The influence of radial undersampling schemes on compressed sensing reconstruction in breast MRI.

Fast imaging applications in magnetic resonance imaging (MRI) frequently involve undersampling of k-space data to achieve the desired temporal resolution. However, high temporal resolution images generated from undersampled data suffer from aliasing artifacts. In radial k-space sampling, this manifests as undesirable streaks that obscure image detail. Compressed sensing reconstruction has been shown to reduce such streak artifacts, based on the assumption of image sparsity. Here, compressed sensing is implemented with three different radial sampling schemes (golden-angle, bit-reversed, and random sampling), which are compared over a range of spatiotemporal resolutions. The sampling methods are implemented in static scenarios where different undersampling patterns could be compared. Results from point spread function studies, simulations, phantom and in vivo experiments show that the choice of radial sampling pattern influences the quality of the final image reconstructed by the compressed sensing algorithm. While evenly undersampled radial trajectories are best for specific temporal resolutions, golden-angle radial sampling results in the least overall error when various temporal resolutions are considered. Reduced temporal fluctuations from aliasing artifacts in golden-angle sampling translates to improved compressed sensing reconstructions overall.

Physics of MRI: a primer.

This article is based on an introductory lecture given for the past many years during the "MR Physics and Techniques for Clinicians" course at the Annual Meeting of the ISMRM. This introduction is not intended to be a comprehensive overview of the field, as the subject of magnetic resonance imaging (MRI) physics is large and complex. Rather, it is intended to lay a conceptual foundation by which magnetic resonance image formation can be understood from an intuitive perspective. The presentation is nonmathematical, relying on simple models that take the reader progressively from the basic spin physics of nuclei, through descriptions of how the magnetic resonance signal is generated and detected in an MRI scanner, the foundations of nuclear magnetic resonance (NMR) relaxation, and a discussion of the Fourier transform and its relation to MR image formation. The article continues with a discussion of how magnetic field gradients are used to facilitate spatial encoding and concludes with a development of basic pulse sequences and the factors defining image contrast.

Transurethral prostate magnetic resonance elastography: prospective imaging requirements.

Tissue stiffness is known to undergo alterations when affected by prostate cancer and may serve as an indicator of the disease. Stiffness measurements can be made with magnetic resonance elastography performed using a transurethral actuator to generate shear waves in the prostate gland. The goal of this study was to help determine the imaging requirements of transurethral magnetic resonance elastography and to evaluate whether the spatial and stiffness resolution of this technique overlapped with the requirements for prostate cancer detection. Through the use of prostate-mimicking gelatin phantoms, frequencies of at least 400 Hz were necessary to obtain accurate stiffness measurements of 10 mm diameter inclusions, but the detection of inclusions with diameters as small as 4.75 mm was possible at 200 Hz. The shear wave attenuation coefficient was measured in vivo in the canine prostate gland, and was used to predict the detectable penetration depth of shear waves in prostate tissue. These results suggested that frequencies below 200 Hz could propagate to the prostate boundary with a signal to noise ratio (SNR) of 60 and an actuator capable of producing 60 µm displacements. These requirements are achievable with current imaging and actuator technologies, and motivate further investigation of magnetic resonance elastography for the targeting of prostate cancer.

The feasibility of endorectal MR elastography for prostate cancer localization.

The objectives of this study were to evaluate the feasibility of using a rigid radio-frequency receiver endorectal coil for intracavitary prostate magnetic resonance elastography (MRE) and to demonstrate the capability of this technique for generating stiffness maps over a typical prostate volume. An endorectal coil is currently used to help improve the signal-to-noise ratio of images acquired with multiparametric magnetic resonance imaging. We propose that this same coil could also serve to generate shear waves in the prostate gland during imaging, opening up the possibility of incorporating prostate stiffness characterization into multiparametric magnetic resonance imaging. Prostate cancer has been shown to change the elasticity of tissue, suggesting that stiffness imaging (elastography) may provide supplementary diagnostic information. A rigid endorectal coil was mechanically coupled to a piezoceramic actuator and used to investigate full volume (27 slices, 2-mm thick) endorectal MRE in a prostate mimicking phantom. The low-amplitude vibrations (± 8-38 µm displacements) necessary to perform endorectal MRE did not affect the signal-to noise ratio of the coil and endorectal MRE was capable of resolving 0.1 cc (0.6 cm diameter) spherical inclusion volumes. Therefore, the results of this study, in combination with current clinical practice, motivate clinical evaluation of endorectal MRE in patients.

