Search

Background Diffusion-weighted MRI (DWI) is currently one of the fastest developing MRI-based techniques in oncology. Histogram properties from model fitting of DWI are useful features for differentiation of lesions, and classification can potentially be improved by machine learning. Purpose To evaluate classification of malignant and benign tumors and breast cancer subtypes using support vector machine (SVM). Study Type Prospective. SUBJECTS Fifty-one patients with benign (n = 23) and malignant (n = 28) breast tumors (26 ER+, whereof six were HER2+). Field Strength/Sequence Patients were imaged with DW-MRI (3T) using twice refocused spin-echo echo-planar imaging with echo time / repetition time (TR/TE) = 9000/86 msec, 90 × 90 matrix size, 2 × 2 mm in-plane resolution, 2.5 mm slice thickness, and 13 b-values. Assessment Apparent diffusion coefficient (ADC), relative enhanced diffusivity (RED), and the intravoxel incoherent motion (IVIM) parameters diffusivity (D), pseudo-diffusivity (D*), and perfusion fraction (f) were calculated. The histogram properties (median, mean, standard deviation, skewness, kurtosis) were used as features in SVM (10-fold cross-validation) for differentiation of lesions and subtyping. Statistical Tests Accuracies of the SVM classifications were calculated to find the combination of features with highest prediction accuracy. Mann–Whitney tests were performed for univariate comparisons. Results For benign versus malignant tumors, univariate analysis found 11 histogram properties to be significant differentiators. Using SVM, the highest accuracy (0.96) was achieved from a single feature (mean of RED), or from three feature combinations of IVIM or ADC. Combining features from all models gave perfect classification. No single feature predicted HER2 status of ER + tumors (univariate or SVM), although high accuracy (0.90) was achieved with SVM combining several features. Importantly, these features had to include higher-order statistics (kurtosis and skewness), indicating the importance to account for heterogeneity. Data Conclusion Our findings suggest that SVM, using features from a combination of diffusion models, improves prediction accuracy for differentiation of benign versus malignant breast tumors, and may further assist in subtyping of breast cancer. Level of Evidence: 3 Technical Efficacy: Stage 3 This is the pre-peer reviewed version of the following article: [Support vector machine for breast cancer classification using diffusion-weighted MRI histogram features: Preliminary study], which has been published in final form at [http://onlinelibrary.wiley.com/doi/10.1002/jmri.25873/abstract]. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Self-Archiving.

Background Steady state susceptibility contrast (SSC)-MRI provides information on vascular morphology but is a rarely used method. Purpose To investigate the utility of the ultrasmall superparamagnetic iron oxide particles (USPIOs) GEH121333 for measuring tumor response to bevacizumab and compare this with gadolinium-based DCE-MRI. Study Type Prospective preclinical animal model study. ANIMAL MODEL Mice bearing subcutaneous TOV-21G human ovarian cancer xenografts treated with bevacizumab (n = 9) or saline (n = 9). Field Strength/Sequence Imaging was performed on a 7T Bruker Biospec. For SSC-MRI with GEH121333 we acquired R1-maps (RARE-sequence with variable TR), R2-maps (multi-spin echo), and math formula-maps (multi-gradient echo). Additionally, R1 and R2 maps were measured on the days after USPIO injection. For DCE-MRI with gadodiamide we acquired 200 T1-weighted images (RARE-sequence). Assessment ΔR1, ΔR2, and math formula maps were computed from SSC-MRI. DCE-MRI was analysed using the extended Tofts model. Statistical Tests Results from pre- and 3 days posttreatment SSC-MRI were compared using paired-sample t-tests. Treatment and control groups were compared using independent sample t-tests. Performance of SSC- and DCE-MRI was compared using multivariate partial least squares discriminant analysis. Results Already one day after treatment and USPIO injection, R1 and R2 values were lower in treated (R1 = 0.49 ± 0.03s−1, R2 = 23.07 ± 1.49s−1) compared with control tumors (R1 = 0.52 ± 0.02s−1, R2 = 24.98 ± 1.01s−1), indicating lower USPIO accumulation. Posttreatment SSC-MRI displayed significantly decreased tumor blood volume (change in ΔR2 = -0.43 ± 0.26s−1, P = 0.001) and vessel density (change in Q = -0.032 ± 0.020s−1/3, P = 0.002). DCE-MRI showed among others lower Ktrans in treated tumors (control = 0.064 ± 0.011min−1, tx = 0.046 ± 0.008min−1, P = 0.002). Multivariate analysis suggests that SSC-MRI was slightly inferior to DCE-MRI in distinguishing treated from control tumors (accuracy = 75%, P = 0.058 versus 80%, P = 0.028), but a combination of both was best (accuracy = 85%; P = 0.003). Data Conclusion SSC-MRI with GEH121333 is sensitive to early (

