹, Melanie Moccaldi¹, Samantha Heller¹, Sungheon Kim¹, and Linda Moy^1
1Radiology, NYU School of Medicine, New York, New York, United States

Dynamic contrast enhanced breast MRI can detect breast cancers that are occult on mammography and sonography. The enhancement pattern of such cancers relative to breast parenchyma has never been quantified. We compared 21 mammographically and sonographically occult breast cancers to control areas of parenchyma. In malignant lesions compared to controls, there was a shorter time to peak and a greater peak enhancement, but no significant difference in signal enhancement ratio. Our findings suggest that time to peak and peak enhancement can help differentiate benign breast parenchyma from otherwise occult malignant lesions, possibly reducing the need for biopsy.

Amy Melsaether¹, Alana R. Amarosa¹, Nathaniel Margolis¹, Melanie Moccaldi¹, Samantha Heller¹, Sungheon Kim¹, and Linda Moy^1
1Radiology, NYU School of Medicine, New York, New York, United States

Malignant breast masses measuring up to 1.0cm are frequently occult on conventional imaging and demonstrate overlapping morphology with benign entities on dynamic contrast enhanced MRI. We therefore investigated whether kinetic markers including time to peak enhancement, peak enhancement, and signal enhancement ratio (SER) can be used to separate these malignancies from benign findings. Thirty up to 1.0cm masses and foci were analyzed for these variables using DynaCAD. Results demonstrate a trend towards increased peak enhancement in malignant masses and suggest benign biopsies could be decreased by establishing a minimum enhancement threshold for biopsy.

Bertine L. Stehouwer¹, Hendrik de Leeuw², Peter R. Seevinck², Fredy Visser^1,3, Peter R. Luijten¹, Dennis W.J. Klomp¹, Paul J. van Diest⁴, Chris J.G. Bakker², and Wouter B. Veldhuis^1
1Radiology, University Medical Center, Utrecht, Utrecht, Netherlands, ²Image Sciences Institute, University Medical Center, Utrecht, Netherlands, ³Philips Healthcare, Best, Netherlands, ⁴Pathology, University Medical Center, Utrecht, Utrecht, Netherlands

We investigated the ability of high-field MRI to detect breast micro-calcifications in a phantom and ex-vivo set-up by exploiting susceptibility differences between calcifications and glandular tissue. Magnitude images depicted several field disturbances, but were clearly distorted by B1-inhomogeneities and, moreover, did not allow characterization of the field inhomogeneities. Phase derivative images were not hampered by B1-inhomogeneities and showed a high sensitivity for detecting the characteristic six-lobed blooming pattern which is seen in case of calcifications. The calcifications were confirmed by CT and mammography, respectively. This study shows the feasibility of the detection of breast micro-calcifications using high-field MRI.

Jeon-Hor Chen^1,2, Siwa Chan³, Dah-Cherng Yeh⁴, Chin-Kai Chang², Li-Kuang Chen¹, Wei-Fan Pan², MuQing Lin¹, Orhan Nalcioglu¹, and Min-Ying Lydia Su^1
1Center for Functional Onco-Imaging, Department of Radiological Science, University of California, Irvine, California, United States, ²Department of Radiology, China Medical University Hospital, Taichung, Taiwan, ³Department of Radiology, Taichung Veterans General Hospital, Taichung, Taiwan, ⁴Department of Surgery, Taichung Veterans General Hospital, Taichung, Taiwan

For assessing the association between MRI-based density and cancer risk, a large dataset is required and combining MRI from multiple centers is the only feasible way to achieve this goal. The purpose of this work is to compare the measurement consistency of breast volume, fibroglandular tissue volume and percent density using 4 different MR scanners. Thirty-four healthy Asian female subjects were studied at two 1.5T and two 3T scanners. The correlation of FV between each pair of two MR scanners was very high, with all R2 ≥ 0.99. For some cases, however, the measurement variation was high, which was due to a large difference of one scanner compared to the other three scanners. The results show that when MR pulse sequences are optimized, and a well-developed segmentation method is used, consistent density parameters from the same women can be obtained on images acquired using different MR scanners. The positioning difference may account for some variation, and further optimization work may be developed to minimize its impact.

Leah C Henze-Bancroft¹, Frederick Kelcz², Kevin M Johnson³, and Walter F Block^1,3
1Department of Biomedical Engineering, University of Wisconsin - Madison, Madison, WI, United States, ²Department of Radiology, University of Wisconsin - Madison, Madison, WI, United States, ³Department of Medical Physics, University of Wisconsin - Madison, Madison, WI, United States

We present initial results of our attempt to determine whether a tight temporal footprint (15 s) and spatial resolution (1 mm isotropic) can better characterize heterogeneously enhancing lesions in the breast. Eight volunteers underwent a standard clinical DCE and a rapid 3D radial (VIPR) SPGR DCE over two days of scanning. The 3D-VIPR SPGR method was able to provide good visualization of a lesion in three orthogonal reformats displaying heterogeneous enhancement as well as good depiction of the margin enhancement. To date this method has only been used on unilateral exams but should be expandable to bilateral imaging.

Thiele Kobus¹, Jack Van Asten¹, Alan Wright¹, Arend Heerschap¹, and Tom Scheenen^1
1Radiology, Radboud University Nijmegen Medical Centre, Nijmegen, Gelderland, Netherlands

In this study the use of ¹H MRSI for the detection of lactate in the prostate was evaluated with the use of a semi-LASER sequence. No convincing lactate signal was detected in patients with highly aggressive prostate cancer. The minimal detectable lactate level was determined and estimated to be around 3 mM.

Jan Weis¹, Francisco Ortiz-Nieto¹, and Håkan Ahlström^1
1Department of Radiology, Uppsala University Hospital, Uppsala, Sweden

Metabolite concentrations may be helpful in monitoring treatment efficacy, and relapse probability associated with prostate cancer. Single-voxel spectroscopy of the prostate at 3 T with surface coil was performed. Metabolite-to-unsuppressed water spectral intensity ratios of healthy volunteers and patients with biopsy-proven prostate cancer were determined with an LCModel. Mean normalized spectra, LCModel fits, and absolute concentrations of Cho, Cr, and Cit were computed for normal volunteers and cancer patients.

Patrik Zamecnik¹, Grzegorz Bauman², Julien Dinkel^1,3, Heinz-Peter Schlemmer¹, and Thomas Gaass^4
1Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany, ²Department of Medical Physics, German Cancer Research Center (DKFZ), Heidelberg, Germany,³Department of Radiology, Massachusetts General Hospital, Boston, Germany, ⁴Zentralinstitut für Medizintechnik, Technische Universität München, Garching, Germany

Validation of the feasibility of TCIR (two compartment inversion recovery) contrast agent-free approach for prostate cancer detection by comparison with the standard T2w, DCE- and DWI-sequences. Using the TCIR technique it was possible to generate good quality high-resolution maps of the fractional blood volume in the prostate without using a contrast agent. TCIR has a high potential for prostate cancer detection because of its high spatial resolution and sensitivity, first of all in patients, which cannot be examined using contrast media.