Supine breast MRI.

PURPOSE: To achieve high-quality unilateral supine breast magnetic resonance imaging (MRI) as a step to facilitate image aiding of clinical applications, which are often performed in the supine position. Contrast-enhanced breast MRI is a powerful tool for the diagnosis of cancer. However, prone patient positioning typically used for breast MRI hinders its use for image aiding. MATERIALS AND METHODS: A fixture and a flexible four-element receive coil were designed for patient-specific shaping and placement of the coil in close conformity to the supine breast. A 3D spoiled gradient sequence was modified to incorporate compensation of respiratory motion. The entire setup was tested in volunteer experiments and in a pilot patient study. RESULTS: The flexible coil design and the motion compensation produced supine breast MR images of high diagnostic value. Variations in breast shape and in tissue morphology within the breast were observed between a supine and a diagnostic prone MRI of a patient. CONCLUSION: The presented supine breast MRI achieved an image quality comparable to diagnostic breast MRI. Since supine positioning is common in many clinical applications such as ultrasound-guided breast biopsy or breast-conserving surgery, the registration of the supine images will aid these applications.

Improvement in DCIS detection rates by MRI over time in a high-risk breast screening study.

Although magnetic resonance imaging (MRI) is much more sensitive than mammography for detecting early invasive breast cancer, in many high-risk screening studies MRI was less sensitive than mammography for detecting ductal carcinoma in situ (DCIS). We reviewed our experience detecting DCIS in our single center study of annual MRI, mammography, ultrasound and clinical breast examination (CBE) for screening very high-risk women. All cases of DCIS±microinvasion and invasive cancer were compared in two time frames: before (period A) and after (period B) July 2001-when we acquired expertise in the detection of DCIS with MRI-with respect to patient demographics, method of detection, and rates of detection of invasive cancer and DCIS. In period A there were 15 cases (3.1% of 486 screens) in 223 women, of which 2 (13%) were DCIS-one with microinvasion-neither detected by MRI. In period B there were 29 cases (3.3% of 877 screens) in 391 women, of which 10 (34%) were DCIS±microinvasion (p=0.04), all 10 detected by MRI but only one by mammography. No DCIS cases were detected by ultrasound or CBE. Specificity was lower in period B than in period A but acceptable. The ability to detect DCIS with screening MRI improves significantly with experience. MRI-guided biopsy capability is essential for a high-risk screening program. In experienced centers the increased sensitivity of MRI relative to mammography is at least as high for DCIS as it is for invasive breast cancer.

Evaluation of multicoil breast arrays for parallel imaging.

PURPOSE: To evaluate three multicoil breast arrays for both conventional and SENSE-accelerated imaging. MATERIALS AND METHODS: Two commercially available 8-element coils and a prototype 16-element coil were compared. One 8-element array had adjustable coils located next to the breast tissue and the other had a fixed coil arrangement; both were designed to allow parallel imaging in the left-right direction. The 16-element coil was designed to have coil sensitivity variation in both the left-right and superior-inferior directions, and also had adjustable coils. Their performance was assessed in terms of signal-to-noise ratio (SNR), g-factor, and uniformity with a custom-built phantom. RESULTS: The 16-element array with adjustable coils provided the highest SNR, while the 8-element coil with a fixed coil arrangement had the best uniformity. All coils performed well for SENSE acceleration in the left-right direction. The 8-element coils did not have the capability for acceleration in the superior-inferior direction across the whole volume. The 16-element coil enabled acceleration in the superior-inferior direction in addition to the left-right direction. CONCLUSION: Smaller, adjustable coil elements located next to breast tissue can provide greater SNR than larger, fixed coil elements. A multicoil breast array with high intrinsic SNR and low g-factors enables high-quality parallel imaging.

