Larry M Cai¹, Jose L Izquierdo-Garcia¹, Myriam M Chaumeil¹, Pia Eriksson¹, Joanna J Phillips², Russell O Pieper², and Sabrina M Ronen^1
1Radiology, UCSF, San Francisco, CA, United States, ²Neurological Surgery, UCSF, San Francisco, CA, United States

The IDH1 mutation (IDHmut) is associated with low-grade gliomas. By taking advantage of glutamate labeling patterns from 2-13C-glucose, we show, using 13C MRS of two glioma cell lines, that IDHmut cells decrease flux through pyruvate dehydrogenase (PDH) compared to pyruvate carboxylase (PC). In both cell lines, this is accompanied by a decrease in PDH expression and activity, an increase in PC expression and activity, and an increase in PDH inhibitory phosphorylation. Taken together, these results reveal potential sites of metabolic reprogramming in IDHmut gliomas.

Victor D. Schepkin¹, Malathy Elumalai², Jason Kitchen³, Chunqi Qian⁴, Peter Gor'kov¹, and William Brey^1
1CIMAR, NHMFL/FSU, Tallahassee, FL, United States, ²AMRIS, NHMFL/UF, Gainesville, FL, United States, ³CIMAR, NHMFL/FSU, Tallahassee, Florida, United States, ⁴NINDS/NIH, Bethesda, MD, United States

Chloride is a unique “window” for investigating brain function and cancer development. The in vivo challenges of low sensitivity, short T2 relaxation time and the small size of the acquisition matrix were evaluated here. The experiments in normal rat brain and glioma revealed: chloride in vivo is as visible as the sodium MR signal. However, the bi-exponential FID and limited sampling size dramatically affect the quantification of images and needs to be corrected according to the developed Matlab model. The finding of increased chloride concentration in glioma correlates with the hypothesis stating a critical role of chloride for tumor progression.

Giselle A. Suero-Abreu^1,2, G. Praveen Raju³, Orlando Aristizabal¹, Eugenia Volkova¹, Edward J. Houston¹, Diane Pham³, Alexandre Wojcinski⁴, Kamila U. Szulc¹, Daniel Colon¹, Alexandra L. Joyner⁴, and Daniel H. Turnbull^1,2
1Skirball Institute of Biomolecular Medicine, NYU School of Medicine, NY, NY, United States, ²Department of Radiology, NYU School of Medicine, NY, NY, United States, ³Department of Pediatrics, Weill Cornell Medical College, NY, NY, United States, ⁴Developmental Biology Program, Sloan-Kettering Institute, Memorial Sloan-Kettering Cancer Center, NY, NY, United States

Mouse models of medulloblastoma have led to important new insights into the etiology of this common malignant pediatric brain tumor. In the current study, Mn-enhanced MRI (MEMRI) was used to characterize tumor progression in mice with a conditional knockout of Ptch1, a mouse model of sporadic medulloblastoma. 3D MEMRI enabled early detection of pre-neoplastic lesions (validated by histology), and longitudinal MEMRI was used to quantify tumor progression in individual mice. Measured tumor growth rates were heterogeneous, leading to the interesting future potential of MEMRI for guiding histological and micro-array analyses of molecular differences between fast and slow progressing medulloblastomas.

Andrea Bianchi¹, Damien Moncelet¹, François Lux², Emeline Julie Ribot¹, Nawal Tassali¹, Veronique Bouchaud¹, Olivier Tillement², Pierre Voisin¹, and Yannick Crémillieux^1
1Centre de Résonance Magnétique des Systèmes Biologiques, Université Bordeaux Segalen, Bordeaux, Bordeaux, France, ²Institut Lumière Matière, Université Claude Bernard, Lyon, France

Glioblastoma is the most aggressive, common and lethal brain tumor. In this context, new noninvasive methods for early detection and therapy are needed in order to improve the prognosis of this pathology. We present here an in vivo MRI longitudinal study of brain cancer detection in tumor-bearing immunodeficient mice through intratracheally- and intravenously- administered multimodal Ultra-Small Rigid Platforms. In this study we showed for the first time that the synergic employment of a strongly T1-weighted MRI UTE sequence and intratracheally-administered gadolinium-based nanoparticles allow the high-precision detection of brain tumor and of its contours.

Marina Radoul¹, Myriam M Chaumeil¹, Pia Eriksson¹, Jose L Izquierdo Garcia¹, and Sabrina M Ronen^1
1Radiology and Biomedical Imaging, UCSF, San Francisco, CA, United States

PI3K/Akt/mTOR, one of the important signaling pathways, is activated in ~88% of GBM and different steps in this pathway can serve as therapeutic targets. This study demonstrates non-invasive monitoring of metabolic response to treatment with XL765/SAR245409, a novel dual PI3K/mTOR inhibitor. The inhibition in PI3K/Akt/mTOR signaling pathway leads to changes in levels of 13C MRS-detectable hyperpolarized lactate production from pyruvate and levels of 1H MRS-detectable total choline. Importantly, metabolic changes are associated with significantly longer survival in the case of XL765/SAR245409-treated animals, independent of tumor size.

Chaohsiung Hsu¹, Zhao Li¹, and Yung-Ya Lin^1
1Chemistry and Biochemistry, UCLA, Los Angeles, CA, United States

Veerle Kersemans¹, Philip D Allen¹, John S Beech¹, Stuart Gilchrist¹, Paul Kinchesh¹, and Sean C Smart^1
1Gray Institute for Radiation Oncology and Biology, Department of Oncology, University of Oxford, Oxford, United Kingdom

The application of prospective cardio-respiratory gating for quantitative DCE-MRI of abdominal and thoracic tumours was investigated. The method was made feasible through the use of cardio-respiratory synchronisation techniques in conjunction with RF calibrations and accurate temporal sampling. As a result, multiple chest tumours which are highly susceptible to motion corruption during DCE-MRI protocols could be screened simultaneously and classified using quantitative DCE-MRI parameters in a manner that is compatible with high-throughput operation. Moreover, volume imaging permitted visualisation and analysis of multiple breast tumours in the chest during the same acquisition, avoiding operator dependent slice pre-selection errors.

Blanca Lizarbe¹, Ana Amor-López¹, Sebastián Cerdán¹, and Pilar López-Larrubia^1
1Instituto Investigaciones Biomédicas "Alberto Sols" CSIC-UAM, Madrid, Madrid, Spain

Brain tumors are associated with tissue inflammation and microvasculature alterations that are MRI detectable. DWI and PWI techniques have been widely used to provide information about blood flow/volume, edema and cellularity. Nevertheless, a better understanding of the relationship between inflammation and microvascular changes during tumoral growth needs to be established. In this work, we present a multimodal evaluation of gliobastoma growth in rat brains using DWI with high and low b-weightings, and three models of diffusion –with or without perfusion effects- fitting. Our results indicate lower water diffusion behavior but higher perfusion contributions in the early stages of tumor growth.

Donna Murrell^1,2, Robbert van Gorkum¹, Amanda Hamilton¹, Christiane Mallett¹, Brunilde Gril³, Ann Chambers^2,4, Patricia Steeg³, and Paula Foster^1,2
1Imaging Research Laboratories, Robarts Research Institute, London, Ontario, Canada, ²Medical Biophysics, Western University, London, Ontario, Canada, ³National Cancer Institute, Maryland, United States, ⁴London Regional Cancer Program, London, Ontario, Canada

Few preclinical models exist to study HER2+ brain metastatic breast cancer. Here, we employ MRI techniques and correlative histology to characterize three murine models of HER2+ brain metastatic breast cancer (SUM190-BR3, JIMT1-BR3, 231BR-HER2). We use 3D anatomical MRI of the mouse brain to illustrate the incidence, distribution and size of brain metastases and contrast-enhanced MRI that provides information about the integrity of the blood-tumour barrier (BTB). Our findings reflect the substantial heterogeneity of this disease; understanding the imaging appearance and underlying biology of these tumours is vital to early diagnosis and advancements in treatment strategies.

Mor Mishkovsky^1,2, Cristina Cudalbu³, Emine Can⁴, Denis Mario⁵, Ivan Radovanovic⁵, Arnaud Comment⁴, Virginie Clément-Schatlo⁵, and Rolf Gruetter^1,6
1Laboratory for Functional and Metabolic Imaging (LIFMET), Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland, ²Department of radiology, Univesity of Lausanne, Lausanne, Switzerland, ³CIBM, Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland, ⁴Institute of Physics of Biological Systems, Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland, ⁵Hôpitaux Universitaires de Genève (HUG), Geneva, Switzerland, ⁶Department of radiology, Univesity of Lausanne and Geneva, Switzerland

Glioblastoma tumorigenesis and its effect on cerebral metabolism were studied longitudinally in a rat model of human glioma initiation cells (GIC). In vivo 1H MRS spectra were measured to characterize brain metabolism and diffusion tensor imaging (DTI) acquisitions allowed to visualize morphological changes between healthy and malignant tissue. Both DTI and 1H MRS indicated the onset of the tumor at similar time point, yet at this early stage the differences between the tumors and the contralateral hemisphere were more evident in the spectroscopic data.

Wenlian Zhu¹, Yoshinori Kato^1,2, and Dmitri Artemov^1,2
1Division of Cancer Imaging Research, The Russell H. Morgan Department of Radiology and Radiological, The Johns Hopkins University School of Medicine, BALTIMORE, Maryland, United States, ²Department of Oncology, The Sidney Kimmel Comprehensive Cancer Center, The Johns Hopkins University School of Medicine, Maryland, United States

Vasculature of solid tumors is intrinsically heterogeneous, which presents challenges to antiangiogenic intervention as well as the evaluation of its therapeutic efficacy. Here we evaluated the tumor vascular changes in response to bevacizumab/paclitaxel therapy using a combination approach of MR angiography and DCE-MRI method. Results showed that macroscopic feeding vessels were not affected by the bevacizumab/paclitaxel treatment. A higher portion of the tumors was within close proximity of these large vessels after the treatment, concomitant with tumor growth retardation. A significant decrease in microvascular permeability and vascular volume in regions near these macroscopic vessels was observed.

Lauriane Juge^1,2, Anne Petiet³, Simon A. Lambert², Pascal Nicole², Simon Chatelin^2,4, Sabrina Doblas², Valerie Vilgrain^2,5, Bernard E. Van Beers^2,5, and Ralph Sinkus^2,6
1Neuroscience Research Australia, Randwick, Sydney, NSW, Australia, ²CRB3-INSERM U773, University Paris Diderot, Paris, France, ³IFR02-CEFI, University Paris Diderot, Paris, France, ⁴Laboratoire Ondes et Acoustique / Institut Langevin, ESPCI, Paris, France, ⁵Radiology, Beaujon Hospital, Clichy, Paris, France, ⁶Division of Imaging Sciences and Biomedical Engineering, King’s College London, King’s Health Partners, St. Thomas’ Hospital, London, United Kingdom

Disease or therapies can change the mechanical integrity and organization of vascular structures. If blood vessels represent a source for wave scattering, Magnetic Resonance Elastography (MRE) should be able to sense these changes. Considering the hypothesis that the presence of an underlying fractal-like stiff structure is capable of generating on the macroscopic scale power law behavior, multi-frequency MRE (100-150Hz) was performed to quantify alteration of the shear wave speed due to the presence of vascular outgrowth using a rat aortic ring model. Results support the ability of using shear wave diffusion parameters to probe the structure of the vascular bed.

Firas Moosvi¹, Jennifer H.E. Baker^2,3, and Stefan A. Reinsberg^1
1Physics and Astronomy, University of British Columbia, Vancouver, BC, Canada, ²BC Cancer Research Institute, Vancouver, BC, Canada, ³Physics and Astronomy, University of British Columbia, BC, Canada

Results from a retrospective reproducibility analysis of basic MR parameters in over 50 control tumour-bearing mice. Parameters such as baseline T1, AUC, and AUGC are considered in three tumour cell lines.