Andriy Fedorov¹, Luis Ibanez², Kemal Tuncali¹, Robert V Mulkern^1,3, William M Wells¹, Clare Tempany⁴, and Fiona Fennessy^1
1Department of Radiology, Brigham and Women's Hospital, Boston, MA, United States, ²Kitware Inc, Clifton Park, NY, United States, ³Department of Radiology, Children's Hospital, Boston, MA, United States, ⁴Brigham and Women's Hospital, Boston, MA, United States

There is strong evidence of the value in using multiparametric MRI (mpMRI) to characterize prostate cancer. However, quantitative analysis of mpMRI is challenging due to the differences in resolution and acquisition-related distortions, which are particularly strong while imaging prostate using endorectal coil placed in air-filled balloon at 3T. Herein we propose a non-rigid registration approach fine-tuned to compensate for such artifacts by incorporating the prior knowledge about the nature of the distortion into the method. We evaluate the robustness and accuracy of the approach, and employ it to quantify distortions present in the DWI MRI of the prostate.

Carlos Felipe Uribe Munoz^1,2, Edward C Jones³, Silvia D. Chang⁴, Larry Goldenberg^5,6, and Piotr Kozlowski^2,5
1Physics & Astronomy, University of British Columbia, Vancouver, British Columbia, Canada, ²MRI Research Centre, University of British Columbia, Vancouver, British Columbia, Canada, ³Pathology and Laboratory Medicine, University of British Columbia, Vancouver, British Columbia, Canada, ⁴Radiology, University of British Columbia, Vancouver, British Columbia, Canada, ⁵Vancouver Prostate Centre, Vancouver, British Columbia, Canada, ⁶Urological Sciences, University of British Columbia, Vancouver, British Columbia, Canada

Diffusion Tensor Imaging (DTI) has been applied in prostate cancer diagnosis. It has been well accepted that water apparent diffusion coefficient (ADC) has a lower value in tumours than in healthy prostatic tissue, while fractional anisotropy (FA) has been reported to be higher, lower, and unchanged. Histology slices were registered with in-vivo and ex-vivo DTI images, and average values of ADC and FA were calculated for certain regions of interest. No significant differences in FA between normal peripheral zone and prostatic carcinoma were found suggesting that FA is not likely to contribute to diagnostic capabilities of DTI in prostate cancer.

Edward M Lawrence¹, Tristan Barrett¹, Andrew N Priest¹, Vincent J Gnanapragasam², Ferdia A Gallagher¹, Andrew J Patterson¹, Anne Warren³, and Evis Sala^1
1Clinical MRI Unit, Addenbrooke's Hospital, Cambridge, United Kingdom, ²Urology, Addenbrooke's Hospital, Cambridge, United Kingdom, ³Histopathology, Addenbrooke's Hospital, Cambridge, United Kingdom

Rosa Rossling¹, Rene Dittrich¹, Emily Decelle², Chin-Lee Wu², W Scott McDougal², and Leo L Cheng^2
1Charité Universitätsmedizin, Berlin, Germany, ²Massachusetts General Hospital, Boston, MA, United States

Implementation of PSA testing has increased the number of early diagnosed cases of prostate cancer. This shift in diagnosis necessitates the evolution of pathology and other biomolecular markers to identify and categorize early stages of the disease. In previous studies, citrate has been identified as a potential marker for malignancy and aggressiveness. Here we correlate citrate levels with PSA velocity, PSA density, percent free PSA, which we are using as surrogate markers for PCa growth and aggressiveness.

Gregor Thörmer¹, Josephin Otto¹, Christian Schröder¹, Nikita Garnov¹, Martin Reiss-Zimmermann¹, Lars-Christian Horn², Minh Do³, Jens-Uwe Stolzenburg³, Michael Moche¹, Thomas Kahn¹, and Harald Busse^1
1Dept. of Diagnostic and Interventional Radiology, Leipzig University Hospital, Leipzig, Saxony, Germany, ²Institute of Pathology, Division of Breast, Urogenital and Perinatal Pathology, University of Leipzig, Leipzig, ³Dept. of Urology, Leipzig University Hospital, Leipzig, Saxony, Germany

Active surveillance (AS) is currently discussed as an alternative to radical therapy for patients with low-risk prostate cancer. This option requires an accurate and reproducible technique to monitor the progress of the disease. Transrectal ultrasound-guided biopsy is only moderately suited because of its invasive nature and a substantial amount (40%) of undergrading. Multiparametric MRI, on the other hand, can provide quantitative parameters of tissue function that may help to assess cancer aggressiveness. This preliminary work evaluates the predictive value of combined DWI and MR spectroscopic parameters to discriminate indolent from aggressive carcinoma.

Rosa Rossling¹, Johannes Kurth¹, Emily Decelle², Chin-Lee Wu², W Scott McDougal², and Leo L Cheng^2
1Charité Universitätsmedizin, Berlin, Germany, ²Massachusetts General Hospital, Boston, MA, United States

A 2005 publication on the investigations of human prostate cancer metabolomics led to two major conclusions that metabolomic profiles can differentiate tissue specimens with and without cancer glands, as well as patient pathological stages. However, this study only proposed the metabolomic structure from the analysis of a training cohort and did not evaluate another independent testing cohort. In this study we followed the concepts of previous report and analyzed tissue samples from prostate cancer patients with constructions of training and testing cohorts.

Mehdi Moradi¹, Andriy Fedorov¹, William M Wells¹, Kemal Tuncali¹, Sandeep N Gupta², Fiona M Fennessy¹, and Clare M Tempany^1
1Radiology, Brigham and Women's Hospital - A Teaching Affiliate of Harvard Medical School, Boston, MA, United States, ²GE Global Research Center, Niskayuna, NY

We propose to use machine learning to enhance the process of target selection for 3T MR-guided transperineal prostate biopsy. Support vector machine and Gaussian classifiers with different combinations of diffusion and DCE MRI are examined. Training is performed on data from 13 prostatectomy cases with histologically confirmed cancer in the peripheral zone. The trained classifier was used to determine the outcome of in ten PZ biopsy samples from five patients. The Bayesian classifier with ADC as the only feature resulted in the ROC area of 0.964 in leave-one-patient-out cross-validation on the training dataset. The outcomes of eight of the ten biopsies, including all three cancer samples, were correctly determined.

Josephin Otto¹, Gregor Thörmer¹, Matthias Seiwerts¹, Jochen Fuchs¹, Nikita Garnov¹, Lars-Christian Horn², Harald Busse¹, Thomas Kahn¹, and Michael Moche^1
1Department of Diagnostic and Interventional Radiology, University Hospital, Leipzig, Saxony, Germany, ²Institute of Pathology, University Hospital, Leipzig, Saxony, Germany

Prostate cancer is characterized by a high incidence but relatively low mortality. Disease prediction, risk stratification and therapeutical decision of histopathologically confirmed prostate cancer are commonly obtained by nomograms, which have a relatively low predictive value. Multiparametric MRI, in contrast, can provide detailed information of unilateral or bilateral extracapsular extension and seminal vesicle invasion, which is important for patients selected for nerve-sparing surgery. The high specificities and relatively high sensitivities observed here suggest that 3 T MRI with an endorectal coil is a suitable diagnostic tool for the local staging of prostate cancer.