Breast Cancer Mortality among Women with a BRCA1 or BRCA2 Mutation in a Magnetic Resonance Imaging Plus Mammography Screening Program.

Annual breast magnetic resonance imaging (MRI) plus mammography is the standard of care for screening women with inherited BRCA1/2 mutations. However, long-term breast cancer-related mortality with screening is unknown. Between 1997 and June 2011, 489 previously unaffected BRCA1/2 mutation carriers aged 25 to 65 years were screened with annual MRI plus mammography on our study. Thereafter, participants were eligible to continue MRI screening through the high-risk Ontario Breast Screening Program. In 2019, our data were linked to the Ontario Cancer Registry of Cancer Care Ontario to identify all incident cancers, vital status and causes of death. Observed breast cancer mortality was compared to expected mortality for age-matched women in the general population. There were 91 women diagnosed with breast cancer (72 invasive and 19 ductal carcinoma in situ (DCIS)) with median follow-up 7.4 (range: 0.1 to 19.2) years. Four deaths from breast cancer were observed, compared to 2.0 deaths expected (standardized mortality ratio (SMR) 2.0, p = 0.14). For the 489 women in the study, the probability of not dying of breast cancer at 20 years from the date of the first MRI was 98.2%. Annual screening with MRI plus mammography is a reasonable option for women who decline or defer risk-reducing mastectomy.

Temporal stability of adaptive 3D radial MRI using multidimensional golden means.

Breast tumor diagnosis requires both high spatial resolution to obtain information about tumor morphology and high temporal resolution to probe the kinetics of contrast uptake. Adaptive sampling of k-space allows images in dynamic contrast-enhanced (DCE)-magnetic resonance imaging (MRI) to be reconstructed at various spatial or temporal resolutions from the same dataset. However, conventional radial approaches have limited flexibility that restricts image reconstruction to predetermined resolutions. Golden-angle radial k-space sampling achieves flexibility in-plane with samples that are incremented by the golden angle, which fills two-dimensional (2D) k-space with radial spokes that have a relatively uniform angular distribution for any time interval. We extend this method to three-dimensional (3D) radial sampling, or 3D-Projection Reconstruction (3D-PR) using multidimensional golden means, which are derived from modified Fibonacci sequences by an eigenvalue approach. We quantitatively compare this technique to conventional 3D radial methods in terms of the fluctuation in error caused by undersampling artifacts, and show that the golden 3D-PR method can substantially improve the temporal stability of quantitative measurements made from dynamic images when compared to conventional 3D radial approaches of k-space sampling.

In vivo MR elastography of the prostate gland using a transurethral actuator.

Conventional approaches for MR elastography (MRE) using surface drivers have difficulty achieving sufficient shear wave propagation in the prostate gland due to attenuation. In this study we evaluate the feasibility of generating shear wave propagation in the prostate gland using a transurethral device. A novel transurethral actuator design is proposed, and the performance of this device was evaluated in gelatin phantoms and in a canine prostate gland. All MRI was performed on a 1.5T MR imager using a conventional gradient-echo MRE sequence. A piezoceramic actuator was used to vibrate the transurethral device along its length. Shear wave propagation was measured transverse and parallel to the rod at frequencies between 100 and 250 Hz in phantoms and in the prostate gland. The shear wave propagation was cylindrical, and uniform along the entire length of the rod in the gel experiments. The feasibility of transurethral MRE was demonstrated in vivo in a canine model, and shear wave propagation was observed in the prostate gland as well as along the rod. These experiments demonstrate the technical feasibility of transurethral MRE in vivo. Further development of this technique is warranted.