Anna M. WANG^1,2, Karrie Mei-Yee Kiang³, GK Leung³, Adrian Tsang^1,2, Victor B. Xie^1,2, Hua Guo⁴, and Ed X. Wu^1,2
1Laboratory of Biomedical Imaging and Signal Processing, The University of Hong Kong, Hong Kong, Hong Kong, ²Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong, Hong Kong, ³Department of Surgery, The University of Hong Kong, Hong Kong, Hong Kong,⁴Center for Biomedical Imaging Research, School of Medicine, Tsinghua University, China

This study explored the capability of Diffusion Weighted Magnetic Resonance Spectroscopy (DW-MRS) for the separation and quantification of the overlapped mobile lipid and lactate signal at 1.3ppm. Both the content and ADC value can be computed from fitting the diffusion weighted signal to a bi-exponential decay model. In this rat model of intracerebral C6 glioma, the spectra from the tumor region was dominated by the mobile lipid signal and the lipid signal intensity is approximately ten times higher than the lactate signal. Our result also shows the lactate ADC in C6 glioma is 2.9(±0.9)×10–4 mm2/s and the lipid ADC is 3.3(±1.3)×10–4 mm2/s. Demonstrated by this study, the DW-MRS provides a feasible way to solve the overlapping problem of the lactate and mobile lipid peak at 1.3ppm, giving an alternative method for the quantification of lipid and lactate content in the clinical study.

Mrignayani Kotecha¹, Shreyan Majumdar¹, Eugene Barth², Boris Epel², and Howard Halpern^2
1Department of Bioengineering, University of Illinois at Chicago, Chicago, IL, United States, ²Center for EPR Imaging in Vivo Physiology, Department of Radiation and Cellular Oncology, University of Chicago, Chicago, IL, United States

The purpose of this work is to establish a relationship between tumor viscosity and partial oxygen pressure (pO2), the two important physiologic parameters that can be channelized to provide targeted radiation therapy. Tumors have a highly heterogeneous environment frequented with areas of low oxygen concentration (hypoxic regions). These hypoxic areas are resistant to radiation and thus, require higher radiation dosage for the destruction of tumor cells. Current practice of ignoring oxygen distribution while applying homogeneous radiation treatment leads to excessive damage of the neighboring healthy tissues, and thereby reduced quality of patient life. Solid tumors have abnormal organization of blood vessels that results in heterogeneous perfusion and extravasation, and a hostile microenvironment with increased interstitial pressure (1). The higher cellularity, tissue disorganization, and increased extracellular space all result in lower apparent diffusion coefficients, equivalent to higher viscosities, for malignant tumors as compared to normal tissue (2). The knowledge of pO2, in conjunction with viscosity and tissue anisotropy, can predict tissue health and may eventually be used in combination with Intensity-Modulated Radiation Therapy (IMRT) for targeted destruction of radiation-resistant areas, while sparing healthy tissues. In this study, we aim to correlate tumor viscosity acquired using diffusion weighted magnetic resonance imaging (DWI) with pO2 obtained by electron paramagnetic resonance oxygen imaging (EPROI). This is first such study correlating these two physiologic parameters at the tissue microstructure level.

Lauriane Juge¹, Miguel Albuquerque², Mouniya MEBARKI², Simon A. Lambert², Sabrina Doblas², Shaokoon Cheng^3,4, Lynne E. Bilston^3,5, Valerie Paradis^2,6, Valerie Vilgrain², Bernard E. Van Beers², and Ralph Sinkus^2
1Neuroscience Research Australia, Randwick, Sydney, NSW, Australia, ²CRB3-INSERM U773, University Paris Diderot, Paris, France, ³Neuroscience Research Australia, Randwick, Sydney, Australia, ⁴Engineering, Macquarie University, North Ryde, Sydney, Australia, ⁵Prince of Wales Clinical School, university of New South Wales, Kensington, Sydney, Australia, ⁶Pathology, Beaujon Hospital, Paris, France

Early detection of changes in the vascularity and cellularity of a tumor could represent a significant advance in treatment management using anti-angiogenic agents. We investigated the potential value of MR-Elastography (800, 900 and 1000 Hz) and Diffusion Weighted MR imaging (6 b-values from 0 to 1000 s/mm²) in the detection of microstructural changes induced by the therapy (Sorafenib, Nexavar ®) in a human liver cancer cell line (HepG2) implanted in immune-deficient mice.. Results showed that potentially, only the biomechanical dispersion properties were sensitive to the changes induced by the anti-angiogenic treatment, while the apparent diffusion properties were not altered.

Jae Mo Park¹, Sui-Seng Tee¹, Ralph Hurd², Kyle Brimacombe³, Matthew Boxer³, Dirk Mayer⁴, Brian Rutt¹, and Daniel Spielman^1
1Radiology, Stanford University, Stanford, CA, United States, ²GE Healthcare, Menlo Park, CA, United States, ³National Center for Advancing Translational Sciences, National Institutes of Health, Bethesda, MD, United States, ⁴Diagnostic Radiology and Nuclear Medicine, University of Maryland, MD, United States

PKM2, the M2 isoform of pyruvate (Pyr) kinase, plays a role in the last step of glycolysis, converting phosphoenolpyruvate (PEP) to Pyr. So far, PKM2 is expressed in all tested cancer cells. We hypothesized that the use of a glucose analogue, 2-deoxyglucose (2DG), in combination with the PKM2 activator will accelerate the uptake of the toxic 2DG in tumors, and observed the therapeutic response in Pyr metabolism of tumor-bearing mice using hyperpolarized 13C Pyr MRSI. Lactate (Lac)-to-Pyr ratio consistently increased in all mice with dual-treatment while single treated mice did not. It suggests that there might be a synergic anti-cancer mechanism of the PKM2 activator and 2DG, and accelerates glucose starvation in tumors.

Egidio Iorio¹, Fabio Ginnari Satriani¹, Alessandro Ricci¹, Emiliano Surrentino¹, Marina Bagnoli², Paola Alberti², Franca Podo¹, Delia Mezzanzanica², and Rossella Canese^1
1Cell Biology and Neurosciences Dept, Istituto Superiore di Sanità, Rome, Italy, ²Experimental Oncology Dept, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy

Epithelial ovarian cancer (EOC) is a heterogeneous disease with a poor prognosis. Evaluation of metabolic effects of anticancer therapies would enhance the capability of non invasive imaging approaches to monitor molecular mechanisms underlying tumour responsiveness. Here we explore the role of MRI/MRS in the detection of the cytotoxic response of trabectedin (ET-743, a new marine-derived antitumor agent, which has shown in vitro and in vivo activity in ovarian cancer) in experimental EOC models, showing previously unexplored trabectedine-induced metabolic and morphofunctional changes.

Jeffrey David Steinberg¹, Justin Lee², Philipp Kaldis³, Jean-Pierre Abastado^2,4, and Valerie Chew^2
1Singapore Bioimaging Consortium, Agency for Science, Technology and Research, Singapore, Singapore, ²Singapore Immunology Network, Agency for Science, Technology and Research, Singapore, Singapore, ³Institute of Molecular and Cell Biology, Agency for Science, Technology and Research, Singapore, Singapore, ⁴Institut de Recherches Internationales Servier, Suresnes, France

Hepatocellular carcinoma (HCC) is the 6th most common cancer with the 3rd highest mortality worldwide. Prognosis for HCC is generally poor with a 5-year survival rate of only 15%. In this study a spontaneous HCC mouse model was treated using poly(I:C), a synthetic TLR3 ligand that activates NK cells and is cytotoxic to HCC cells. Using a 3D T1-weighted MPRAGE MRI sequence, the tumor progression could be monitored. Average tumor growth was 8 times and 183 times the pre-therapy tumor volume for the treatment and control mice respectively. Thus, poly(I:C) was effective in delaying HCC tumor growth.

Carole D. Thomas^1,2, Florent Poyer^1,2, Philippe Maillard^1,3, Mihaela Lupu^1,2, and Joel Mispelter^1,2
1Institut Curie, Orsay, France, ²INSERM U759, Orsay, France, ³CNRS UMR176, Orsay, France

The aim of this study was double. Firstly, to determine if photodynamic therapy associated with nitroglycerin ointment was able to induce a major cellular death on a tumoral line of retinoblastoma that was less responsive to treatment. Secondly to determine if extracellular MRI follow-up is able to give rapid information about tumor cells destruction. Nitroglycerin increased the photosensitzer concentration at the tumor level and hence the treatment efficiency. Sodium MRI monitored non-invasively the cellular destruction, showing a local increase of sodium concentration specific to extracellular sodium amount all over where cells were irreversible damaged by PDT.

Hyeon-Man Baek^1,2, Yun-Ju Lee¹, Gregory Hyung Jin Park¹, Eun-Hee Kim¹, Gyunggoo Cho¹, and Chaejoon Cheong^1
1Division of MR Research, Korea Basic Science Institute, Ochang, Chungbuk, Korea, ²Department of Bio-Analytical Science, University of Science & Technology, Yuseong-gu, Korea

This study represents, to our knowledge, the first in vitro measurements of absolute quantification of 2HG levels in IDH1/2 mutated tumors using high resolution 1H-900MHz (21.1 Tesla) NMR spectroscopy. Our analysis revealed that a significant increase in the concentrations of 2HG, Iso, Leu, Ala, Glu, Gln, Tau, m-Ins, and Gly and Tau were observed in the IDH1/2 mutated cells (P < 0.05). This result reflects that levels of amino acids and choline derivatives were altered in the IDH1/2 mutated cells, possibly associated with IDH gene mutation. However, our findings are not consistent with the previously published Mass spectroscopy results. Further studies are needed.

Lloyd Lumata¹, Chendong Yang², Bookyung Ko², Ralph J. Deberardinis², and Matthew E Merritt^1
1AIRC, UTSW Medical Center, Dallas, TX, United States, ²Children's Med. Ct. Res. Inst., UTSW Med. Ctr., Dallas, TX, United States

Glutamine metabolism can satisfy both energetic and biosynthetic demands of cancer cells. Glutamine oxidation can be accentuated when glucose metabolism is blocked by any intervention. Here the action of an inhibitor of mitochondrial pyruvate transport is studied with hyperpolarized (HP) pyruvate. The inhibitor blocked formation of HP alanine and bicarbonate in a glioblastoma cell line while leaving lactate formation largely unperturbed. Further analysis showed that when pyruvate transport is inhibited glutamate dehydrogenase is upregulated, resulting in increased glutamine oxidation.