Tryggve Holck Storås¹, and Kjell-Inge Gjesdal^2
1Intervention Centre, Oslo University Hospital, Oslo, Norway, Norway, ²Sunnmøre MR-klinikk, Ålesund, Norway, Norway

Prostate tissues have been shown to exhibit multiexponential T2 and ADC decay. In this paper we used a multiecho interleaved SE and IR technique to assess T1 and T2 values of the two compartments. We found a long T1 component of about 3000 ms and a short T1 component of about 1200ms.

Harald Busse¹, Josephin Otto¹, Gregor Thörmer¹, Nikita Garnov¹, Thomas Kahn¹, and Michael Moche^1
1Diagnostic and Interventional Radiology Department, Leipzig University Hospital, Leipzig, Germany

A multiparametric MRI examination of the prostate typically generates a large number of individual images (over 500) and MR spectra (several hundreds). Analysis and review of these data can be a laborious and time-consuming procedure for the radiologist. This work describes and evaluate a non-commercial software solution for the analysis and simultaneous display of data from T2-weighted, diffusion-weighted, dynamic contrast-enhanced (DCE) and chemical shift imaging (CSI). Besides an automated calculation of semiquantitative DCE parameter maps, the tool also allows for a transparent overlay of selected low-error metabolite ratios from CSI data analyzed with a well-established spectroscopy software (LCModel).

Emily Decelle¹, Phillip Knape², Chin-Lee Wu¹, W Scott McDougal¹, and Leo L Cheng^1
1Massachusetts General Hospital, Boston, MA, United States, ²Charité Universitätsmedizin, Berlin, Germany

Prostate cancer exhibits a racial disparity such that African American men have a significantly higher incidence and mortality rate than Caucasian men. The possible ethological or biological factors responsible for these differences remain unknown but may reveal much about the disease that has become the most frequently diagnosed cancer in men. In this study, we compare tissue metabolic intensities and the metabolomic profiles of prostate cancer from African American patients with those from matched Caucasian patients in order to expose metabolic similarities and differences between the two patient populations.

Eline Vos¹, Geert Litjens¹, Thiele Kobus¹, Thomas Hambrock¹, Christina Hulsbergen-Van de Kaa², Henkjan Huisman¹, and Tom Scheenen^1
1Radiology, Radboud University Nijmegen Medical Center, Nijmegen, Netherlands, ²Pathology, Radboud University Nijmegen Medical Center, Nijmegen, Netherlands

Dynamic contrast enhanced MR imaging for retrospective assessment of prostate cancer aggressiveness was explored in fifty-one patients at 3T with histopathological Gleason scores of resected prostates as the gold standard. Calibration for pharmacokinetic modeling was done by using non-cancer peripheral zone tissue to estimate patient-specific arterial input functions. For semi-quantitative parameters LateWash and Relative Enhancement there was a significant difference between low aggressive and high aggressive prostate cancer in the peripheral zone. For quantitative parameters Ktrans, Kep and Ve there was no correlation with aggressiveness, however, they showed significant difference between non-cancer tissue and prostate cancer for the peripheral zone

Nyoman D. Kurniawan¹, Gary Cowin¹, Paul Sved², Geoffery Watson³, and Roger Bourne^4
1Centre for Advanced Imaging, University of Queensland, Brisbane, Queensland, Australia, ²Department of Urology, Royal Prince Alfred Hospital, The University of Sydney, Sydney, New South Wales, Australia, ³Department of Anatomical Pathology, Royal Prince Alfred Hospital, University of Sydney, Sydney, New South Wales, Australia, ⁴Discipline of Medical Radiation Sciences, Faculty of Health Sciences, The University of Sydney, Sydney, New South Wales, Australia

16.4T high-resolution diffusion tensor imaging (DTI) of fixed prostate cancer biopsy has revealed distinct microscopic diffusion environments and tissue architecture consistent with that seen on light microscopy. In this study, we present a new methodology for the visualization of fibromuscular stromal matrix from a prostate biopsy using high-resolution DTI-tractography.

Wenchao Cai¹, Feiyu Li¹, Jing Wang², Jue Zhang^2,3, Xiaoying Wang^1,2, and Xuexiang Jiang^1
1Department of Radiology, Peking University First Hospital, Beijing, Beijing, China, ²Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, Beijing, China,³College of Engineering, Peking University, Beijing, Beijing, China

Introduction:Pulsed arterial spin labeling (PASL) MRI is a non-invasive imaging tool capable of quantitatively measuring the microvascular perfusion characteristics of tissue. Purpose:We applied the PASL technique to detect the prostate cancer and to compare the differences blood flow(BF) values between the malignant and normal prostate peripheral zones. Methods:Six patients with pathologically confirmed prostate cancer were recruited to undergo the PASL examination with different invertion time(TI 1000/1200/1400/1600msec).Results: The mean BFs determined by PASL MRI with different TI in the prostate cancer were significantly higher than those in noncancerous regions (P<0.05, Paired T-test).Conclusion: This study demonstrates that PASL sequence can be used to evaluate the difference of BF value between cancer and noncancerous tissue in prostate.

Maysam Jafar¹, Sharon Giles², Veronica Morgan², and Nandita deSouza^3
1Clinical Magnetic Resonance, Institute of Cancer Research, London, Surrey, United Kingdom, ²Clinical Magnetic Resonance, Royal Marsden NHS Foundation Trust, London, United Kingdom, ³Clinical Magnetic Resonance, Institute of Cancer Research and Royal Marsden NHS Foundation Trust, London, United Kingdom

In prostate cancer DWI is additionally proving useful as a predictive biomarker of disease aggressiveness but severe artefacts and low resolution can hamper accurate quantification and reduce value of the biomarker particularly where suspected lesions are <1cm2. High-resolution DWI techniques potentially improve detection of small lesions, but suffer from reduced SNR compared to conventional techniques. The purpose of this study therefore was to investigate the differences in estimated ADC values in normal prostate and prostate cancer derived from a conventional diffusion-weighted protocol vs. a high-resolution protocol.

Mohammad Haris¹, Kavindra Nath², Anup Singh¹, Kejia Cai¹, David S Nelson², Dennis B Leeper³, Hari Hariharan¹, Jerry D Glickson², and Ravinder Reddy^1
1Radiology, University of Pennsylvania, Philadelphia, Pennsylvania, United States, ²Molecular Imaging Lab, University of Pennsylvania, ³Radiation Oncology, Thomas Jefferson University Hospitals

In this preliminary study, we explored the chemical saturation transfer effect from amino acids particularly glutamate and alanine present in the mouse model of melanoma and image them at high spatial resolution and discuss the potential significance of using this method in the management of human melanoma.

Weiguo Li¹, Zhuoli Zhang¹, Jodi Nicolai¹, Guang-Yu Yang², Reed Omary¹, and Andrew Larson^1
1Radiology, Northwestern University, Chicago, IL, United States, ²Pathology, Northwestern University, Chicago, IL, United States

Magnetization transfer (MT) MRI measurements were performed in 3 pancreatic ductal adenocarcinoma (PDAC) mouse xenograft models. For each of 28 PDAC xenografts, magnetization transfer ratios (MTR) were calculated and compared to histologic fibrosis levels from reference standard trichrome staining. MTR measurements were well correlated to quantitative fibrosis levels (r = 0.69, P = 0.01). Results indicated that MTR measurements offer the potential to serve as a valuable in vivo biomarker of desmoplasia in PDAC.