MR acoustic radiation force imaging: in vivo comparison to ultrasound motion tracking.

MR acoustic radiation force (ARF) imaging was developed for measuring tiSsue elastic properties using focused ultrasound to deliver a localized tissue motion. In this study, an imaging ultrasound transducer was mounted on the focused ultrasound transducer and ultrasound motion tracking was performed simultaneously to MR ARF imaging to validate the measurement results. In vivo studies on rabbit thigh muscle were performed and results showed a general agreement between the two modalities (slope=0.96 and R2=0.67). The temporal information by the ultrasound measurement indicates that the parameters in MR ARF imaging should be optimized according to the tissue type, acoustic power, and envelope and frequency of the ARF modulation.

Systematic review: using magnetic resonance imaging to screen women at high risk for breast cancer.

BACKGROUND: A sensitive and acceptable screening regimen for women at high risk for breast cancer is essential. Contrast-enhanced magnetic resonance imaging (MRI) of the breast is highly sensitive for diagnosis of breast cancer but has variable specificity. PURPOSE: To summarize the sensitivity, specificity, likelihood ratios, and posttest probability associated with adding MRI to annual mammography screening of women at very high risk for breast cancer. DATA SOURCES: English-language literature search of the MEDLINE, EMBASE, and Cochrane databases from January 1995 to September 2007, supplemented by hand searches of pertinent articles. STUDY SELECTION: Prospective studies published after 1994 in which MRI and mammography (with or without additional tests) were used to screen women at very high risk for breast cancer. DATA EXTRACTION: Methods and potential biases of studies were assessed by 2 reviewers, and data were extracted and entered into 2 x 2 tables that compared American College of Radiology Breast Imaging Reporting and Data System (BI-RADS) scores of MRI plus mammography, mammography alone, or MRI alone with results of breast tissue biopsies. DATA SYNTHESIS: Eleven relevant, prospective, nonrandomized studies that ranged from small single-center studies with only 1 round of patient screening to large multicenter studies with repeated rounds of annual screening were identified. Characteristics of women that varied across study samples included age range, history of breast cancer, and BRCA1 or BRCA2 mutation status. Studies used dynamic contrast-enhanced MRI with axial or coronal plane images (European studies) or sagittal images (North American studies) that were usually interpreted without knowledge of mammography results. The summary negative likelihood ratio and the probability of a BI-RADS-suspicious lesion (given negative test findings and assuming a 2% pretest probability of disease) were 0.70 (95% CI, 0.59 to 0.82) and 1.4% (CI, 1.2% to 1.6%) for mammography alone and 0.14 (CI, 0.05 to 0.42) and 0.3% (CI, 0.1% to 0.8%) for the combination of MRI plus mammography, using a BI-RADS score of 4 or higher as the definition of positive. LIMITATIONS: Differences in patient population, center experience, and criteria for positive screening results led to between-study heterogeneity. Data on patients with nonfamilial high risk were limited, and no data were available on recurrence or survival. CONCLUSION: Screening with both MRI and mammography might rule out cancerous lesions better than mammography alone in women who are known or likely to have an inherited predisposition to breast cancer.

Is mammography adequate for screening women with inherited BRCA mutations and low breast density?

BACKGROUND: Several observational studies have shown that magnetic resonance imaging (MRI) is significantly more sensitive than mammography for screening women over age 25 at high risk for hereditary breast cancer; however, MRI is more costly and less specific than mammography. We sought to determine the extent to which the low sensitivity of mammography is due to greater breast density. METHODS: Breast density was evaluated for all patients on a high-risk screening study who were diagnosed with breast cancer between November 1997 and July 2006. Density was measured in two ways: qualitatively using the four categories characterized by the Breast Imaging Reporting and Data System and quantitatively using a computer-aided technique and classified as (a) 50% density. Comparison of sensitivity of mammography (and MRI) for each individual density category and after combining the highest two and lowest two density categories was done using Fisher's exact test. RESULTS: A total of 46 breast cancers [15 ductal carcinoma in situ (DCIS) and 31 invasive] were diagnosed in 45 women (42 with BRCA mutations). Mean age was 48.3 (range, 32-68) years. Overall, sensitivity of mammography versus MRI was 20% versus 87% for DCIS and 26% versus 90% for invasive cancer. There was a trend towards greater mammographic sensitivity for invasive cancer in women with fattier breasts compared with those with greater breast density (37-43% versus 8-12%; P = 0.1), but this trend was not seen for DCIS. CONCLUSION: It is necessary to add MRI to mammography for screening women with BRCA mutations even if their breast density is low.