Mounia Beloueche-Babari¹, Carol Box², Harry G Parkes¹, Melanie Valenti², Alexis De Haven Brandon2², Liz Jackson¹, Vaitha Arunan¹, Sue Eccles², and Martin O Leach^1
1CRUK & EPSRC Cancer Imaging Centre, Institute of Cancer Research, Sutton, Surrey, United Kingdom, ²CRUK Cancer Therapeutics Unit, Institute of Cancer Research, Sutton, Surrey, United Kingdom

Acquired resistance to molecular therapeutics, including tyrosine kinase inhibitors (TKIs), is a key challenge in personalized cancer medicine. Identifying mechanisms and biomarkers of resistance could help detect patient relapse early and improve disease management. Here we show that acquired resistance to multiple epidermal growth factor receptor (EGFR) TKIs in human head and neck squamous cell carcinoma is associated with altered glycolytic, choline phospholipid and amino acid metabolism as detected by 1H MRS of cell line and xenograft tumor extracts. Such effects could provide potential metabolic imaging biomarkers of acquired resistance to EGFR TKIs

Sveva Grande¹, Alessandra Palma¹, Antonella Rosi¹, Anna Maria Luciani¹, Mauro Biffoni², Lucia Ricci-Vitiani², Daniele Runci², Roberto Pallini³, Laura Guidoni⁴, and Vincenza Viti^4
1Dipartimento di Tecnologie e Salute, Istituto Superiore di Sanità and INFN Sanità Group, Roma, Italy, Italy, ²Dipartimento di Ematologia, Oncologia e Medicina Molecolare, Istituto Superiore di Sanità, Italy, Italy, ³Dipartimento di Neurochirurgia, Università Cattolica del Sacro Cuore, Roma, Italy, Italy,⁴INFN Sanità Group, Roma, Italy, Italy

High recurrence rate and failure of conventional treatments in patients with glioblastoma multiforme is attributed to the presence of stem-like cells in these tumors. A dramatic accumulation of α-aminoadipate (αAAD) has been detected in some glioblastoma stem-like cells derived from primary Glioblastoma grade IV by means of 1H NMR and it was related to tumor aggressiveness.. The study suggests a role of αAAD as biomarker of cancer. Expression and high activity of ALDH7A1 could be envisaged in these cells. The indication that, similarly to prostate cancer, ALDH7A1 activity in glioblastoma may correlate with tumour invasiveness, is of potential diagnostic importance.

Ji Soo Choi¹, Hyeon-Man Baek², Suhkmann Kim³, Min Jung Kim⁴, Hee Jung Moon⁴, and Eun-Kyung Kim^4
1Samsung Medical Center, Seoul, Seoul, Korea, ²Korea Basic Science Institute, Chungbuk, Korea, ³Pusan National University, Busan, Korea, ⁴Yonsei University College of Medicine, Seoul, Korea

We performed metabolic profiling of 37 core needle biopsy samples collected from locally advanced breast cancer before neoadjuvant chemotherapy (NAC) using HR-MAS MRS. Various metabolites including choline-containing compounds were identified and quantified by HR-MAS MRS in all tissue samples. In multivariate analysis, the OPLS-DA models built with HR-MAS MR metabolic profiles showed visible discrimination between the pathologic response groups. This study showed OPLS-DA multivariate analysis using metabolic profiles of pretreatment CNB samples assessed by HR- MAS MRS may be used to predict pathologic response before NAC.

Natsuko Onishi¹, Shotaro Kanao¹, Masako Kataoka¹, Mami Iima¹, Rena Sakaguchi¹, Makiko Kawai¹, Tatsuki Kataoka², Yoshiki Mikami², Masakazu Toi³, and Kaori Togashi^1
1Department of Diagnostic Imaging and Nuclear Medicine, Kyoto Univerisity Graduate School of Medicine, Kyoto, Kyoto, Japan, ²Department of Diagnostic Pathology, Kyoto Univerisity Graduate School of Medicine, Kyoto, Kyoto, Japan, ³Department of Breast Surgery, Kyoto Univerisity Graduate School of Medicine, Kyoto, Kyoto, Japan

ADC is known to have inverse correlation with cellularity. Considering that cellularity is linked with prognosis in breast mucinous carcinoma (MBC), we examined the association between ADC and Ki-67 index (a marker of tumor proliferation) in MBC comparing with invasive carcinoma of no special type. ADC showed inverse correlation with cellularity(r=-0.802, p=<0.0001) and with Ki-67 index (r=-0.825, p=<0.0001) in MBC. The ability of ADC to classify highly proliferating MBC from low proliferating one was also demonstrated. ADC can be a promising non-invasive surrogate marker for Ki-67 index in the risk stratification of MBC.

Charles S Springer, Jr.¹, Xin Li¹, Luminita A. Tudorica², Karen Y. Oh², Nicole Roy², Stephen Y-C. Chui³, Arpana M. Naik⁴, Megan L. Holtorf⁵, Aneela Afzal¹, William D. Rooney¹, and Wei Huang^1
1Advanced Imaging Research Center, Oregon Health & Science University, Portland, Oregon, United States, ²Diagnostic Radiology, Oregon Health & Science University, Portland, Oregon, United States, ³Hematology/Oncology, Oregon Health & Science University, Portland, Oregon, United States, ⁴Surgical Oncology, Oregon Health & Science University, Portland, Oregon, United States, ⁵Clinical Trials Office, Oregon Health & Science University, Portland, Oregon, United States

Tumors exhibit metabolic heterogeneity. This requires individualized metabolic imaging with intra-tumor resolution. Shutter-speed pharmacokinetic analyses of DCE-MRI data provide tau_i, the mean intracellular water lifetime, with high-resolution. We present results for human breast cancer in vivo. Comparison of tau_i heterogeneity with that of other biomarkers and in response to therapy shows that tau_i is inversely related to on-going Na⁺/K⁺ATPase activity.

Mami Iima¹, Masako Kataoka¹, Masaki Umehana², Yuto Nakanishi², Takayuki Ito², Kojiro Yano³, Shotaro Kanao¹, Kaori Togashi¹, and Denis Le Bihan^4,5
1Diagnostic Imaging and Nuclear Medicine., Kyoto University Graduate School of Medicine, Kyoto, Kyoto, Japan, ²Kyoto University Faculty of Medicine, Kyoto, Kyoto, Japan, ³Information Science and Technology, Osaka Institute of Technology, Hirakata, Osaka, Japan, ⁴Human Brain Research Center, Kyoto University Graduate School of Medicine, Kyoto, Kyoto, Japan, ⁵Neurospin, CEA Saclay, Gif-sur-Yvette, Ile-de-France, France

A computer-assisted diagnostic tool was evaluated in 36 patients with breast lesions. The IVIM signal of each voxel was fitted using a kurtosis diffusion model and with the IVIM model. A parametric map was then generated by ascribing each voxel a value from 0 to 3 according to the number of parameters falling beyond a given threshold (K>0.80, ADCo<1.40 x 10-3mm²/s, fIVIM>2.07%) established from a previous study, and displayed using a color scale. The diagnostic accuracy of this computer-assisted diagnosis tool (3 or 2: malignant likely, 1 or 0: benign likely) was found very high (97 % sensitivity and 100 % specificity).

Elizabeth J Sutton¹, Elizabeth A Morris², Monica Morrow³, Michelle Stempel³, Amita Shukla-Dave⁴, Jung Hun Oh⁴, Joseph O Deasy⁴, and Yousef Mazaheri^5
1Radiology, Memorial Sloan-Kettering Cancer Center, New York, New York, United States, ²Radiology, Memorial Sloan-Kettering Cancer Center, New York, United States, ³Surgery, Memorial Sloan-Kettering Cancer Center, New York, United States, ⁴Medical Physics, Memorial Sloan-Kettering Cancer Center, New York, United States, ⁵Medical Physics, Memorial Sloan-Kettering Cancer Center, New York, New York, United States

Purpose: The purpose of this study is to correlate MR features (kinetics, morphology, and image texture) of triple-negative breast carcinomas (TNBC) on contrast-enhanced imaging with histological tumor grade.

Elizabeth AM O'Flynn¹, Matthew Blackledge², David Collins², Simon Doran², Hardik Patel³, Martin O Leach², and Dow-Mu Koh^4
1Clinical Magnetic Resonance, Insitute of Cancer Research and Royal Marsden Hospital, Sutton, Surrey, United Kingdom, ²Clinical Magnetic Resonance, Insitute of Cancer Research, Sutton, Surrey, United Kingdom, ³Radiology, Hammersmith Hospitals, London, United Kingdom, ⁴Radiology, Royal Marsden Hospital, Sutton, Surrey, United Kingdom

Computed diffusion weighted (DW) imaging (cDWI) calculates a high b value image from acquired DW MR images and can improve image quality and tumour detection by showing better suppression of benign tissues. 41 women underwent breast MRI, 20 patients with breast cancer and 21 normal cases. Breast images with a computed b value of 2000s/mm2 resulted in a higher overall diagnostic sensitivity of 82.4% compared to images acquired at a b value of 1150s/mm2 (sensitivity 17.7%) and equal sensitivity to DCE-MR of 82.4%. cDWI holds potential as an alternative fast, non-contrast diagnostic MR technique in breast cancer diagnosis.

Sara Vigano'^1,2, Andrew J. Patterson³, Mary McLean⁴, Elena Provenzano⁵, Louise Hiller⁶, Janet Dunn⁶, Anne-Laure Vallier⁷, Louise Grybowicz⁷, Reem Bedair⁸, Matthew G Wallis⁹, Martin J Graves¹⁰, Helena Earl¹¹, and Fiona J Gilbert^8
1Cambridge University Hospitals NHS Foundation Trust, Cambridge, United Kingdom, ²Universita' degli Studi di Milano, Milano, Italy, ³Radiology, Cambridge University Hospitals NHS Foundation Trust, Cambridge, United Kingdom, ⁴CRUK Cambridge Institute, University of Cambridge, United Kingdom, ⁵Department of Histopathology and Cambridge Breast Unit, Cambridge University Hospital NHS Foundation Trust and NIHR Cambridge Biomedical Research Centre, United Kingdom, ⁶Warwick Clinical Trials Unit, University of Warwick, Coventry, United Kingdom, ⁷Department of Oncology, Cambridge Cancer Trials Centre, Cambridge Breast Unit, Cambridge University Hospitals NHS Foundation Trust, United Kingdom, ⁸Radiology, University of Cambridge, Cambridge, United Kingdom, ⁹Cambridge Breast Unit and NIHR Biomedical Research centre, Cambridge University Hospitals NHS Foundation Trust, Cambridge, United Kingdom,¹⁰Radiology, Cambridge University Hospital NHS Foundation Trust and NIHR Cambridge Biomedical Research Centre, United Kingdom, ¹¹Department of Oncology, NIHR Cambridge Biomedical Research Centre and Cambridge Breast Unit, University of Cambridge, Cambridge, United Kingdom

Assessing and predicting neoadjuvant chemotherapy response is extremely important. Our aim was to compare changes in morphologic MR parameters and ADC values between responders and non-responders over time. Forty-five breast cancer patients had MRI at baseline, after the third cycle (mid-treatment), and at end-treatment. Tumor diameter, total volume and ADC values were compared between responders and non-responders. Changes in tumour volume and ADC significantly differed at end-treatment (p=0.007 and p<0.001), but at mid-treatment only ADC values showed significant changes (p=0.001). Morphologic parameters and ADC may be used in monitoring response but ADC may be an earlier predictor of pathological outcome.

Gene Young Cho^1,2, Linda Moy^1,3, Sungheon Kim¹, Ana Paula Klautau Leite⁴, Steven Baete¹, Jim Babb¹, Daniel K Sodickson¹, and Eric E Sigmund^1
1Radiology - Bernard and Irene Schwartz Center for Biomedical Imaging, New York University School of Medicine, New York, Select, United States, ²Sackler Institute of Graduate Biomedical Sciences, New York University, New York, NY, United States, ³Radiology, New York University Cancer Institute, New York, NY, United States, ⁴Radiology, Hospital das Clínicas, School of Medicine, University of São Paulo, São Paulo, SP, Brazil

An accurate, simple method using imaging to determine cancer type has long been a goal in oncology. Determination of tumor subtypes through imaging can be advantageous in the strategic planning of therapy and limit the need for invasive biopsy procedures. In MRI, intravoxel incoherent motion (IVIM) can play an important role as it is most commonly employed due to its sensitivity to tumor cell density and vascularity, both components of aggressiveness. In this study, highly sampled DWI data is used to perform IVIM analysis in a cohort of breast cancer patients in a 3T clinical scanner.

