[en] Mammographic interpretation often uses symmetry between left and right breasts to indicate the site of potential tumour masses. This approach has not been applied to breast images obtained from MRI. We present an automatic technique for breast symmetry detection based on feature extraction techniques which does not require any efforts to co-register breast MRI data. The approach applies computer-vision techniques to detect natural biological symmetries in breast MR scans based on three objective measures of similarity: multiresolution non-orthogonal wavelet representation, three-dimensional intensity distributions and co-occurrence matrices. Statistical distributions that are invariant to feature localization are computed for each of the extracted image features. These distributions are later compared against each other to account for perceptual similarity. Studies based on 51 normal MRI scans of randomly selected patients showed that the sensitivity of symmetry detection rate approached 94%. The symmetry analysis procedure presented in this paper can be applied as an aid in detecting breast tissue changes arising from disease

[en] Over the past decade, there has been increasing interest in modelling soft tissue deformation. This topic has several biomedical applications ranging from medical imaging to robotic assisted telesurgery. In these applications, tissue deformation can be very large due to low tissue stiffness and lack of physical constraints. As a result, deformation modelling of such organs often requires a treatment, which reflects nonlinear behaviour. While computational techniques such as nonlinear finite element methods are well developed, the required intrinsic nonlinear mechanical parameters of soft tissues that are critical to develop reliable tissue deformation models are not well known. To address this issue, we developed a system to measure the hyperelastic parameters of small ex vivo tissue samples. This measurement technique consists of indenting an unconfined small block of tissue using a computer controlled loading system while measuring the resulting indentation force. The nonlinear tissue force-displacement response is used to calculate the hyperelastic parameters via an appropriate inversion technique. This technique is based on a nonlinear least squares formulation that uses a nonlinear finite element model as the direct problem solver. The features of the system are demonstrated with two samples of breast tissue and typical hyperelastic results are presented

[en] 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

[en] Imaging of the microvasculature is often performed using contrast agents in combination with either ultrasound (US) or magnetic resonance (MR) imaging. Contrast agents are used to enhance medical imaging by highlighting microvascular properties and function. Dynamic signal changes arising from the passage of contrast agents through the microvasculature can be used to characterize different pathologies; however, comparisons across modalities are difficult due to differences in the interactions of contrast agents with the microvasculature. Better knowledge of the relationship of contrast enhancement patterns with both modalities could enable better characterization of tissue microvasculature. We developed a co-registration platform for multi-modal US and MR imaging using clinical imaging systems in order to study the relationship between US and MR contrast enhancement. A preliminary validation study was performed in phantoms to determine the registration accuracy of the platform. In phantoms, the in-plane registration accuracy was measured to be 0.2 ± 0.2 and 0.3 ± 0.2 mm, in the lateral and axial directions, respectively. The out-of-plane registration accuracy was estimated to be 0.5 mm ±0.1. Co-registered US and MR imaging was performed in a rabbit model to evaluate contrast kinetics in different tissue types after bolus injections of US and MR contrast agents. The arrival time of the contrast agent in the plane of imaging was relatively similar for both modalities. We studied three different tissue types: muscle, large vessels and fat. In US, the temporal kinetics of signal enhancement were not strongly dependent on tissue type. In MR, however, due to the different amounts of agent extravasation in each tissue type, tissue-specific contrast kinetics were observed. This study demonstrates the feasibility of performing in vivo co-registered contrast US and MR imaging to study the relationships of the enhancement patterns with each modality.

[en] A methodology for imposing approximate plane strain conditions in magnetic resonance elastography through physical constraint is described. Under plane strain conditions, data acquisition and analysis may be conducted in two dimensions, which reduces imaging and reconstruction time significantly compared with three-dimensional analysis. Simulations and experiments are performed to illustrate the constraint concept. A signal/noise analysis of a two-dimensional linear inversion technique for relative elastic modulus is undertaken, and modifications to the numerical method are described which can reduce the SNR requirements by a factor of two to four. Experimentally measured data are reconstructed to illustrate the performance of the method. (author)

[en] 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

[en] 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 μm 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. (note)

[en] 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 μg 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

[en] Over the past decade, several methods have been proposed to image tissue elasticity based on imaging methods collectively called elastography. While progress in developing these systems has been rapid, the basic understanding of tissue properties to interpret elastography images is generally lacking. To address this limitation, we developed a system to measure the Young's modulus of small soft tissue specimens. This system was designed to accommodate biological soft tissue constraints such as sample size, geometry imperfection and heterogeneity. The measurement technique consists of indenting an unconfined small block of tissue while measuring the resulting force. We show that the measured force-displacement slope of such a geometry can be transformed to the tissue Young's modulus via a conversion factor related to the sample's geometry and boundary conditions using finite element analysis. We also demonstrate another measurement technique for tissue elasticity based on quasi-static magnetic resonance elastography in which a tissue specimen encased in a gelatine-agarose block undergoes cyclical compression with resulting displacements measured using a phase contrast MRI technique. The tissue Young's modulus is then reconstructed from the measured displacements using an inversion technique. Finally, preliminary elasticity measurement results of various breast tissues are presented and discussed

[en] Breast MRI is frequently performed prior to breast conserving surgery in order to assess the location and extent of the lesion. Ideally, the surgeon should also be able to use the image information during surgery to guide the excision and this requires that the MR image is co-registered to conform to the patient’s position on the operating table. Recent progress in MR imaging techniques has made it possible to obtain high quality images of the patient in the supine position which significantly reduces the complexity of the registration task. Surface markers placed on the breast during imaging can be located during surgery using an external tracking device and this information can be used to co-register the images to the patient. There remains the problem that in most clinical MR scanners the arm of the patient has to be placed parallel to the body whereas the arm is placed perpendicular to the patient during surgery. The aim of this study is to determine the accuracy of co-registration based on a surface marker approach and, in particular, to determine what effect the difference in a patient’s arm position makes on the accuracy of tumour localization. Obtaining a second MRI of the patient where the patient’s arm is perpendicular to body axes (operating room position) is not possible. Instead we obtain a secondary MRI scan where the patient’s arm is above the patient’s head to validate the registration. Five patients with enhancing lesions ranging from 1.5 to 80 cm"3 in size were imaged using contrast enhanced MRI with their arms in two positions. A thin-plate spline registration scheme was used to match these two configurations. The registration algorithm uses the surface markers only and does not employ the image intensities. Tumour outlines were segmented and centre of mass (COM) displacement and Dice measures of lesion overlap were calculated. The relationship between the number of markers used and the COM-displacement was also studied. The lesion COM-displacements ranged from 0.9  to 9.3 mm and the Dice overlap score ranged from 20% to 80%. The registration procedure took less than 1 min to run on a standard PC. Alignment of pre-surgical supine MR images to the patient using surface markers on the breast for co-registration therefore appears to be feasible. (paper)