Rajiv Ramasawmy^1,2, Simon Walker-Samuel¹, Adrienne Campbell¹, Sean P Johnson², Jack Wells¹, Rosamund B Pedley², and Mark F Lythgoe^1
1UCL Centre for Advanced Biomedical Imaging, Division of Medicine and Institute of Child Health, University College London, London, United Kingdom, ²Cancer Institute, University College London, London, United Kingdom

Vascular targeting therapies induce acute changes in tumour perfusion, and imaging biomarkers of response are of particular interest in the clinic. We present a novel use of a respiratory-triggered flow-sensitive alternating inversion recovery (FAIR) Look-Locker arterial spin labelling (ASL) technique to measure perfusion within orthotopic liver metastases. Perfusion within the metastases was measured to be significantly lower than in normal liver tissue. We argue that this technique is suitable for follow-up and repeated measurements to track patient response as it does not require injected contrast agents.

Louisa Bokacheva¹, Khushali Kotedia^1,2, Megan Reese¹, Carl Le¹, Jason Koutcher^1,3, and Sean Carlin^1,3
1Department of Medical Physics, Memorial Sloan-Kettering Cancer Center, New York, New York, United States, ²Baylor College of Medicine, Houston, Texas, United States,³Department of Radiology, Memorial Sloan-Kettering Cancer Center, New York, New York, United States

Six athymic nu/nu rats bearing HT29 human colorectal xenograft tumors (700 mm³) were imaged at 7 T twice and diffusion-weighted images were acquired at five b-values between 0 and 900 s/mm². After imaging, rats were sacrificed, tumors were excised and histologically analyzed to determine the necrotic fraction. ADC voxel maps were calculated using monoexponential equation. K-means clustering was applied to the ADC data pooled from all tumors and ADC maps were segmented into two clusters: (1) viable and (2) necrotic. The threshold ADC value between clusters was found to be 0.88x10^-3 mm²/s. The fractional area of the necrotic cluster 2 correlated well with the histological necrotic fraction (R² = 0.91), although for half of the tumors, necrotic fraction was slightly overestimated. The spatial agreement between the cluster maps and histological necrotic areas was better for tumors with larger contiguous necrotic areas than for tumors with multiple small necrotic regions.

Benedicte F Jordan¹, Julie Magat¹, Elif Ozel¹, Florence Colliez¹, Anne-Catherine Fruytier¹, Valerie Marchand¹, Lionel Mignion¹, Caroline Bouzin², Olivier Feron², and Bernard Gallez^1
1Biomedical Magnetic Resonance Group, Université Catholique de Louvain, Brussels, Belgium, ²Pole of Pharmacotherapy, Université Catholique de Louvain, Brussels, Belgium

There is a critical need for methods able to monitor dynamically and noninvasively tumor oxygenation. The purpose of the current work was to compare the MOBILE technique, a method developed to map variations in oxygenation based on the changes in the relaxation properties of the tissue lipids by exploiting the higher solubility property of oxygen in lipids than in water, with R2*, R1 H2O, and simultaneous quantitative oxygen measurements using fluorescence quenching fiber optic probes. Changes in tumor oxygenation were induced by an hyperoxic breathing challenge in order to determine correlations between the response assessed using each technique.

Laura Glass^1,2, Yann Jamin¹, Rani George³, Louis Chesler², and Simon P Robinson^1
1CRUK and ESPRC Cancer Imaging Centre, The Institute of Cancer Research, Sutton, London, United Kingdom, ²Paediatric Oncology, The Institute of Cancer Research, Sutton, London, United Kingdom, ³Pediatric Oncology, Harvard University School of Medicine, Dana Faber Cancer Institute, Boston, MA, United States

Neuroblastoma is the most common extra-cranial solid tumour of infancy. It is well established that MCYN amplified patients are strongly correlated with a poor prognosis and increased likelihood of disease relapse. More recently, patients with ALK mutated cases of neuroblastoma have been associated with a very poor prognosis and an ultra-high risk of disease relapse. This study uses R2* to identify the presence of an ALK mutation in transgenic murine models. If this phenomenon translates to clinical situation, multi-gradient echo sequences could be incorporated into diagnostic scanning in neuroblastoma to stratify patients for targeted therapy.

Jessica K. R. Boult¹, Lara Perryman^2,3, Alexa Jury^2,3, Gary Box³, Paola Porcari^1,4, Sergey Popov^2,3, Suzanne A. Eccles³, Chris Jones^2,3, and Simon P. Robinson^1
1CRUK and EPSRC Cancer Imaging Centre, The Institute of Cancer Research and Royal Marsden NHS Foundation Trust, Sutton, Surrey, United Kingdom, ²Division of Molecular Pathology, The Institute of Cancer Research, Sutton, Surrey, United Kingdom, ³CRUK Cancer Therapeutics Unit, The Institute of Cancer Research, Sutton, Surrey, United Kingdom, ⁴Physics Department, Sapienza University of Rome, Rome, Italy

Delineation of diffuse infiltrative gliomas, in which the blood brain barrier remains intact, using conventional Gd-DTPA-enhanced MRI can be problematic. We aimed to develop a model of diffuse infiltrative brain tumours and to interrogate it using a multiparametric MRI approach. Primary paediatric glioblastoma xenografts presented with a diffuse growth pattern, clinically relevant observations on T₂-weighted and FLAIR images, and demonstrated no signal change following Gd-DTPA or USPIO administration. Areas of invasion in MDA-MB-231 tumours in the brain correlated with regions of low fBV and Gd-DTPA-induced R₁. These tumours represent an ideal platform for evaluating novel therapeutics targeted towards invasive disease.

Julio Cárdenas-Rodríguez¹, Christy M Howison², Terry O Matsunaga³, and Mark D. Pagel^4
1Chemistry and Biochemistry, TheUniversity of Arizona Cancer Center, University of Arizona, Tucson, AZ - Arizona, United States, ²Arizona Research Labs, University of Arizona, Tucson, Arizona, United States, ³Radiology, University of Arizona, Tucson, Arizona, United States, ⁴Departments of Biomedical Engineering, Chemistry and Biochemistry, The University of Arizona Cancer, University of Arizona, Tucson, Arizona, United States

The Reference Region Model has been proposed as an alternative method for evaluating DCE-MRI results without requiring an AIF, yet this method cannot account for changes in blood flow and hematocrit that affect the evaluation of vascular permeability. Instead of comparing DCE of a single contrast agent in two tissues, we propose a new model that compares two contrast agents in a single tissue, termed the Reference Agent Model. This report describes the selective detection of two 19F contrast agents in vivo under dynamic and multislice conditions, and opens the possibility of estimating the relative permeability of two CA because this new model is independent of blood flow and hematocrit.

Natenael Semmineh¹, and C. Chad Quarles^1
1Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States

The use of DSC-MRI in brain tumors is known to be influenced by contrast agent extravasation, which can result in additional extravascular T₁ and T₂ * effects well after the contrast agent initial bolus has passed through tissue. By separating and quantifying these T₁ and T₂ * effects in the period of time after the contrast agent’s first pass we propose that a new metric, termed the extravascular susceptibility calibration factor (K_e), can be derived and used to assess cellular packing heterogeneity. In this study we demonstrate that K_e may be used to differentiate between tumor types with varying cellular features.