Lian Wang¹, Yiyi Chen², Alina Tudorica², Karen Oh², Nicole Roy², Mark Kettler², Dongseok Choi², and Wei Huang^2
1Providence Health and Services, Portland, Oregon, United States, ²Oregon Health & Science University, Portland, Oregon, United States

Pre-biopsy breast DCE-MRI pharmacokinetic parameters obtained from three institutions were supplied as inputs to a classification tree algorithm to identify imaging biomarkers and corresponding cut-off values for accutae breast cancer diagnosis. The results validate that the DeltaKtrans parameter is the single most accurate diagnostic marker among all DCE-MRI parameters. Incorporation of additional parameters in the classification tree approach further improves diagnostic sensitivity and specificity.

Noam Nissan¹, Edna Furman-Haran², Myra Shapiro-Feinberg³, Dov Grobgeld¹, and Hadassa Degani^1
1Department of Biological Regulation, Weizmann Institute of Science, Rehovot, Israel, ²Unit of Biological Services, Weizmann Institute of Science, Israel,³Meir Medical Center, Kfar Saba, Israel

In this study, we investigated the hormonal regulation of breast diffusion tensor imaging (DTI) throughout the menstrual cycle phases, during lactation and in post-menopause with and without hormonal replacement therapy (HRT). Our findings suggest that DTI parameters are not sensitive to the menstrual cycle changes, while menopause, long term HRT and the presence of milk in lactating women affected DTI parameters. Therefore, the timing for performing breast DTI is not restricted throughout the menstrual cycle, whereas the modulations in diffusion parameters due to HRT and lactation should be taken into account upon DTI evaluation

Peter Gibbs¹, Michael Fox¹, Martin Pickles¹, and Lindsay Turnbull^1
1MRI Centre, HYMS at University of Hull, Hull, East Yorkshire, United Kingdom

Statistical methods of texture analysis are widely used in image classification due to their computational ease and high level of discrimination. However, the most appropriate statistical method is unknown. In this work run length matrices have been calculated for a series of patients with locally advanced breast cancer prior to receiving neoadjuvant chemotherapy. Significant differences in run length based parameters were noted between low grade (I/II) and high grade (III) lesions pre-contrast and 5 minutes post contrast.

Cheng-Liang Liu¹, Matthew L Olson¹, Peixian Liu¹, Marko K Ivancevic², Constance D Lehman¹, and Savannah C Partridge^1
1Department of Radiology, University of Washington, Seattle, Washington, United States, ²Philips Healthcare, Best, Netherlands

Misalignment within DWI sequences due to eddy-current based distortions in diffusion gradient images reduces the accuracy of computed DWI and DTI parametric maps. Our study in 21 breast cancer patients showed that image-based distortion correction of DTI acquisitions improves spatial alignment and lesion conspicuity and may be essential for quantification of breast DTI parameters beyond ADC.

Jeon-Hor Chen^1,2, Shadfar Bahri¹, Rita S. Mehta³, Philip M. Carpenter⁴, and Min-Ying Su^1
1Center for Functional Onco-Imaging, University of California, Irvine, California, United States, ²Department of Radiology, Eda Hospital and I-Shou University, Kaohsiung, Taiwan, ³Department of Medicine, University of California, Orange, California, United States, ⁴Department of Pathology, University of California, Orange, California, United States

This study attempted to investigate if the change of biomarker status after adjuvant chemotherapy (NAC) will affect MR accuracy in evaluation of residual tumor size. 59 breast cancer patients who received NAC and had residual tumor size in MRI and pathology and pre- and post-NAC biomarker information were analyzed. Our study noted biomarker conversion following NAC did not impact accuracy of MRI in determining residual tumor size in Her-2 negative and PR positive breast cancer. In Her-2 positive cancer, when converted into Her-2 negative cancer, the MR-pathology tumor size difference was remarkably higher than Her-2 positive cancer without biomarker conversion.

Jose Ramon Teruel^1,2, Hans Erikssønn Fjøsne^3,4, Agnes Østlie⁵, Pål Erik Goa^2,6, and Tone Frost Bathen^1
1Department of Circulation and Medical Imaging, NTNU, Trondheim, Norway, ²St. Olavs University Hospital, Trondheim, Norway, ³Department of Surgery, St. Olavs University Hospital, Trondheim, Norway, ⁴Institute of Cancer Research and Molecular Medicine, NTNU, Trondheim, Norway, ⁵Department of Radiology, St. Olavs University Hospital, Trondheim, Norway, ⁶Department of Physics, NTNU, Trondheim, Norway

In our study we evaluate the potential of a 4th order diffusion tensor imaging (DTI) model for breast cancer differentiation. Our results reveal how this model outperforms standard DTI for differentiation of malignant and benign lesions, and healthy fibroglandular tissue. In particular, fractional anisotropy derived from this model is found to increase the potential of diffusivity markers to differentiate malignant and benign lesions.

Jeon-Hor Chen^1,2, Tsung-Lung Yang³, Huei-Lung Liang³, Chen-Pin Chou³, Jer-Shyung Huang³, Yifan Li¹, Min-Ying Su¹, and Huay-Ben Pan^3
1Center for Functional Onco-Imaging, University of California, Irvine, California, United States, ²Department of Radiology, Eda Hospital and I-Shou University, Kaohsiung, Taiwan, ³Department of Radiology, Kaohsiung Veterans General Hospital, Kaohsiung, Taiwan

Despite of the development of a new volumetric breast density (Quantra) analysis tool, still it is not known how accurate the acquired density results are. This study aimed to compare the results of density measurement using Quantra and 3D MRI in the same women. 56 women were studied. The MR density analysis was based on a novel semi-automatic method. Overall, moderate correlation (r=0.55 for the right breast and r=0.65 for the left breast) between the two modalities was noted. Huge measurement variations in the two modalities were noted in a few women, unexplained by the breast density or breast morphology.

Jeon-Hor Chen^1,2, Jia Shen Hong³, Po-Chuan Tseng¹, Peter T. Fwu¹, Celine M Vachon⁴, and Min-Ying Su^1
1Center for Functional Onco-Imaging, University of California, Irvine, California, United States, ²Department of Radiology, Eda Hospital and I-Shou University, Kaohsiung, Taiwan, ³Department of Medical Imaging, China Medical University, Taichung, Taiwan, ⁴Department of Health Sciences Research, Mayo Clinic, Rochester, Minnesota, United States

In this study we investigated the density distribution in the four quadrants of the breast and compare between a Western and an Asian cohort. Breast MRI from 250 Western women and 156 Asian women was semi-automatically segmented for the quantification of quadrant breast density. In total 91 right breasts and 65 left breasts from Asian cohort, and 144 right breasts and 106 left breasts from Western cohort were analyzed. Our study noted that Asian women have the most common highest density in the inner upper quadrant and Western women have the most common highest density in the outer lower quadrant.

David K Woolf¹, Sonia P Li¹, N. Jane Taylor², Andreas Makris¹, Andrew Gogbashian², Mark J Beresford³, Mei-Lin W Ah-See¹, J. James Stirling², David J Collins⁴, and Anwar R Padhani^2
1Academic Department of Oncology, Mount Vernon Hospital, Northwood, Middlesex, United Kingdom, ²Paul Strickland Scanner Centre, Mount Vernon Hospital, Northwood, Middlesex, United Kingdom, ³Royal United Hospital Bath, Bath, United Kingdom, ⁴CR-UK-EPSRC Cancer Imaging Centre, Institute of Cancer Research & Royal Marsden Hospital, Sutton, Surrey, United Kingdom

Quantitative DCE-MRI can predict both response and outcome in breast cancer patients after 2 cycles of neoadjuvant chemotherapy (NAC). Quantitative parameters are time-consuming to calculate, requiring expensive software and interpretive expertise. This study shows that the easier to use, signal intensity-time curve (SITC) shapes were significantly associated with K^trans values at baseline and after two cycles of NAC (both p = 0.000). Changes in curve type and K^trans were significantly associated (² = 53.5, p = 0.000). Reductions of >1 in SITC shape predicts improved overall 5 year survival (81% vs 69% (p = 0.048)).

Scott James Peltier¹, Marc Berman², Mary Kathleen Askren³, Bratislav Misic⁴, Mi Sook Jung¹, Anthony Randal McIntosh⁴, Lynn Ossher¹, Min Zhang¹, Patricia Reuter-Lorenz¹, and Bernadine Cimprich^1
1University of Michigan, Ann Arbor, MI, United States, ²University of South Carolina, SC, United States, ³University of Washington, WA, United States,⁴University of Toronto, Ontario, Canada

This study investigates resting-state network correlations in women treated for breast cancer and age-matched healthy controls. In addition, a partial-least squares analysis was performed. Univariate and multivariate analyses both exhibited differential patterns of connectivity across groups and time. This may form the basis for improved diagnostic and monitoring techniques for cognitive changes associated with breast cancer and its treatment.

Samata Kakkad^1,2, Jiangyang Zhang¹, Alireza Akhbardeh¹, Desmond Jacob¹, Meiyappan Solaiyappan¹, Michael A. Jacobs^1,3, Venu Raman^1,3, Dieter Leibfritz², Kristine Glunde^1,3, and Zaver M. Bhujwalla^1,3
1The Russell H. Morgan Department of Radiology and Radiological Science, The Johns Hopkins University School of Medicine, Baltimore, Maryland, United States, ²Department of Chemistry and Biology, University of Bremen, Bremen, Germany, ³Sidney Kimmel Comprehensive Cancer Center, The Johns Hopkins University School of Medicine, Baltimore, Maryland, United States

We investigated the influence of collagen 1 (Col1) fibers on water diffusion, using a breast cancer xenograft fluorescing under hypoxia. High Col1 fiber density correlated with increased apparent diffusion coefficient (ADC) and fractional anisotropy (FA). Hypoxic regions contained significantly fewer Col1 fibers, and were characterized by lower ADC and FA compared to normoxic tumor regions. Diffusion patterns observed in vivo were spatially similar to those observed ex vivo, suggesting that noninvasive DTI can be used to evaluate Col1 fiber density, an important biomarker of tumor progression, and highlighting the importance of Col1 fibers in molecular transport through the extracellular matrix.

Jessica Kwong¹, Farouk Nouizi¹, Yifan Li¹, Gultekin Gulsen¹, and Min-Ying Lydia Su^1
1Tu & Yuen Center for Functional Onco-Imaging, Department of Radiological Sciences, University of California, Irvine, CA, United States

Diffuse Optical Imaging can potentially be used to measure quantitative breast density. We used MRI-based density segmented from healthy breasts to reconstruct a 3D model of the breast and the dense tissue for optical imaging simulation. Near infrared lights of various patterns were illuminated to the breast from one side and received from the other side. The light propagation through the 3D tissue model was simulated and reconstructed to generate the 3D absorption maps. The percentage of the high-absorbing dense tissue from the optical reconstructions is highly correlated with the true percent density measured by MRI, with r= 0.9866.

Sheye Aliu¹, Kathryn E Keenan², Lisa Wilmes¹, David Newitt¹, Ella Jones¹, and Nola Hylton^1
1University of California San Francisco, San Francisco, CA, United States, ²National Institute of Standards and Technology, Boulder, United States

Using a novel breast phantom, we evaluated variability in T1 relaxation and ADC measurements across four scanner/coil configurations.

Wybe JM van der Kemp¹, Bertine L Stehouwer¹, Jurgen H Runge², Jannie P Wijnen¹, Aart J Nederveen², Peter R Luijten¹, and Dennis WJ Klomp^1
1Radiology, UMC Utrecht, Utrecht, Netherlands, ²Radiology, AMC, Amsterdam, Netherlands

The ratio between phosphocholine (PC), phosphoethanolamine (PE) and their glycerol compounds (GPE and GPC) are highly valuable biomarkers. However, when obtained in vivo, these signals can overlap with signals from membrane phospholipids (i.e. glycerophosphatidyl-choline (GPtC) and –ethanolamine (GPtE)). Using adiabatic multi-echo (AMESING) and polarization transfer (BINEPT) techniques at 7T we demonstrate that based on chemical shift, absence of polarization transfer and reduced T2, the majority of phosphodiester signals obtained in breast tissue originate from GPtC and GPtE rather than from the GPE and GPC.