Purpose To improve early diagnosis of prostate cancer to aid clinical decision-making. Diffusion-weighted magnetic resonance imaging (DW-MRI) is sensitive to water diffusion throughout tissues, which correlates with Gleason score, a histological measure of prostate cancer aggressiveness. In this study the ability of DW-MRI to detect prostate cancer onset and development was evaluated in transgenic adenocarcinoma of the mouse prostate (TRAMP) mice. Materials and Methods T2-weighted and DW-MRI were acquired using a 7T MR scanner, 200 mm bore diameter; 10 TRAMP and 6 C57BL/6 control mice were scanned every 4 weeks from 8 weeks of age until sacrifice at 28–30 weeks. After sacrifice, the genitourinary tract was excised and sectioned for histological analysis. Histology slides registered with DW-MR images allowed for validation of DW-MR images and the apparent diffusion coefficient (ADC) as tools for cancer detection and disease stratification. An automated early assessment tool based on ADC threshold values was developed to aid cancer detection and progression monitoring. Results The ADC differentiated between control prostate ((1.86 ± 0.20) × 10−3 mm2/s) and normal TRAMP prostate ((1.38 ± 0.10) × 10−3 mm2/s) (P = 0.0001), between TRAMP prostate and well-differentiated cancer ((0.93 ± 0.18) × 10−3 mm2/s) (P = 0.0006), and between well-differentiated cancer and poorly differentiated cancer ((0.63 ± 0.06) × 10−3 mm2/s) (P = 0.02). Conclusion DW-MRI is a tool for early detection of cancer, and discrimination between cancer stages in the TRAMP model. The incorporation of DW-MRI-based prostate cancer stratification and monitoring could increase the accuracy of preclinical trials using TRAMP mice. J. Magn. Reson. Imaging 2016;43:1207–1217. This is the pre-peer reviewed version of the following article: Diffusion-weighted MRI for early detection and characterization of prostate cancer in the transgenic adenocarcinoma of the mouse prostate model, which has been published in final form at http://onlinelibrary.wiley.com/doi/10.1002/jmri.25087/abstract . This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Self-Archiving.

Purpose To study cancer associated with abnormal metabolism of phospholipids, of which several have been proposed as biomarkers for malignancy or to monitor response to anticancer therapy. We explored 3D 31P magnetic resonance spectroscopic imaging (MRSI) at high magnetic field for in vivo assessment of individual phospholipids in two patient-derived breast cancer xenografts representing good and poor prognosis (luminal- and basal-like tumors). Materials and Methods Metabolic profiles from luminal-like and basal-like xenograft tumors were obtained in vivo using 3D 31P MRSI at 11.7T and from tissue extracts in vitro at 14.1T. Gene expression analysis was performed in order to support metabolic differences between the two xenografts. Results In vivo 31P MR spectra were obtained in which the prominent resonances from phospholipid metabolites were detected at a high signal-to-noise ratio (SNR >7.5). Metabolic profiles obtained in vivo were in agreement with those obtained in vitro and could be used to discriminate between the two xenograft models, based on the levels of phosphocholine, phosphoethanolamine, glycerophosphocholine, and glycerophosphoethanolamine. The differences in phospholipid metabolite concentration could partly be explained by gene expression profiles. Conclusion Noninvasive metabolic profiling by 3D 31P MRSI can discriminate between subtypes of breast cancer based on different concentrations of choline- and ethanolamine-containing phospholipids. © 2014 Wiley Periodicals, Inc. This is the authors' manuscript to the article.

PURPOSE: Phospholipid metabolites are of importance in cancer studies, and have been suggested as candidate metabolic biomarkers for response to targeted anticancer drugs. The purpose of this study was to develop a phosphorus (31P) high resolution magic angle spinning (HR-MAS) magnetic resonance spectroscopy (MRS) protocol for quantification of phosphorylated metabolites in intact cancer tissue. METHODS: 31P spectra were acquired on a 14.1 T spectrometer with a triplet 1H/13C/31P MAS probe. Quantification of metabolites was performed using the PULCON principle. Basal-like and luminal-like breast cancer xenografts were treated with the dual PI3K/mTOR inhibitor BEZ235, and the impact of treatment on the concentration of phosphocholine (PCho), glycerophosphocholine (GPC), phosphoethanolamine (PEtn) and glycerophosphoethanolamine (GPE) was evaluated. RESULTS: In basal-like xenografts, BEZ235 treatment induced a significant decrease in PEtn (-25.6%, P = 0.01) whilst PCho (16.5%, P = 0.02) and GPC (37.3%, P < 0.001) were significantly increased. The metabolic changes could partially be explained by increased levels of phospholipase A2 group 4A (PLA2G4A). CONCLUSION: 31P HR-MAS MRS is a useful method for quantitative assessment of metabolic responses to PI3K inhibition. Using the PULCON principle for quantification, the levels of PCho, GPC, PEtn and GPE could be evaluated with high precision and accuracy. © 2013 Wiley Periodicals, Inc. This is the peer reviewed version of the article, which has been published in final form at dx.doi.org/10.1002/mrm.24869. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Self-Archiving.