Preliminary in vivo validation of a dedicated breast MRI and sonographic coregistration imaging system.

OBJECTIVE: Sonographic correlation of breast MRI findings is often challenging. We present a preliminary in vivo feasibility study evaluating the degree of error of a new MRI-sonography coregistration system for showing MRI and sonographically visible breast lesions. CONCLUSION: In 10 patients with 13 lesions, the system was found to be an accurate means for targeting sonography to MRI of the same breast lesions.

Breast cancers detected with imaging screening in the BRCA population: emphasis on MR imaging with histopathologic correlation.

The benefit of screening with breast magnetic resonance (MR) imaging for certain patient populations at high risk for breast cancer, most notably patients with a genetic mutation in the BRCA1 or BRCA2 gene, has been established in numerous studies and is now becoming part of routine clinical practice. Despite the lower sensitivity of mammography compared with that of MR imaging, the former remains the standard of care for screening any patient population. In the BRCA1 and BRCA2 populations, the inferior sensitivity and specificity of ultrasonography (US) limit its role as a screening tool, but US remains a vital diagnostic tool because of its ability to provide guidance for biopsy of many suspicious lesions detected with MR imaging. Important features of a screening program with breast MR imaging include the following: optimization of the MR imaging technique, an awareness of the imaging features of invasive and noninvasive breast cancers detected with MR imaging, an understanding of the limitations of the various imaging modalities in both the initial screening and subsequent diagnostic work-up evaluations, and the requirement for MR imaging-guided biopsy.

An inverse problem solution for measuring the elastic modulus of intact ex vivo breast tissue tumours.

Soft tissue elasticity has been a subject of interest in biomedical applications as an aid to medical diagnosis since the dawn of medicine. More recently, this has led to the concept of elastography with the aim of imaging the spatial distribution of tissue elasticity. Interpreting elastography images requires reliable information pertaining to elastic properties of normal and pathological tissues. Such information is either very limited or not available in the literature. Elastic modulus measurement techniques developed for soft tissues generally require tissue excision to prepare samples for testing. While this may be done with normal tissues, tumour tissue excision is generally not permissible because tumour pathological assessment requires that the tumour be kept intact. To address this limitation, we developed a system to measure the Young's modulus of tumour specimens. The technique consists of indenting the tumour specimen while measuring indentation force and displacements. To obtain the Young's modulus from the measured force-displacement slope, we developed an iterative inversion technique that uses a finite element model of the piecewise homogeneous tissue slice in each iteration. Preliminary elasticity measurement results of various breast tumours are presented and discussed. These results indicate that the proposed method is robust and highly accurate. Furthermore, they indicate that a benign lesion and malignant tumours are roughly five times and ten times stiffer than normal breast tissues respectively.

Elastic moduli of normal and pathological human breast tissues: an inversion-technique-based investigation of 169 samples.