He Zhu¹, Lori Arlinghaus¹, Jennifer G. Whisenant¹, John C. Gore¹, and Thomas Yankeelov^1
1VUIIS, Vanderbilt University, Nashville, TN, United States

Diffusion weighted imaging (DWI) provides quantitative and non-invasive assessments of cell density in breast tissue. We recently developed a water selective DWI acquisition obviating the need for fat suppression using inversion recovery or pre-saturation. Instead, our method relies on image-based shimming to identify the water resonance during a pre-scan and then applies frequency selective excitation on the water resonance. In this abstract, we report a test-retest study to investigate if the added complexity of this water selective DWI method compromised reproducibility.

Martin D Pickles¹, Peter Gibbs¹, Anne Hubbard², Ayesha Rahman², Joanna Wieczorek², Ronjabati Roychaudhury², and Lindsay W Turnbull^1
1Centre for Magnetic Resonance Investigations, HYMS at University of Hull, Hull, East Yorkshire, United Kingdom, ²Breast Care Unit, Hull & East Yorkshire Hospitals NHS Trust, Hull, East Yorkshire, United Kingdom

The aim of this study was to evaluate the feasibility of registered breast MRI and ultrasound data in the planning of breast conserving surgery via guide wire insertion. Following MRI examination participants underwent US/MRI registration proceeding to guide wire localization. To aid co-registration 3-4 common points were identifiable in both the US and MR images. Root mean square deviation values and a qualitative assessment of global registration were recorded. All US/MRI registrations were successful. These results demonstrate that the registration of MR and US data to aid in the insertion of surgical guide wires is feasible.

Yonggang Lu¹, Jacobus F.A. Jansen², Gaorav Gupta¹, Nancy Lee¹, Hilda E. Stambuk¹, Yousef Mazaheri¹, Joseph O. Deasy¹, and Amita Shukla-Dave^1
1Memorial Sloan-Kettering Cancer Center, NEW YORK, New York, United States, ²Maastricht University Medical Center, Maastricht, Netherlands

Reliable prediction for treatment response will help to make optimized treatment planning more effectively in head and neck cancers. In the present study we evaluated the merits of texture analysis on parametric maps derived from pharmacokinetic modeling of dynamic contrast enhanced magnetic resonance imaging in the usage of treatment response prediction. The results demonstrated that the energy (E) of parametric maps of ve (volume fraction of the extravascular extracellular space) was significantly higher during treatment compared with pretreatment, suggesting chemo-radiation treatment significantly reduces the heterogeneity of tumors. Future studies with larger patient populations are required to validate this finding.

Aneela Afzal¹, Xin Li¹, Mohan Jayatilake¹, Yiyi Chen¹, Zunqiu Chen¹, William Woodward¹, William Fleming¹, and Wei Huang^1
1Oregon Health & Science University, Portland, Oregon, United States

Pre-chemotherapy DCE-MRI exams were performed on 13 acute myelogenous leukemia (AML) patients. The mean bone marrow kep parameter from vertebral body (L2 - L4) and iliac crest provided excellect prediction of complete remission status after the chemotherapy. Bone marrow DCE-MRI may be a useful noninvasive imaging tool in personalized care of leukemia patient.

Junyu Guo¹ and Wilburn E Reddick^1
1St Jude Children's Research Hospital, Memphis, TN, United States

We present a quantile-quantile (QQ) plot quantification method to measure the changes between the baseline and the following DCE-MRI examinations on a single phase II trial for pediatric osteosarcoma (OS) patients. The QQ quantification analysis could generate five QQ parameters to describe the different characteristics of the change of histograms in tumors between two serial DCE-MRI examinations. We found that the QQ quantification could potentially provide very early biomarkers for histologic response and event free survival analysis, but the mean value of DCE-MRI parameters in the tumors didn’t provide such early biomarkers in a statistical analysis of 31 OS patients.

Hyunki Kim¹, Desiree Morgan¹, David Sarver², Kyle Lee¹, T. Beasley¹, and Kimberly Keene^1
1University of Alabama at Birmingham, Birmingham, AL, United States, ²University of Arkansas for Medical Sciences, AR, United States

DCE-MRI/DWI was successfully applied for patients with HCCs to quantitate the perfusion and diffusion parameters of HCCs. Significant decreases of Ktrans and kep values were observed after sequential combination therapy with radiation and sorafenib, while tumor ADC values were significantly increased. Tumor Ktrans change was significantly correlated with tumor-volume change, and therefore it may serve as an effective surrogate biomarker to assess the therapeutic efficacy of radiation therapy alone or in combination with sorafenib.

Carolin Reischauer^1,2, Johannes M. Froehlich¹, Dow-Mu Koh³, René Patzwahl⁴, Christoph A. Binkert⁴, Sebastian Kos¹, and Andreas Gutzeit^1,5
1Institute of Radiology and Nuclear Medicine, Clinical Research Unit, Hirslanden Klinik St. Anna, Lucerne, Switzerland, ²Institute for Biomedical Engineering, ETH and University Zurich, Zurich, Switzerland, ³Department of Radiology, Royal Marsden Hospital, Sutton, United Kingdom, ⁴Department of Radiology, Cantonal Hospital Winterthur, Winterthur, Switzerland, ⁵Department of Radiology, Paracelsus Medical University, Salzburg, Austria

Patients suffering from prostate cancer-related bone metastases initially respond well to antihormonal treatment. However, after a period of 2-3 years resistance is usually observed. The present work investigates whether diffusion-weighted imaging permits monitoring this process. It has been previously shown that increased apparent diffusion coefficients (ADCs) are observed at 1 month after commencement of therapy in responders to androgen deprivation. Using ADCs and functional diffusion maps, the present work shows that this initial increase is followed by decreasing ADCs with onset of antihormonal resistance.

Dania Daye¹, Anil Chauhan¹, Sarah Englander¹, Thomas Ferrara¹, Colleen Redlinger², Naomi Haas², Hee-Kwon Song¹, Stephen Keefe², and Mark Rosen^1
1Radiology, University of Pennsylvania, Philadelphia, PA, United States, ²Medicine, University of Pennsylvania, Philadelphia, PA, United States

We compared the sensitivity of DCE-MRI and multi-B-value DWI to detect changes in renal cell carcinoma tumor physiology following initiation of anti-angiogenic therapy with Sunitinib. Both DCE-MRI and DWI were able to detect therapy-induced changes in tumor physiology. Statistically significant decreases in tumor permeability and plasma volume (DCE), and fast diffusion component and perfusion fraction (DWI) were shown. Positive correlations were identified between DCE and DWI methods for quantifying baseline RCC tumor vascular physiology, and for detecting vascular changes in RCC tumors early after therapy. DWI can be used as a non-contrast method of tumor vascular monitoring during anti-angiogenic therapy.

Carolin Reischauer^1,2, Johannes M. Froehlich¹, Miklos Pless², Christoph A. Binkert², Sebastian Kos¹, and Andreas Gutzeit^1,3
1Institute of Radiology and Nuclear Medicine, Clinical Research Unit, Hirslanden Klinik St. Anna, Lucerne, Switzerland, ²Department of Radiology, Cantonal Hospital Winterthur, Winterthur, Switzerland, ³Department of Radiology, Paracelsus Medical University, Salzburg, Austria

Therapy response in patients suffering from non-small lung cancer is usually measured by tumor shrinkage assessed using computed tomography after two cycles of chemotherapy. If the treatment turns out to be ineffective, patients undergo toxic therapy for weeks without benefit. The present work investigates whether mean apparent diffusion coefficients (ADC) and functional diffusion maps (fDMs) permit predicting response at an earlier stage. Thereby, fDMs allow evaluating heterogeneous treatment effects by quantifying the fractions of the tumor volume that show either a significant increase or decrease in ADCs compared with pretreatment values. Statistical analysis reveals that tumor shrinkage can be predicted using fDMs but not mean ADCs.

Seung Hyun Cho¹, Gab Chul Kim¹, Hye Jung Kim¹, Kyung-Min Shin¹, Yun-Jin Jang², Hunkyu Ryeom², and See Hyung Kim^3
1Radiology, Kyungpook National University Medical Center, Daegu, Korea, ²Radiology, Kyungpook National University Hospital, Daegu, Korea, ³Department of Radiology, Dongsan Hospital, College of Medicine, Keimyung University, Daegu, Korea

Locally Advanced Rectal Cancer: Post-chemoradiotherapy Apparent Diffusion Coefficient (ADC) Histogram Analysis for Predicting a Complete Response

Mihaela Rata¹, Nina Tunariu¹, Dow M Koh¹, Stan Kaye², Angela George³, Martin O Leach¹, David J Collins¹, and Matthew D Blackledge^1
1Radiotherapy and Imaging, CR-UK and EPSRC Cancer Imaging Centre, Institute of Cancer Research and Royal Marsden Hospital, Sutton, Surrey, United Kingdom, ²Drug Development Unit, CR-UK and EPSRC Cancer Imaging Centre, Institute of Cancer Research and Royal Marsden Hospital, Sutton, Surrey, United Kingdom, ³Gynaecology Unit, Royal Marsden Hospital, Sutton, Surrey, United Kingdom

Intra- or inter- tumour heterogeneity is a major challenge for anticancer treatment. Our work proposes a novel whole-body DWI analysis methodology to measure therapy response in disseminated disease, as exemplified on a metastatic ovarian cancer patient. The longitudinal tumour behaviour (over 10 visits) was continuously assessed during therapy at multiple sites of metastasis. Such analysis allows full characterization and visualisation of disease burden and observation/quantification of individual tumour response to therapy at each individual site. Histogram and volumetric assessment of individual lesions provides additional information regarding treatment compared to overall assessments of total tumour burden and global ADC distributions.

Eimear Ann Joyce¹, Sylvia A O'Keeffe¹, Andrew Fagan^1,2, Jason McMorrow¹, Danielle Byrne¹, John Kennedy³, and James F Meaney^1,2
1Centre for Advanced Medical Imaging, St. James's Hospital, Dublin 8, Dublin, Ireland, ²Trinity College Dublin, Dublin 2, Ireland, ³HOPE Directorate, St. James' Hospital, Dublin, Ireland

The purpose of this study was to evaluate the performance of tumor apparent diffusion coefficient (ADC) values in predicting response early in the course of chemotherapy, for patients with breast cancer. We also assessed the effect of tumor marker clip placement on the ADC value. Our results confirm that changes in ADC values early in the course of treatment can predict treatment response. The gel-containing clip, used for tumor marking prior to neoadjuvant chemotherapy, results in a potential source of error when calculating ADC tumor values and should be avoided when drawing tumor regions of interest.

Mohan Jayatilake¹, Xubo Song¹, Alina Tudorica¹, Yiyi Chen¹, Karen Oh¹, Nicole Roy¹, Mark Kettler¹, and Wei Huang^1
1Oregon Health & Science University, Portland, Oregon, United States

Pre-biopsy DCE-MRI data from 82 mammography- and/or ultrasound-detected suspicious lesions were collected and analyzed using the Standard and Shutter-Speed PK models. Mean, higher order moments, and entropy of the pixel PK parameters were calculated. The higher order moments and entropy of the Ktrans and DeltaKtrans parameters provided comparable diagnostic accuracy as the mean metric, suggesting tumor perfusion heterogeneity can be used to discriminate benign and malignant breast lesions. The utility of spatial heterogeneity of the PK parameters will be investigated in future studies.