Septian Hartono¹, Tong San Koh², Quan Sing Ng², Richard Ong², Hung The Huynh², Sidney Yu³, Sotirios Bisdas⁴, Laurent Martarello⁵, and Choon Hua Thng^2
1Nanyang Technological University, Singapore, Singapore, ²National Cancer Centre, Singapore, Singapore, ³Singapore General Hospital, Singapore, ⁴Eberhard Karls University, Tübingen, Germany, ⁵F. Hoffmann-La Roche, Switzerland

DCE-MRI has been used extensively in preclinical and clinical trials as a biomarker of drug effect. It is commonly assumed that a decrease in blood volume and blood flow is related to drug effect. This study examines the pitfall of such an assumption and illustrates decrease in perfused blood volume in untreated tumor as derived by DCE-MRI

Kelly C McPhee¹, Jennifer HE Baker^1,2, Katayoun Saatchi³, Urs O Häfeli³, and Stefan A Reinsberg^1
1Physics and Astronomy, University of British Columbia, Vancouver, BC, Canada, ²Radiation Biology Unit, BC Cancer Research Centre, Vancouver, BC, Canada, ³Faculty of Pharmaceutical Sciences, University of British Columbia, Vancouver, BC, Canada

The bolus arrival time (BAT) is the time at which signal enhancement commences, following the injection of a contrast agent. Two MRI studies were performed one day apart, the first using Bayer Healthcare’s Gadovist, and the second with a high-molecular-weight contrast agent, Hyperbranched polyglycerol (HPG) derivatized with p-NH2-benzyl-DOTA. The BAT maps for the same tumour are compared to corresponding whole tumour histology sections showing necrotic and healthy tissue. Areas where BAT is delayed relative to surrounding tissue match necrotic areas in histology, with longer delays in the BAT of HPG-Gd than Gadovist. Thus, BAT is not only an important input parameter for pharmacokinetic modelling, but also a reproducible, phenomenological parameter in its own right.

Marc S Ramirez¹, Yunyun Chen², Stephen Y Lai², and James A Bankson^1
1Imaging Physics, The University of Texas M.D. Anderson Cancer Center, Houston, TX, United States, ²Head & Neck Surgery, The University of Texas M.D. Anderson Cancer Center, Houston, TX, United States

To promote the use of MRI for preclinical evaluation of experimental cancer therapies, cost should be reduced, scanner access should increase, and more animals should be scanned near a common longitudinal time point. A 20-coil array system permits a combination of multi-animal and parallel imaging accelerations to substantially improve the throughput of a preclinical DCE-MRI study of a thyroid tumor model compared to that which can be achieved with a commercially-available single-mouse hardware setup. A comparison of throughput, image quality, and consistency of resulting pharmacokinetic parameters are described. Furthermore, the logistics of simultaneously preparing and monitoring five animals are discussed.

James Russell Ewing¹, Madhava Aryal², Hassan Farhad Bagher-Ebadian¹, Swayamprava Panda³, Kishor Karki¹, Nagaraja Tavarekere⁴, Glauber Cabral⁵, Joseph Fenstermacher⁴, and Tom Mikkelsen^1
1Neurology, Henry Ford Hospital, Detroit, Michigan, United States, ²Physics, Oakland University, Rochester Hills, Michigan, United States, ³Neurology, Henry Ford Hospital, Detroit, United States, ⁴Anesthesiology, Henry Ford Hospital, Detroit, Michigan, United States, ⁵Neurology, Henry Ford Hospital, Detroit, m, United States

In DCE studies in a rat model of cerebral glioma, test-retest variability is examined when a Model Selection paradigm is employed

Stephanie L. Barnes^1,2, Jennifer G. Whisenant^2,3, Gregory D. Ayers⁴, and Thomas E. Yankeelov^1,2
1Radiology and Radiological Sciences, Vanderbilt University, Nashville, TN, United States, ²Institute of Imaging Sciences, Vanderbilt University, Nashville, TN, United States,³Chemical and Physical Biology, Vanderbilt University, Nashville, TN, United States, ⁴Cancer Biostatistics, Vanderbilt University, Nashville, TN, United States

MRI techniques have been developed that have the potential to evaluate the response of tumors to treatment in a more sensitive manner than physical change in tumor size. However, confident implementation of these methods for tumor evaluation requires knowledge of their repeatability. The focus of this work is the evaluation of the repeatability of DCE-MRI in mice. Both 128² and 64² acquisition matrices were evaluated, with 64² consistently demonstrating higher repeatability. Specifically in the 64² case, a statistically significant change in K^trans and v_e for a group of 12 mice is 7.3% and 3.9%, respectively.

Zhuoli Zhang¹, Weiguo Li¹, Nichole R Blatner², Kristen Dennis², Daniele Procissi¹, Khashayarsha Khazaie^2,3, and Andrew C Larson^1,4
1Radiology, Northwestern University, Chicago, IL, United States, ²Medicine, Northwestern University, Chicago, IL, United States, ³R. Lurie Comprehensive Cancer Center, Northwestern University, Chicago, IL, United States, ⁴Biomedical Engineering, Northwestern University, Evanston, IL, United States

The mouse model of hereditary intestinal cancer based on haploinsufficiency of the adenomatous polyposis coli gene (APCΔ468) has been widely validated for studying the pathophysiology and carcinogenesis of human disease in pre-clinical research settings. By providing a organ specific (colon) tumor it offers the opportunity to test new therapies and observe tumor progression and response in a “realistic” tissue microenvironment. In this study, we investigated the ability to use MR techniques to 1) rapidly and reliably detect colorectal tumors in the transgenic APCΔ468 mouse model and 2) identify tumor-specific and potentially therapeutically relevant imaging biomarkers.

Benjamin Lemasson¹, Stefanie Galbán², Fei Li¹, Kevin Heist¹, Alnawaz Rehemtulla^1,2, Craig Galbán¹, and Brian Ross^1
1Radiology, University of Michigan, Ann Arbor, Michigan, United States, ²Radiation Oncology, University of Michigan, Ann Arbor, Michigan, United States

The goal of this study was to evaluate the functional diffusion map (fDM) as a biomarker of tumor recurrence on an animal-by-animal basis following different doses of radiation (1G, 2G and 4G) on a genetically engineered murine GBM model. Tumor size and fDM were monitor daily for a week and then every two days. We observed a large inter- and intra-group variability of time-to-recurrence. The maximum fDM_rADC+ values measured during radiation therapy was correlated strongly to time-to-recurrence (R2=0.85). fDM_rADC+ technique may serve as a biomarker of tumor recurrence that is insensitive to the response heterogeneity typically observed in GBM.

Jean-Christophe Brisset¹, Benjamin A Hoff¹, Stefanie Galbán², Benjamin Lemasson¹, Alnawaz Rehemtulla², Craig J Galbán¹, and Brian D Ross^1
1Department of Radiology, University of Michigan, Ann Arbor, Michigan, United States, ²Department of Oncology, University of Michigan, Ann Arbor, Michigan, United States

Molecularly targeted therapies are seen as alternatives to chemotherapy in treating cancer. Although promising, typical volumetric based approaches for response assessment prove inadequate for targeted agents. This study aims to determine the effectiveness of DW-MRI at assessing response to an oral kinase inhibitor, Cabozantinib, after one week of treatment in a prostate bone metastasis model. We observed a significant increase in tumor ADC values following treatment by Cabozantinib, which correlated with the reduction in tumor growth rate. These results demonstrate the potential of quantitative ADC maps for assessing tumor response early following treatment completion.