Understanding and quantifying the mechanical properties of breast tissues has been a subject of interest for the past two decades. This has been motivated in part by interest in modelling soft tissue response for surgery planning and virtual-reality-based surgical training. Interpreting elastography images for diagnostic purposes also requires a sound understanding of normal and pathological tissue mechanical properties. Reliable data on tissue elastic properties are very limited and those which are available tend to be inconsistent, in part as a result of measurement methodology. We have developed specialized techniques to measure tissue elasticity of breast normal tissues and tumour specimens and applied them to 169 fresh ex vivo breast tissue samples including fat and fibroglandular tissue as well as a range of benign and malignant breast tumour types. Results show that, under small deformation conditions, the elastic modulus of normal breast fat and fibroglandular tissues are similar while fibroadenomas were approximately twice the stiffness. Fibrocystic disease and malignant tumours exhibited a 3-6-fold increased stiffness with high-grade invasive ductal carcinoma exhibiting up to a 13-fold increase in stiffness compared to fibrogalndular tissue. A statistical analysis showed that differences between the elastic modulus of the majority of those tissues were statistically significant. Implications for the specificity advantages of elastography are reviewed.

A positive contrast, tri-modality tissue marker for breast tumour localization.

A new interstitial breast localization marker is proposed which exhibits positive contrast in T1-weighted MRI, ultrasound and x-ray mammography. Unlike previous markers which provide MRI contrast on the basis of a susceptibility-induced signal void, this marker provides a clear positive contrast without any loss of signal or spatial distortion. The marker is composed of 400 microm diameter copper microspheres suspended in a Gd-DTPA-doped gel matrix. Optimal contrast in T1-weighted spoiled gradient recalled MRI was found to occur with the addition of 10 mM Gd-DTPA. Ultrasound contrast was generated on the basis of scattering from the copper microspheres. X-ray contrast was provided by the high x-ray attenuation properties of the copper microspheres. The study demonstrates potential suitability of the marker for use as a breast localization marker based on ex vivo studies of chicken breast.

A novel microbubble construct for intracardiac or intravascular MR manometry: a theoretical study.

It has been demonstrated that gas-filled microbubble contrast agents, based on their volume changes, can serve as pressure probes in an MR field. It was recently reported that such an MR-based pressure measurement with microbubbles at 1.5 T must make use of microbubbles that have a volumetric magnetic susceptibility difference with the blood of at least 34 ppm in SI units. In this work, we show through analytical approximations and numerical simulations that such a microbubble formulation can be achieved by coating typical lipid-shelled microbubbles with particles of high dipole moment. Through finite-element simulations we demonstrate that the effective volumetric magnetic susceptibility of a coated microbubble is dependent on the radius, the shell volume fraction and the magnetic susceptibility of the particulates on the shell. Our calculations suggest that a suitable microbubble formulation which will be MR-sensitive to small pressure changes at 1.5 T must be 2-3 microm in radius and be uniformly coated with single-domain magnetic nanoparticles, such as magnetite, at shell volume fractions below 5%.

Development of an MRI/x-ray/ultrasound compatible marker for pre-operative breast tumour localization.

This paper describes an in vitro investigation into the composition, structure and development of an magnetic resonance imaging (MRI), ultrasound (US) and x-ray imaging compatible marker for breast tumour localization. The marker is composed of 0.4-0.6 mm glass and iron-containing aluminium microspheres suspended in a gelatin matrix. The final form of the marker is a cylindrical shape 7 mm long with 2.05 mm diameter to facilitate delivery through a 12 gauge biopsy needle. To get optimal reflectivity for the US contrast, the glass microsphere concentration was found to be 40% by weight. US contrast is independent of marker orientation and the cylindrical shape made its US signal appearance distinctive thus ensuring confident identification. To control the MRI contrast, iron content was varied to generate a clear and local susceptibility signal void to reflect the marker position. Optimal iron content was found to be 52 microg iron which produced a clear signal void in spoiled gradient recalled MR images. The appearance of the susceptibility artefact is determined by the marker's shape, orientation and echo time. The final marker produces a dark artefact in MRI while appears as a clear hyperintense structure with acoustic shadowing in US images. The x-ray image showed the marker as a radio-opaque structure. This in vitro study demonstrates that the marker forms an alternative to traditional wire localization currently used for breast surgical procedures and creates new opportunities for US guided surgical procedures.