Hassan Bagher-Ebadian^1,2, James R Ewing^2,3, Siamak P. Nejad-Davarani^4,5, Hamed Moradi⁶, Reza Faghihi⁶, Rajan Jain⁷, Tom Mikkelsen⁸, Lisa Scarpace⁸, and Hamid Soltanian-Zadeh^1,9
1Radiology, Henry Ford Hospital, Detroit, MI, United States, ²Physics, Oakland University, Rochester, MI, United States, ³Neurology, Henry Ford Hospital, MI, United States, ⁴Neurology, Henry Ford Hospital, Detroit, MI, United States, ⁵Biomedical Engineering, University of Michigan, Ann Arbor, MI, United States, ⁶Mechanical Engineering, Shiraz University, Fars, Iran, ⁷Radiology, NYU Langone Medical Center, NY, United States, ⁸Neurosurgery, Henry Ford Hospital, Detroit, MI, United States, ⁹CIPCE, ECE Dept., University of Tehran, Tehran, Iran

In this study, Logan plot analysis was applied to dynamic-contrast-enhanced MRI data of 15 patients with Glioblastoma-Multiforme to estimate the tumor distribution volume (VD). BDS (W.A.Brock, W.Dechert and J.Scheinkman) statistic was used to identify the equilibrium condition of the Logan curve. Nested-Model-Selection (NMS) technique was also applied to the same dataset. Results confirm that the VD values estimated by the two techniques are quite in agreement (0.946,p<0.001) while there is considerable variation between subjects in both methods (VD:5% to 46% in Logan-plot with mean and STD of VD=0.23%±0.13% and 7% to 53% in NMS with mean and STD of VD=0.27%±0.14%).

Rafal Panek¹, Maria A Schmidt¹, Marco Borri¹, Dow-mu Koh¹, Angela Riddell², Liam Welsh¹, Ceri Powell², Shreerang A. Bhide¹, Christopher M. Nutting³, Kevin J. Harrington⁴, Kate L. Newbold¹, and Martin O. Leach^1
1Royal Marsden NHS FT and Institute of Cancer Research, Sutton, United Kingdom, ²Royal Marsden NHS FT, Sutton, United Kingdom, ³Royal Marsden NHS FT, London, United Kingdom, ⁴Royal Marsden NHS FT and Institute of Cancer Research, London, United Kingdom

The use of the TWIST view-sharing sequence for DCE pharmacokinetic parameter calculation was investigated. High temporal resolution DCE data obtained without view-sharing was used to simulate the effects of different TWIST k-space undersampling patterns. Absolute percentage differences of DCE parameters were calculated for 15 different combinations of sampled central and peripheral parts of k-space. Calculations were carried out for a group of H&N cancer patients (n=8) with varying primary and nodal disease sites to maximize vascular parameter heterogeneity. Optimal parameters allowing for reliable DCE calculations using TWIST were found allowing for high temporal and spatial resolution measurements in H&N cancer.

Huyen T Nguyen¹, Guang Jia², Kamal S Pohar³, Amir Mortazavi⁴, Zarine K Shah⁵, Debra Zynger⁶, Lai Wei⁷, Xiangyu Yang¹, Daniel Clark¹, and Michael V Knopp^1
1Wright Center of Innovation in Biomedical Imaging, Department of Radiology, The Ohio State University, Columbus, Ohio, United States, ²Department of Physics and Astronomy, Louisiana State University, Baton Rouge, Louisiana, United States, ³Deparment of Urology, The Ohio State University, Columbus, Ohio, United States, ⁴Deparment of Internal Medicine, The Ohio State University, Columbus, Ohio, United States, ⁵Deparment of Radiology, The Ohio State University, Columbus, Ohio, United States, ⁶Deparment of Pathology, The Ohio State University, Columbus, Ohio, United States, ⁷Center for Biostatistics, The Ohio State University, Columbus, Ohio, United States

This study is to evaluate the value of k-means clustering of DCE-MRI pharmacokinetic parameters in T staging of bladder tumors. k-means clustering was performed on the non-dimensionalized Amp and kep values of all twenty-four patients in the study to determine three cluster centers. The volume fractions (VFs) of three clusters were correlated with the tumor stage. Significant difference in the VF of cluster 2 was found between T1/lower vs. T2, T1/lower vs. T3, and T3 vs. T4. The differences in all three cluster VFs were also statistically significant. Fat-invasive tumors had significantly higher VFs of cluster 1 and 3 and a significantly lower VF of cluster 2 than did non-fat-invasive tumors. The VF of cluster 2 had area-under-the-curve (AUC) value of 0.83 in the differentiation of fat-invasive from non-fat-invasive tumors. k-means clustering of DCE-MRI pharmacokinetic parameters can be a useful tool for the quantitative assessment of T stages to improve the accuracy of the T staging of bladder cancer.

Hassan Bagher-Ebadian^1,2, Siamak P. Nejad-Davarani^3,4, James R Ewing^2,3, Tom Mikkelsen⁵, Rajan Jain⁶, Lisa Scarpace⁵, and Hamid Soltanian-Zadeh^1,7
1Radiology, Henry Ford Hospital, Detroit, MI, United States, ²Physics, Oakland University, Rochester, MI, United States, ³Neurology, Henry Ford Hospital, Detroit, MI, United States, ⁴Biomedical Engineering, University of Michigan, Ann Arbor, MI, United States, ⁵Neurosurgery, Henry Ford Hospital, Detroit, MI, United States, ⁶Radiology, NYU Langone Medical Center, NY, United States, ⁷CIPCE, ECE Dept., University of Tehran, Tehran, Iran

A nested model selection (NMS) technique along with physiological concepts of the models is introduced and a ‘model-averaging’ technique in Dynamic-Contrast-Enhanced (DCE)-MR model selection using the Akaike-Information-Criterion (AIC) is constructed. The Models in NMS are recruited in the AIC and applied to an exemplary DCE-MR data of a patient with Glioblastoma-Multiforme. Model-choice and probability maps estimated from both techniques are compared. The AIC and NMS provide unique set of probability maps for estimating the contribution of each model in a specific voxel. These probabilities allow combining the estimations from different models, thus generating a more accurate estimate of permeability parameters.

Donald Robinson Cantrell¹, Thomas Anthony Gallagher¹, Bruce Spottiswoode², Timothy Carroll^1,3, Charles Fasanati¹, and Dingxin Wang^4,5
1Department of Radiology, Northwestern University, Chicago, IL, United States, ²Cardiovascular MR R&D, Siemens Healthcare, Chicago, IL, United States,³Department of Biomedical Engineering, Northwestern University, Chicago, IL, United States, ⁴Siemens Medical Solutions USA, Inc., Minneapolis, MN, United States, ⁵CMRR, University of Minnesota, Minneapolis, MN, United States

Perfusion Weighted Imaging (PWI) provides functional information on the hemodynamic status of CNS malignancies. Dynamic Susceptibility Contrast (DSC) enhanced MRI is an implementation of PWI that rapidly acquires images following contrast administration. However, standard clinical DSC-MRI protocols have limited spatial coverage and temporal resolution. Our previous work demonstrated the feasibility of using Slice Accelerated Echo Planar Imaging (SA-EPI) for DSC-MRI measurements in healthy volunteers. We now report the first large-scale clinical implementation of this promising new protocol for the evaluation of CNS malignancies. Faster data acquisition achieved with SA-EPI allows for increased spatial coverage while maintaining the required temporal resolution.

Milica Medved¹, Aytekin Oto¹, Xiaobing Fan¹, Federico D Pineda¹, Gregory S Karczmar¹, and Russell Z Szmulewitz^2
1Radiology, University of Chicago, Chicago, Illinois, United States, ²Medicine, University of Chicago, Chicago, Illinois, United States

Cabozantinib shows promising results in extension of progression-free survival for men with castration-resistant prostate cancer. In this preliminary report, we document the effect of cabozantinib on perfusion in bone metastases in castration-resistant prostate cancer and in muscle tissue using the Tofts model parameters K^trans and v_e. We find a statistically significant decrease in lesion K^trans (37% on average, p < 0.01) during the first two weeks of therapy, while other quantities were not significantly changed. It remains to be seen whether individual changes in lesion K^trans can be correlated with treatment response.

Yonggang Lu¹, Ashok R. Shaha¹, Hilda E. Stambuk¹, Andre Moreira¹, Yousef Mazaheri¹, Joseph O. Deasy¹, R. Michael Tuttle¹, and Amita Shukla-Dave^1
1Memorial Sloan-Kettering Cancer Center, NEW YORK, New York, United States

There is an urgent need to non-invasively assess tumor aggressiveness in patients with papillary thyroid carcinoma (PTC). In this study, pretreatment intravoxel incoherent motion imaging (IVIM) MRI was used to quantify water molecular diffusion and blood perfusion in tumor tissue of 15 patients with PTC. All patients were treated with surgery and histopathological features of aggressiveness were used as the standard of reference. The results show that IVIM MRI derived metrics were able to differentiate between aggressive and non-aggressive tumors. The study concludes that diffusion coefficient is a surrogate biomarker of tumor aggressiveness in patients with PTC.

Karen Mooney¹, Tejan Diwanji¹, Xiutao Shi¹, Warren D D'Souza¹, and Nilesh Mistry^1
1Radiation Oncology, University of Maryland School of Medicine, Baltimore, Maryland, United States

Simulated 1D navigator echoes were retrospectively generated in two directions from 2D MR images of 5 patients with lung tumors. The respiratory motion of the tumors was tracked using the navigators, and the results were compared to manual tracking of the tumors using the full images.

Carri K Glide-Hurst¹, Ning Wen¹, David Hearshen¹, Milan Pantelic¹, Bo Zhao¹, Yanle Hu², Tina Kunkel¹, Kenneth Levin¹, Benjamin Movsas¹, Indrin J. Chetty¹, and M. Salim Siddiqui^1
1Henry Ford Health System, Detroit, MI, United States, ²Washington University, St Louis, MO, United States

Due to its excellent soft tissue contrast, MRI is integrated as an adjunct to computed tomography simulation (CT-SIM) for radiotherapy (RT) treatment planning to assist in tumor and organ at risk delineation. Recently, dedicated MR simulation (MR-SIM) platforms for radiation oncology have been introduced, although paucity in the literature exists on how to fully implement MR-SIM into the clinic. This work describes our experience with characterizing system performance, establishes quality assurance (QA) programs, and sets the context for dedicated MR-SIM for RT. We developed QA procedures and workflow necessary to implement MR-SIM into treatment planning and demonstrate its clinical use.

Basetti Madhu¹, Greg Shaw¹, David Neal¹, and John R Griffiths^1
1Cancer Research UK Cambridge Institute, University of Cambridge, Cambridge, England, United Kingdom

Human prostate biopsies were analysed with HRMAS ¹H NMR spectroscopy. Absolute metabolite concentrations of the following metabolites were estimated by a modified LC-Model basis set in benign, low grade and high grade prostate cancer samples: alanine, lactate, glutamine, glutamate, citrate, choline, phosphocholine (PC)+ glycerophosphocholine(GPC), creatine + phosphocreatine, taurine, myo-inositol and spermine . Alanine was significantly elevated in both low and high-grade prostate cancer biopsies; total choline(choline+PC+GPC) and glutamate were significantly higher in high-grade malignant prostate. None of the macromolecule (0.9ppm, 2.0ppm) and lipid signals (0.9ppm, 1.3ppm and 2.0ppm) showed any statistical differences between the benign and cancer samples

Navneeta Bansal¹, Ashish Gupta², Anil Mandhani³, Abbas Ali Mahdi⁴, and Satya Narain Sankhwar^1
1Department of Urology, King George's Medical University, Lucknow, UP, India, ²Department of Metabolomics, Centre of Biomedical Research, Lucknow, UP, India, ³Department of Urology, SGPGIMS, Lucknow, UP, India, ⁴Department of Biochemistry, King George’s Medical University, Lucknow, UP, India

A novel approach of 1H NMR spectroscopy of serum metabolite profile and multivariate statistical approach—orthogonal partial least-squares discriminant analysis (PLS-DA) was carried out to identify differential biomarkers of urinary bladder cancer (UBC) comprising, low grade (LG) and high grade (HG). The study was carried out on 67 UBC patients and 32 healthy volunteers to differentiate among healthy control (HC), LG and HG. PLS-DA-derived serum metabolomics were able to precisely discriminate 95% of cases of BC with 96% sensitivity and 94% specificity when compared to HC and 98% of cases of LG from HG with 97% sensitivity and 99% specificity.