Yu-Chun Lin^1,2, Chun-Chieh Wang³, Yi-Ping Lin³, Yun-Han Lin⁴, and Jiun-Jie Wang^5
1Department of Diagnostic Radiology, Chang Gung Memorial Hospital, Linkou, Taiwan, Taiwan, ²Department of Electrical Engineering, Chang Gung University, Taoyuan, Taiwan, Taiwan, ³Department of Radiation Oncology, Chang Gung Memorial Hospital, ⁴Molecular Imaging Center, Chung Gung Memorial Hospital, ⁵Department of Medical Imaging and Radiological Sciences, Chang Gung University

Transgenic adenocarcinoma of the mouse prostate (TRAMP)-C1 tumors were irradiated on half of the tumor. Treatment response was assessed by using DCE and BOLD MRI on the 6th day after radiotherapy. Results showed that the Giant cell formed on the irradiated portion of the tumor. The irradiated area showed increased Ktrans and reduced BOLD response to carbogen breathing as compared with those at non-irradiated area. The reduced BOLD response to carbogen is related to the decreased microvascular density, which was evidenced by immunohistochemistry.

Scott Jones^1,2, Anshuman Panda^1,2, Higinia Cardenes³, James Fletcher², Gary Hutchins², and Ulrike Dydak^1,2
1School of Health Sciences, Purdue University, West Lafayette, Indiana, United States, ²Department of Radiology and Imaging Sciences, Indiana University School of Medicine, Indianapolis, IN, United States, ³Department of Radiation Oncology, Indiana University School of Medicine, Indianapolis, IN, United States

The purpose of this study is to evaluate the diagnostic and/or prognostic value of 3D 31P MRSI to assess a spatially resolved early metabolic response to ionizing radiation in both healthy and malignant liver tissue. A 3D 31P MRSI protocol using a dual-tuned 1H/31P 8-channel phased array liver coil was used to obtain 31P metabolic information throughout the whole human liver at 3.0T. 31P MRSI data from patients with hepatocellular cancer acquired before and 24 hours after undergoing stereotactic body radiotherapy are presented and show clear differential metabolic responses to radiation between healthy and malignant tissue.

Stefanie J.C.G. Hectors¹, Igor Jacobs¹, Gustav J. Strijkers¹, and Klaas Nicolay^1
1Biomedical NMR, Department of Biomedical Engineering, Eindhoven University of Technology, Eindhoven, Netherlands

To advance the clinical application of HIFU treatment of malignant tumors, a time-efficient multislice multiparametric MRI protocol, consisting of quantification of T₁, T₂, ADC and MTR, was developed. Image clustering techniques were used to differentiate non-treated and HIFU-treated tumor tissue. Results showed that after HIFU ablation, a region emerged in which pixels were assigned to clusters identified as HIFU-treated tumor tissue. MRI parameter values in this region were significantly different from parameter values in tissue identified as non-treated tumor tissue. Furthermore, a strong correlation was observed between MRI-derived HIFU-treated tumor tissue fractions and histologically derived non-viable tumor tissue fractions.

Jessica K.R. Boult¹, Yann Jamin¹, Vivien Jacobs², Lesley D. Gilmour¹, Simon Walker-Samuel¹, Jane Halliday², Paul Elvin², Anderson J. Ryan², John C. Waterton², and Simon P. Robinson^1
1CRUK and EPSRC Cancer Imaging Centre, Division of Radiotherapy and Imaging, The Institute of Cancer Research and Royal Marsden NHS Foundation Trust, Sutton, Surrey, United Kingdom, ²AstraZeneca, Alderley Park, Cheshire, United Kingdom

The use of pharmacodynamic biomarkers, including imaging biomarkers, is now desirable for the evaluation of novel molecularly targeted therapeutics in oncology. Before their deployment in clinical trials, such imaging biomarkers require evaluation. Here we describe two preclinical studies in which emerging MRI biomarkers were correlated with histology to assess tumour response to Src kinase or VEGFR2 inhibition. In both studies, treatment resulted in pathological target inhibition, but the corresponding imaging biomarkers failed to show the anticipated change. Reporting of such negative preclinical imaging biomarker responses is informative for clinical trial design.

Septian Hartono^1,2, Tong San Koh^1,2, Quan Sing Ng², Richard Ong², Hung The Huynh², Sidney Yu³, Sotirios Bisdas⁴, Laurent Martarello⁵, and Choon Hua Thng^2
1Nanyang Technological University, Singapore, Singapore, ²National Cancer Centre, Singapore, Singapore, ³Singapore General Hospital, Singapore, ⁴Eberhard Karls University, Tübingen, Germany, ⁵F. Hoffmann-La Roche, Switzerland

We aim to determine if DCE MRI is a sensitive enough technique to detect changes in fractional intravascular volume and blood flow after administration of a single dose of bevacizumab. The results showed that the effects of bevacizumab are dose-related. A more gradual drop of blood volume was observed in group treated with bevacizumab 1mg/kg. On the other hand, a steep drop of blood volume was observed in group treated with bevacizumab 10mg/kg. The results suggested potential of DCE MRI as a sensitive biomarker of microvascular density.

Yosuke Otake¹, Koji Hirata¹, Yoshihisa Soutome¹, and Yoshitaka Bito^1
1Hitachi, Ltd., Central Research Laboratory, Kokubunji, Tokyo, Japan

A method for evaluating efficacy of ¹⁹F-containing drugs on the basis of an image of active metabolites obtained by ¹⁹F-MRI and a k-means class map of tumor status obtained by ¹H-MRI was developed. This method was used to evaluate the drug efficacy of a 5-FU in a tumor-bearing rat. The obtained image of active metabolites and the k-means map of the rat were evaluated by using binary classification. The binary-classification result shows the regions of positive effect and negative effect of the 5-FU in the rat. The developed method will therefore be a powerful tool for pharmacokinetics and pharmacodynamics research.

Rheal A. Towner¹, David L. Gillespie², Debra Saunders¹, Osama Abdullah³, Edward W. Hsu⁴, Andrea Schwager⁵, and Randy L. Jensen^2
1Advanced Magnetic Resonance Center, Oklahoma Medical Research Foundation, Oklahoma City, OK, United States, ²Huntsman Cancer Institute, University of Utah, Salt Lake City, UT, United States, ³Bioengineering, University of Utah, Salt Lake City, UT, ⁴Bioengineering, University of Utah, Salt Lake City, UT, United States, ⁵Interdepartmental Program in Neuroscience, University of Utah, Salt Lake City, UT, United States

Magnetic resonance imaging and spectroscopy was used to follow the regression of rat F98 gliomas following treatment with the nitrone compound, OKN-007. Following MR and bioluminescence image detection of F98 gliomas, OKN-007 (administered orally) was found to decrease tumor growth. MR spectroscopy analysis (metabolite/tCr (total creatine) peak area ratios, and LCModel) indicated F98 glioma-induced alterations in tumor metabolites (tCho (total choline), tCr, NAA (N-acetyl aspartate), Lip1.3 (methylene hydrogens in lipid acyl groups), Lip5.3 (olefinic hydrogens in unsaturated lipid acyl groups)), were found to revert back to normal levels following OKN-007 treatment. OKN-007 may be considered as a promising therapeutic addition or alternative for the treatment of human gliomas.