Nathalie Just^1
1CIBM-AIT, EPFL, Lausanne, Switzerland

DCE-MRI and DWI have been used to estimate various physiological parameters on a voxel by voxel basis allowing a full visualisation of the tumour heterogeneity. The methodologies for calculating and interpreting voxel by voxel values are poorly known and may be time-consuming. Unfortunately, mean and median quantitative values are not always significantly sensitive to small changes and may not represent the precise status of the tumour owing to their intrinsic chaotic environment. Owing to advances in both high-resolution MRI and signal processing methods, histogram analyses of tumours showed their usefulness for investigating the distributions of various tumour parameters.

Akira Nishikori^1,2, Masaaki Hori^1,3, Fumitaka Kumagai^1,2, Yoshitaka Masutani⁴, Ryuji Nojiri³, Katsutoshi Murata⁵, Kohei Kamiya^1,4, Koji Kamagata¹, Mariko Yoshida¹, Michimasa Suzuki¹, and Shigeki Aoki^1
1Radiology, Juntendo University School of Medicine, Tokyo, Japan, ²Graduate School of Human Health Sciences, Tokyo, Japan, ³Tokyo Medical Clinic, Tokyo, Japan, ⁴The University of Tokyo, Tokyo, Japan, ⁵Siemens Japan K.K., Tokyo, Japan

The purpose of this exhibit is to explain methods of histogram analyses of diffusion metrics and to demonstrate clinical usefulness of the analysis in evaluation of brain tumors. Histogram analyses of diffusion metrics showed promise for classification of the grades and subtypes of brain tumors, demonstration of intratumoral microstructures and determination of peritumoral invasion In case it is difficult for radiologists to reach comprehensive diagnosis with conventional imaging technique, complementary combination use of histogram analysis of diffusion metrics will be helpful for more precise diagnosis of brain tumors.

Andrew B Rosenkrantz¹, Anne-Kristin Vahle¹, Christian Geppert², Christopher Glielmi², Kent P Friedman¹, Rachel M Bartlett¹, Samir S Taneja³, Yu-Shin Ding¹, and Thomas Koesters^1
1Center for Advanced Imaging Innovation and Research, Department of Radiology, NYU Langone Medical Center, New York, NY, United States, ²Siemens Healthcare, Erlangen, Germany, ³Urologic Oncology, NYU Langone Medical Center, New York, New York, United States

By providing truly simultaneous spatial and temporal acquisitions, hybrid PET/MRI using dynamic PET imaging allows robust comparison of the kinetics of MRI and PET tracers. In this study, 12 prostate cancer patients underwent hybrid PET/MRI with simultaneous DCE-MRI and dynamic PET acquisitions following consecutive injections of gadolinium-chelate and 18F-FDG. Early post-injection PET data was reconstructed using 30-second bins, and ROIs were placed on lesions using fused DCE-MRI/PET images to generate matching time-activity-curves. Versus gadolinium-chelate, FDG exhibited significantly later time-to-peak and greater maximal slope of uptake. Such kinetic differences between the two agents have not been previously reported to our knowledge.

Radka Stoyanova¹, Andres Parra¹, Kyle Padgett¹, Matthew Abramowitz¹, and Alan Pollack^1
1Radiation Oncology, University of Miami, Miami, Florida, United States

Three clinical trials are initiated for targeted radiation treatment of prostate cancer, based on the hypothesis that: (i) the dominant lesions recognized on multiparametric MRI (MP-MRI) determine outcome; (ii) MP-MRI-directed biopsies are critical to accurately assessing pre-treatment (pre-Tx) histopathologic and molecular characteristics; (iii) MP-MRI parameters are related to tumor response and molecular abnormalities; (iv) early MP-MRI changes after treatment will correlate with response and (v) targeting these lesions will improve control rates without increasing toxicity. The workflow for integration of MP-MRI at multiple points in the trails design is discussed, together with preliminary results.

Maarten Leonard Smits¹, Bruce L Daniel², Sharon E. Clarke², Jesse McKenney², Andrew Wentland², Manoj Saranathan², Lewis Shin², Kyung Sung², Brian Hargreaves², Emine U. Saritas², Dwight G. Nishimura², and Graham Sommer^2
1University Medical Center Utrecht, Utrecht, Netherlands, ²Stanford University, California, United States

This study assessed prospectively the ability of multiparametric (mp) MRI to map intraprostatic adenocarcinoma (Pca) of varying Gleason scores and the mapping accuracy of the individual components of mp-MRI. All sequences were much more accurate in imaging higher grades (significant tumor) than lower grade tumor. DWI and ADC maps were most accurate in imaging PCa overall, and DCE imaged higher grade tumor most accurately. T2W gave poorer results than either DWI or DCE. The multiparametric approach resulted in marginally better results than DWI in this study.

Ara Alconchel Pilar¹, Sarah Alessi¹, Michele Colombi², Paul Summers¹, Massimo Bellomi^1,3, Luca Antiga², and Giuseppe Petralia^1
1European Institute of Oncology, Milan, MI, Italy, ²Orobix S.r.l., Bergamo, BG, Italy, ³Univerity of Milan, Milan, MI, Italy

Since its publication in the 2012 ESUR guidelines, PI-RADS (Prostate Imaging Reporting And Data System) scoring of prostate Multiparametric MRI (mp-MRI) is increasingly used in radiological practice. We present our experience with RADcommunicator, a freely-available iPad application that provides standardized reporting with axial and coronal prostate templates and a table to record mp-MRI assessments of each lesion. In our institution, RADcommunicator has facilitated routine use of PI-RADS scoring in >300 patients and enhanced the communication of clinical results derived from prostate mp-MRI to all physicians involved in the care of prostate cancer patients.

Marnix C. Maas¹, Alan J. Wright¹, Kirsten M. Selnæs^2,3, Masoom A. Haider⁴, Katarzyna J. Macura⁵, Daniel J.A. Margolis⁶, Berthold Kiefer⁷, Jurgen J. Fütterer¹, and Tom W.J. Scheenen^1
1Radiology, Radboud University Medical Center, Nijmegen, Netherlands, ²Department of Circulation and Medical Imaging, Norwegian University of Science and Technology, Trondheim, Norway, ³St. Olavs Hospital, Trondheim University Hospital, Trondheim, Norway, ⁴Sunnybrook Health Sciences Centre, University of Toronto, Toronto, Ontario, Canada, ⁵Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins University, Baltimore, MD, United States, ⁶UCLA David Geffen School of Medicine, Los Angeles, CA, United States, ⁷Siemens AG Healthcare Sector, Erlangen, Germany

We present initial results of a multi-center trial aimed at assessing the diagnostic accuracy of 3T multi-parametric MR imaging and spectroscopy in distinguishing clinically significant prostate cancer (PCa) from other prostatic tissue, with whole-mount section histopathology as the gold standard. Good separation between PCa and benign tissues was found for multi-center ADC data; DCE and MRSI need further analysis before solid conclusions about these methods’ performance can be drawn. The validation part of this prospective trial will be used to determine the parameters contributing most to the detection and localization of clinically significant PCa as well as their optimal thresholds.

Nelly Tan¹, Daniel J Margolis¹, David Y Lu², Robert E Reiter³, and Steven S Raman^1
1Radiology, UCLA, Los Angeles, CA, United States, ²Pathology, UCLA, Los Angeles, CA, United States, ³Urology, UCLA, Los Angeles, CA, United States

Our findings have implications for emerging focal therapy of prostate cancers. A large proportion of secondary tumors and a small proportion of index lesions will not be detected by MR imaging. The results underscore the need for close follow up post therapy.

Fusun Citak Er¹, Metin Vural², Omer Acar³, Tarik Esen⁴, Aslihan Onay², and Esin Ozturk-Isik^5
1Genetics and Bioengineering, Yeditepe University, Istanbul, Turkey, ²Department of Radiology, VKF American Hospital, Istanbul, Turkey, ³Department of Urology, VKF American Hospital, Istanbul, Turkey, ⁴School of Medicine, Koc University, Istanbul, Turkey, ⁵Department of Biomedical Engineering, Yeditepe University, Istanbul, Turkey

This study aims to evaluate the performances of linear and quadratic discriminant analysis and linear and non-linear support vector machine (SVM) for estimation of final Gleason score preoperatively for prostate cancer. The digital rectal examination (DRE) findings, age, prostate specific antigen (PSA) level, index lesion size, biopsy Gleason score, ADC, Likert scales of T2, diffusion weighted, and dynamic contrast enhanced (DCE) MRI were used as predictors for estimating the final Gleason score based on the pathologic analysis after prostatectomy. The results of our study indicated that linear SVM and linear discriminant analysis performed well in predicting final Gleason score.

Shonit Punwani¹, Salvatore Benigno², Balaji Ganeshan², Ashley Groves², and Mark Emberton^2
1University College London, London, UK, United Kingdom, ²University College London, United Kingdom

This study explores the utility of an entropy textural analysis metric for detection of significant tumour within the transition zone of the prostate on multi-parametric MRi

Rajakumar Nagarajan¹, Zohaib Iqbal¹, Brian Burns¹, Daniel A Margolis¹, Manoj K Sarma¹, Robert E Reiter², Steven S Raman¹, and M.Albert Thomas^1
1Radiological Sciences, University of California Los Angeles, LOS ANGELES, CA, United States, ²Urology, University of California Los Angeles, LOS ANGELES, CA, United States

Detection of more metabolites in prostate cancer is demonstrated using non-uniformly undersampled (NUS) echo-planar J-resolved spectroscopic imaging (EP-JRESI) data processed using maximum entropy (MaxEnt) non-linear reconstruction method than conventional one-dimensional (1D) MR Spectroscopic imaging (MRSI). We have quantified metabolites changes observed in spermine, myo-inositol and citrate in cancer locations. Based on the ADC values, the diffusion weighted imaging (DWI) was able to differentiate the cancer and non-cancer locations. There were positive and negative correlations between the ADC values and metabolite ratios in cancer locations investigated in our study.

Isabel Dregely¹, Daniel JA Margolis¹, Kyung H Sung¹, and Holden H Wu^1
1University of California Los Angeles, Los Angeles, California, United States

The purpose of this work was to apply diffusion weighted 3D DESS MRI to prostate imaging with the goal to achieve simultaneously quantitative T2- and ADC-mapping. Diffusion weighted DESS in the prostate is challenging due to motion, which limits SNR especially on the diffusion weighted echo-signal. Therefore only moderate diffusion gradient moments could be applied. Preliminary results in a healthy volunteer showed that quantitative T2 and ADC-values were within expected range for healthy prostate. Monte Carlo simulations showed good accuracy and precision for T2, however less for ADC-mapping. Simulated scenarios showed that improving SNR and/or diffusion sensitivity will improve quantification.

Feiyu Li¹, Wenchao Cai¹, Jintang Ye¹, Queenie Chan², Xiaoying Wang¹, and Xuexiang Jiang^1
1Department of Radiology, Peking University First Hospital, Beijing, Beijing, China, ²Philips Healthcare, Hong Kong, China

Accurate identification of the recurrent tumor in prostate allows better selection of patients for salvage or adjuvant RT after endocrine therapy. The purpose of this study was to assess the diffusion and perfusion coefficients changes of prostate cancer (PCa) and noncancerous areas after endocrine therapy using DWI biexponential model. The mean D and f values of cancerous and noncancerous foci in the endocrine treatment group both showed statistically different from that in the non-treatment group. The changes of the D and f were accordance with the pathological and physiological degeneration of cells and vessels after endocrine therapy in the histopathologic study. IVIM maybe a promising technique in the better detection of the prostate recurrent lesion after therapy.