Rama Jayasundar¹, Gaurav Sharma¹, Shyam S Chauhan², and Thirumurty Velpandian^3
1Department of NMR and MRI, All India Institute of Medical Sciences, New Delhi, India, ²Department of Biochemistry, All India Institute of Medical Sciences, New Delhi, India,³Department of Occular Pharmacology, All India Institute of Medical Sciences, New Delhi, Delhi, India

The study evaluated the metabolic profiling, elemental analysis and anticancer potential on four polyherbal formulations namely KG, VK, MK and GTK using NMR, LIBS and MTT assay, respectively. Antioxident potential of formulations was carried out using FRAP assay. Amongst four formulations maximum cytotoxicity was observed for KG (IC50 22.48 lower case Greek mug/ml)followed by MK (IC50 27.58 lower case Greek mug/ml). Using FRAP assay maximum antioxidant potential was achieved for VK (17.0 lower case Greek muM Fe++g-1) amongst four. KG and MK formulations were selected for characterization through NMR Spectroscopy, which revealed the chemical indentification of metabolites to explore pharmacological action. LIBS data supports antioxident properties of formulations.

Sean Peter Johnson¹, Rajiv Ramasawmy², Adrienne Campbell², Mathew Robson¹, Mario Mazzantini¹, Vineeth Rajkumar¹, Barbara Pedley¹, Mark Lythgoe², and Simon Walker-Samuel^2
1Cancer Institute, UCL, London, United Kingdom, ²Centre for Advanced Biomedical Imaging, UCL, London, United Kingdom

Liver metastases are the major cause of mortality in patients with colorectal carcinoma (CRC). The vascular disrupting agent OXi4503 has been tested in phase I clinical trials of metastatic disease with R2* recommended as a biomarker of response. R2* response however may not accurately reflect response to treatment exclusively. Arterial spin labeling (ASL) is able to assess changes in tumour perfusion. We show that in a clinically relevant liver metastatic model of CRC (n=10 metastases) acquisition of combined R2* and ASL at baseline and 90min following 40mg/kg OXi4503 i.v. is more adequate at assessing response to treatment.

Jannie P Wijnen*^1,2, Kannie WY Chan*^1,3, Peter CM van Zijl^1,4, Michael T McMahon**^1,4, and Kristine Glunde**^1
1Russell H. Morgan Department of Radiology and Radiological Sciences, Johns Hopkins University School of Medicine, Baltimore, MD, United States, ²Department of Radiology, University Medical Centre Utrecht, Utrecht, Netherlands, ³Cellular Imaging Section and Vascular Biology Program, Institute for Cell Engineering, Baltimore, MD, United States,⁴F.M. Kirby Research Center for Functional Brain Imaging, Kennedy Krieger Institute, Baltimore, MD, United States

Magnetic Resonance Spectroscopy (MRS) detects high levels of total choline-containing metabolites (tCho) in malignant cancers, which is important for cancer diagnosis and therapy. ¹H MRS measures the tCho levels, but cannot resolve the overlapping signals of free choline (Cho), PC, and GPC that the tCho signal consists of. To develop novel imaging approaches with high molecular specificity, we have explored Chemical Exchange Saturation Transfer (CEST) to monitor exchangeable protons of phospholipid metabolites. The CEST contrast at 1.2 ppm correlated with the Cho level measured by high-resolution MRS in cell extracts, and could be used to assess malignancy in breast cancers. *Authors contributed equally ** Authors share the corresponding authorship

Maria Dung Cao¹, Lu Jiang², Tiffany R. Greenwood², Balaji Krishnamachary², Zaver M. Bhujwalla², Ingrid S. Gribbestad¹, and Kristine Glunde^2
1Department of Circulation and Medical Imaging, Norwegian University of Science and Technology (NTNU), Trondheim, Norway, ²Russell H. Morgan Department of Radiology and Radiological Sciences, The Johns Hopkins University School of Medicine, Baltimore, Maryland, United States

We previously demonstrated that the glycerophosphocholine phosphodiesterase (GPC-PDE) encoded by GDPD5 is partially responsible for the relatively low glycerophosphocholine (GPC) levels in human breast tumor samples, and that it is associated with cancer malignancy. Here we transiently silenced GDPD5 in human breast cancer cells using siRNA. Magnetic resonance spectroscopy revealed increased GPC levels following GDPD5 silencing, suggesting that GDPD5 catalyzes the degradation of GPC in human breast cancer cells. Total choline and phosphocholine levels increased in highly malignant MDA-MB-231 cells, but not in weakly malignant MCF-7 cells upon GDPD5 silencing, indicating different roles of GDPD5 in these two cell lines.

Balaji Krishnamachary¹, Marie-France Penet¹, Mayur M Gadiya¹, Noriko Mori¹, Yelena Mironchik¹, kristine glunde¹, and Zaver M Bhujwalla^1
1Radiology, JHU ICMIC Program, The Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins University, Baltimore, Maryland, United States

Synopsis: Stem-like breast cancer cells (SBCCs) are drug resistant, invasive, and likely to lead to recurrence and repopulation. High CD44 adhesion molecule expression and high expression of the drug transporter ABCG2 are two markers associated with populations enriched with SBCCs. Altered choline metabolism is one of the hallmarks of cancer. Cancers typically exhibit elevated phosphocholine mostly due to increased choline kinase expression and activity. Here we have examined the relationship between choline kinase and these two markers of SBCCs. Downregulating choline kinase expression resulted in decreased CD44 and ABCG2. These studies provide insight into potential metabolic targeting of SBCCs.

Caleb Roberts¹, Geoff J Parker¹, Ahmad Mirza², Andrew Jackson³, Yvonne Watson¹, Susan Cheung¹, Giovanni Buonaccorsi¹, and Josephine H Naish^1
1Biomedical Imaging Institute, The University of Manchester, Manchester, Greater Manchester, United Kingdom, ²Departments of Gastrointestinal Surgery and Histopathology, University Hospital of South Manchester, Manchester, Greater Manchester, United Kingdom, ³Clinical Oncology, University Hospitals Southampton NHS Foundation Trust, Southampton, Hampshire, United Kingdom

There is a need to develop imaging methods that can assess treatment response and help individualise therapy in patients with oesophageal cancer. We explore the utility of dynamic contrast-enhanced MRI in a pilot study of 5 patients with oesophageal cancer. Each patient underwent two 3D DCE-MRI scans, each separated by a course of radical chemo radiotherapy. All patients showed significant treatment response demonstrated by changes in post-treatment pharmacokinetic parameters. We have successfully applied 3D DCE-MRI in oesophageal cancer and demonstrated its potential to assess treatment response.

Saada A M Abujarada^1,2, James P B O'Connor^1,3, and Chris J Rose^1,2
1University of Manchester Biomedical Imaging Institute, Manchester, Greater Manchester, United Kingdom, ²University of Manchester Academic Health Science Centre, Manchester, Greater Manchester, United Kingdom, ³Department of Radiology, Christie Hospital NHS Trust, Manchester, Greater Manchester, United Kingdom

MRI-derived biomarkers, measured before treatment, may predict tumor response. Such research tends to use linear regression that may overlook differences in biomarker reliability (interscan repeatability). Where reliability is considered, it is often quantified separately to the regression task, rather than within the same conceptual framework. We describe a novel nonlinear model that explicitly models biomarker reliability and compare its predictions to those from a linear errors-in-variables model in a study of liver metastases treated with bevacizumab and FOLFOX6. While the median prediction error of the methods is equal, our method makes substantially fewer large errors.

E.G.W. ter Voert¹, L. Heijmen², C.J.A. Punt², A. Heerschap¹, and H.W.M. van Laarhoven^2
1Radiology, Radboud University Nijmegen Medical Centre, Nijmegen, Gelderland, Netherlands, ²Medical Oncology, Radboud University Nijmegen Medical Centre, Nijmegen, Gelderland, Netherlands

Fifty percent of the patients with colorectal cancer develop distant metastases, mainly in the liver. Early response monitoring is desirable as only a subset of patients responds to the potentially toxic and expensive systemic treatment. The aim of this study is to predict response to systemic treatment in patients with colorectal liver metastases using pre-treatment measurements of the apparent diffusion coefficient (ADC) and the T₂^*. The preliminary results indicate, after analyzing 79 lesions in 25 patients, that high pre-treatment ADC values and possibly low pre-treatment T₂^* values in tumor tissue predict good treatment outcome.

Peter Brotchie^1,2, and Shalini Amukotuwa^1
1MRI, Geelong Hospital, Geelong, Victoria, Australia, ²Radiology, University of Melbourne, Melbourne, Victoria, Australia

Local recurrence is a major obstacle in curative surgery for rectal carcinoma. The main factor affecting local recurrence is tumour involvement of the circumferential resection margin (CRM). If CRM is involved a long-course of chemoradiotherapy is given to reduce local recurrence rates. If not, a short-course of neoadjuvant radiotherapy decreases local recurrence in patients with transmural invasion of tumour. For these reasons, accurate pre-operative identification of transmural invasion and CRM involvement is needed. Rectal MRIs were performed on 101 consecutive patients with histologically proven rectal adenocarcinoma. Histopathologically, transmural invasion (T3 or T4) was present in 29 of 51 (T1 or T2) tumours and absent in 22 of 51. 25 of the 29 tumours with transmural invasion were correctly identified with 4 false negatives. 19 of the 22 tumours without transmural invasion were correctly identified with 3 false positives (Accuracy: 86%, Sens: 86%, Spec: 86 %, PPV: 89%, NPV: 83 %). Histopathologically CRM was involved in 7 of 48 tumours and not involved in 41 of 48 tumours. With MRI, CRM involvement was correctly identified in 7 of 7 cases and correctly identified as absent in 37 with 4 false positives (Sens 100%, Spec 90%, PPV 64%, NPV 100%).

Jadegoud Yaligar¹, Wei W Teoh², Swee Shean Lee¹, Kanaga Sabapathy², Philip William Kuchel¹, and S. Sendhil Velan^1
1Laboratory of Molecular Imaging, Singapore Bioimaging Consortium, Singapore, Singapore, Singapore, ²Laboratory of Molecular Carcinogenesis, National Cancer Center, Sinapore, Singapore, Singapore

Aflatoxin-B1 and hepatitis-B virus are the most common etiological factors associated with high incidence of hepatocellular carcinoma (HCC) in Asia. Epidemiological studies have also suggested a strong correlation between AFB1 exposure and a signature mutation on the tumor suppressor gene, p53. We have investigated the HCC model that closely mimics human HCC, induced by expression of the Hepatitis surface antigen as a transgene from the albumin promoter (HBsAg) after treatment with AFB. We performed in vivo MRI/MRS and in-vitro MAS experiments and analyzed longitudinal changes of diffusion and tissue biochemistry. We detected carnitine for the first time, to our knowledge, in HCC tumors of HBsAg mice. Therefore relatively high concentrations of carnitine may be a biomarker for the early detection of HCC.

Yanping Sun¹, Shakti Ramkissoon², Matthew A Theisen², Amy S Freund³, Kristen L Jones¹, Marika Hayashi², Juan Wang¹, Keith Ligon², and Andrew L Kung^1,4
1Lurie Family Imaging Center, Dana-Farber Cancer Institute, Boston, MA, United States, ²Center for Molecular Oncologic Pathology, Dana-Farber Cancer Institute, ³Bruker BioSpin Corp., Billerica, ⁴Pediatric Oncology, Children’s Hospital, Boston

Mutations in isocitrate dehydrogenase (IDH1/2) and excessive production of the 2-hydroxyglutarate (2HG) have been found in low grade gliomas and secondary glioblastomas (GBMs). In this study, we developed an intracranial xenograft GBM model harboring the IDH1-R132H mutation. MRI showed that T1 and T2 in tumor were higher, while tumor perfusion was lower, than normal tissue. 1D and 2D MRS demonstrated the presence of 2HG in IDH tumor but absent in control tumor. The tumor was characterized as a densely cellular glial neoplasm composed of large pleomorphic cells, which were positive by immunohistochemistry for the mutant form of IDH1

Jelena Lazovic¹, Horacio Soto², David Piccioni², Arthur Chou², Sichen Li², Robert Prins², Linda Liau², Timothy Cloughesy², Albert Lai², and Whitney Pope^3
1Radiology, University of California, Los Angeles, CA, United States, ²University of California, ³Radiology, University of California, Los Angeles

The presence of isocitrate dehydrogenase 1 (IDH1) R132 mutation in gliomas including secondary glioblastoma is associated with more favorable outcome and increased overall survival compared to patients with functional IDH1. The aim of this study was to investigate metabolic changes and tumor growth rates in U87 glioma cells transformed to express mutated IDH1-R132. We found IDH1-R132 overexpression to be associated with an increase in tumor growth rate. Metabolic changes associated with the IDH1-R132 mutation included elevated levels of 2-hydroxyglutaric acid and reduced glutamate levels.

Simon Walker-Samuel¹, Rajiv Ramasawmy¹, Francisco Torrealdea², Marilena Rega², Peter Johnson³, Vineeth Rajkumar³, Simon Richardson¹, Dave Thomas², Barbara Pedley³, Xavier Golay², and Mark F. Lythgoe^1
1Centre for Advanced Biomedical Imaging, University College London, London, United Kingdom, ²Institute of Neurology, University College London, United Kingdom, ³Cancer Institute, University College London, United Kingdom

GlucoseCEST has been shown to allow the detection of exogenously administered, unlabelled glucose in tumours and we have recently extended this technique to quantify glucose concentration in vivo. In turn, this has enabled us to perform pharmacokinetic modelling of regional tumour glucose uptake. This provided estimates of the rate of glucose extravasation from blood plasma and the fractional volume accessible to glucose, using blood plasma glucose measurements from direct blood sampling. The current study is a proof-of-principle demonstration and the results open the possibility for applying more complex multi-compartmental analysis to measure regional tumour glucose metabolism.