Veronica A Morgan¹, Chris C Parker², Sharon L Giles¹, and Nandita M de Souza^3
1Clinical Magnetic Resonance Unit, Royal Marsden NHS Foundation Trust, Sutton, Surrey, United Kingdom, ²Academic Urology, Royal Marsden NHS Foundation Trust, Sutton, Surrey, United Kingdom, ³Clinical Magnetic Resonance Unit, Institute of Cancer Research, Sutton, Surrey, United Kingdom

This study evaluates the relationship between ADC and tumor growth in low risk prostate cancer patients. Tumor volume was calculated on baseline and follow-up scans of 23 patients. From the baseline scan a representative tumor ADC was also derived from a ROI drawn around the largest area of tumor. There was a significant negative correlation between tumor growth rate and ADC at presentation. In this low risk group doubling time is around 3years. Cancers with lower ADC at the outset tended to grow more quickly. Determining threshold ADCs from a larger study could influence decisions around timing of treatment.

Yousef Mazaheri¹, Andreas M Hoetker², Yonggang Lu³, Amita Shukla-Dave³, Oguz Akin⁴, and Hedvig Hricak^4
1Medical Physics, Memorial Sloan-Kettering Cancer Center, New York, NY, United States, ²Radiology, Memorial Sloan-Kettering Cancer Center, New York, NY, United States, ³Medical Physics, Memorial Sloan-Kettering Cancer Center, New York, United States, ⁴Radiology, Memorial Sloan-Kettering Cancer Center, New York, United States

The aim of this work was to compare the non linear least squares (NLLS) algorithm to the maximum likelihood (ML) algorithm for fitting diffusion decay curves to extract parameters to the measured MR signal intensities as a function of b-value for bi-exponential, stretched exponential, and non-Gaussian (Kurtosis) models.

Miriam W. Lagemaat¹, Marnix C. Maas¹, Adam B. Kerr², Andreas K. Bitz^3,4, Stephan Orzada⁴, Mark J. van Uden¹, Eline K. Vos¹, and Tom W.J. Scheenen^1,4
1Radiology, Radboud university medical center, Nijmegen, Netherlands, ²MRSRL, Electrical Engineering, Stanford University, Stanford, CA, United States,³Medical Physics in Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany, ⁴Erwin L. Hahn Institute for Magnetic Resonance Imaging, University Duisburg-Essen, Essen, Germany

A ¹H-MRSI sequence with spectral-spatial refocusing pulses was evaluated for prostate applications at 7T. The combination of an external Tx array coil and an Rx endorectal coil allowed the use of these non-adiabatic pulses, while retaining sensitivity within the prostate. Well-resolved spectra in the prostate of patients were obtained, with some variation due to B₀ and B₁ inhomogeneities.

Kirsten Margrete Selnæs^1,2, Riyas Vettukattil¹, May-Britt Tessem^1,2, Helena Bertillson^3,4, Alan Wright⁵, Arend Heerschap^1,5, Anders Angelsen^3,4, and Tone Frost Bathen^1
1Department of Circulation and Medical Imaging, Norwegian University of Science and Technology, Trondheim, Norway, ²St. Olavs Hospital, Trondheim University Hospital, Trondheim, Norway, ³Department of Urology, St. Olavs Hospital, Trondheim University Hospital, Trondheim, Norway, ⁴Department of Cancer Research and Molecular Medicine, Norwegian University of Science and Technology, Trondheim, Norway, ⁵Department of Radiology, Radboud University Nijmegen Medical Center, Nijmegen, Netherlands

Metabolic and morphologic changes due to prostate cancer lead to changes in MR imaging and MR spectroscopic parameters. This study aims to assess the relationship between MRI parameters (T2 intensity and ADC) measured on patients in vivo, individual metabolites measured on prostatectomy tissue ex vivo with HR-MAS MRS and quantitative histopathological features (percentage nuclei and luminal space). ADC was positively correlated to lumen and negatively correlated to amount of nuclei (ñ = 0.54 and -0.36 respectively, p<0.01). There is a positive correlation between total choline and amount of nuclei (ñ=0.38, p<0.01) and between citrate and amount of lumen (ñ=0.37, p<0.01)

Renuka Sriram¹, Kayvan R Keshari², Robert Bok¹, Subramaniam Sukumar¹, Mark Van Criekinge¹, Daniel B Vigneron¹, and John Kurhanewicz^1
1University of California, San Francisco, San Francisco, California, United States, ²Memorial Sloan-Kettering Cancer Center, New York, United States

This study demonstrates that hypoxia is significantly increased in TRAMP tumors, and is responsible for driving metabolic and micro-environmental changes that favor disease progression. Moreover, these metabolic and micro-environmental changes can be imaged using a multi-probe hyper polarized ¹³C MRI approach. The increase in the hyperpolarized ¹³C signals of Lac/Pyr ratio as well as urea in late stage compared to early stage tumors is synchronized with the increased expression of hypoxic gene regulators. These pre-clinical findings recapitulate the human situation where increasing levels of hypoxia have been measured with increasing clinical stage, and correlated with poor clinical outcomes.

Nassim Tayari¹, Arend Heerschap¹, and Alan J. Wright^1
1Dept. of Radiology, RadboudUMC, Nijmegen, Gelderland, Netherlands

Proton MR spectroscopic imaging is commonly performed with water signal suppression to avoid artifacts from side bands of the large water signal, but this signal can be very useful for estimating metabolite concentrations by providing a line-shape reference and a quantification of the water signal. In this study we developed water unsuppressed 3D MRSI of the prostate and used it to provide choline concentration maps for localization of prostate cancer. We observed that high choline concentrations co-localised to tumour foci as confirmed on histopathological staining of radical prostatectomy specimens.

Gregory J. Metzger¹, Benjamin Fossen¹, Patrick J. Bolan¹, Chrisopher Warlick², Badrinath Konety², Stephen C. Schmechel³, Chaitanya Kalavagunta¹, and Ivan Tkac^1
1Center for Magnetic Resonance Research, University of Minnesota, Minneapolis, MN, United States, ²Urologic Surgery, University of Minnesota, Minneapolis, MN, United States, ³Laboratory Medicine, University of Washington, WA, United States

Model-based fitting of prostate spectra can produce more selective metabolite ratios than conventional peak integration. The model-based ratio tCho/Cit was found to have a strong correlation with cancer aggressiveness (grade), comparable with the standard approach of CSC/Cit ratio, but with the advantage of greater potential selectivity at higher fields.

Andrew B Rosenkrantz¹, Christian Geppert², Robert Grimm², Tobias K Block¹, Christopher Glielmi², Li Feng¹, Ricardo Otazo¹, Justin M Ream¹, Melanie Moccaldi Romolo¹, Samir S Taneja³, Daniel K Sodickson¹, and Hersh Chandarana^1
1Center for Advanced Imaging Innovation and Research, Department of Radiology, NYU Langone Medical Center, New York, NY, United States, ²Siemens Healthcare, Erlangen, Germany, ³Urologic Oncology, NYU Langone Medical Center, New York, New York, United States

 

Andrew David Nicholson^1,2, Viraj A Master^3,4, Tracy E Powell^1,2, Jian Kang^1,5, Adeboye O Osunkoya^3,6, Martin G Sanda^3,4, and Sherif G Nour^1,2
1Department of Radiology and Imaging Science, Emory University Hospital, Atlanta, GA, United States, ²Interventional MRI Program, Emory University Hospital, Atlanta, GA, United States, ³Department of Urology, Emory University Hospital, Atlanta, GA, United States, ⁴School of Medicine, Emory University, Atlanta, GA, United States, ⁵Department of Biostatistics and Bioinformatics, Emory University, Atlanta, GA, United States, ⁶Department of Pathology and Laboratory Medicine, Emory University Hospital, Atlanta, GA, United States

Our presentation details a retrospective analysis, correlating multiparametric prostate MRI findings to pathologic results of MRI-guided core needle biopsy. We present a simple but statistically sound method for scoring lesions, that can be used to select lesions for targeted biopsy and therapy.

Subashini Srinivasan^1,2, Holden H Wu^1,2, Kyunghyun Sung^1,2, Daniel JA Margolis¹, and Daniel B Ennis^1,2
1Department of Radiological Sciences, University of California, Los Angeles, California, United States, ²Department of Bioengineering, University of California, Los Angeles, California, United States

3D T2-weighted imaging is used clinically for high resolution imaging of small tumors. Current clinical standards for 3D T2-weighted imaging are limited by long acquisition durations. We propose a 3D T2-weighted variable flip angle transition into driven equilibrium balanced SSFP (3D T2-TIDE) technique for fast T2-weighted imaging at high field strengths. Images were acquired using a 3D Cartesian trajectory with interleaved ky-kz spiral sampling such that the center of k-space was acquired with increased T2-weighting. 3D T2-TIDE prostate images from five healthy subjects reduced the acquisition duration by 59% while improving the SNR efficiency compared to 3D FSE.

Daniel Margolis¹, Edward Chang², Frederick Dorey³, Jesse Le², Patricia Lieu², and Leonard Marks^2
1Department of Radiological Sciences, UCLA, Los Angeles, CA, United States, ²Urology, UCLA, Los Angeles, CA, United States, ³Biostatistics, UCLA, Los Angeles, CA, United States

MRI-ultrasound fusion targeted prostate biopsy has shown improved yield for significant cancer in men with prior negative biopsies and undergoing active surveillance. It has also been shown to improve yield for the initial biopsy session. We present a comparison of systematic versus targeted biopsy yield in men with elevated PSA but no prior biopsies using 4 definitions of significant disease. The rates of detection were similar, although both systematic and targeted biopsies miss significant disease for all definitions. However, this detection rate occurred with nearly twice as many systematic as targeted biopsy cores.

Ahmed El-Shater Bosaily¹, Massimo Valerio¹, Yipeng Hu², Alex Freeman³, Charles Jameson³, Louise Brown⁴, Richard Kaplan⁴, Mark Emberton¹, Chris Parker⁵, Richard Hindley⁶, and Hashim Ahmed^1
1Division of Surgery and interventional science, University College London, London, United Kingdom, ²Centre for Medical Image Computing, University college London, London, United Kingdom, ³Pathology, University College London Hospitals NHS Foundation Trust, London, United Kingdom, ⁴Clinical Trials Unit, Medical Research Council, London, United Kingdom, ⁵oncology, Royal Marsden Hospital/Institute of Cancer Research., London, United Kingdom,⁶Urology, Hampshire Hospitals NHS Foundation Trust, Hampshire, United Kingdom

Currently, MRI guidance and planning is part of a rapidly growing trend of targeted prostate biopsies. Despite the assumption that the maximum cancer grade (grade hotspot) lies within the maximum dimension of the lesion (volume hotspot), some argue that it might not always be true and that areas of higher cancer grade may show different signal characteristics and may be identifiable on MRI.The aim of this study is to assess the concordance between the grade hotspot and the volume hotspot using TPM biopsies outputs from the pilot phase of the multicenter MRC PROMIS study

Xiaoxi Chen¹, Lianming Wu¹, Yongming Dai², and Jianrong Xu^1
1Department of Radiology, Renji Hospital, Shanghai Jiao Tong University, School of Medicine, Shanghai, Shanghai, China, ²Philips(China) Investment Co.Ltd, Shanghai, China

Feasibility and Preliminary Experience of Quantitative T2 star mapping at 3.0 T for Detection and Assessment of Aggressiveness of Prostate Cancer