Radiomics software for breast imaging optimization and simulation studies.

BACKGROUND AND OBJECTIVE: The development, control and optimisation of new x-ray breast imaging modalities could benefit from a quantitative assessment of the resulting image textures. The aim of this work was to develop a software tool for routine radiomics applications in breast imaging, which will also be available upon request. METHODS: The tool (developed in MATLAB) allows image reading, selection of Regions of Interest (ROI), analysis and comparison. Requirements towards the tool also included convenient handling of common medical and simulated images, building and providing a library of commonly applied algorithms and a friendly graphical user interface. Initial set of features and analyses have been selected after a literature search. Being open, the tool can be extended, if necessary. RESULTS: The tool allows semi-automatic extracting of ROIs, calculating and processing a total of 23 different metrics or features in 2D images and/or in 3D image volumes. Computations of the features were verified against computations with other software packages performed with test images. Two case studies illustrate the applicability of the tool - (i) features on a series of 2D 'left' and 'right' CC mammograms acquired on a Siemens Inspiration system were computed and compared, and (ii) evaluation of the suitability of newly proposed and developed breast phantoms for x-ray-based imaging based on reference values from clinical mammography images. Obtained results could steer the further development of the physical breast phantoms. CONCLUSIONS: A new image analysis toolbox was realized and can now be used in a multitude of radiomics applications, on both clinical and test images.

Development of breast lesions models database.

PURPOSE: We present the development and the current state of the MaXIMA Breast Lesions Models Database, which is intended to provide researchers with both segmented and mathematical computer-based breast lesion models with realistic shape. METHODS: The database contains various 3D images of breast lesions of irregular shapes, collected from routine patient examinations or dedicated scientific experiments. It also contains images of simulated tumour models. In order to extract the 3D shapes of the breast cancers from patient images, an in-house segmentation algorithm was developed for the analysis of 50 tomosynthesis sets from patients diagnosed with malignant and benign lesions. In addition, computed tomography (CT) scans of three breast mastectomy cases were added, as well as five whole-body CT scans. The segmentation algorithm includes a series of image processing operations and region-growing techniques with minimal interaction from the user, with the purpose of finding and segmenting the areas of the lesion. Mathematically modelled computational breast lesions, also stored in the database, are based on the 3D random walk approach. RESULTS: The MaXIMA Imaging Database currently contains 50 breast cancer models obtained by segmentation of 3D patient breast tomosynthesis images; 8 models obtained by segmentation of whole body and breast cadavers CT images; and 80 models based on a mathematical algorithm. Each record in the database is supported with relevant information. Two applications of the database are highlighted: inserting the lesions into computationally generated breast phantoms and an approach in generating mammography images with variously shaped breast lesion models from the database for evaluation purposes. Both cases demonstrate the implementation of multiple scenarios and of an unlimited number of cases, which can be used for further software modelling and investigation of breast imaging techniques. The created database interface is web-based, user friendly and is intended to be made freely accessible through internet after the completion of the MaXIMA project. CONCLUSIONS: The developed database will serve as an imaging data source for researchers, working on breast diagnostic imaging and on improving early breast cancer detection techniques, using existing or newly developed imaging modalities.

Models of breast lesions based on three-dimensional X-ray breast images.

This paper presents a method for creation of computational models of breast lesions with irregular shapes from patient Digital Breast Tomosynthesis (DBT) images or breast cadavers and whole-body Computed Tomography (CT) images. The approach includes six basic steps: (a) normalization of the intensity of the tomographic images; (b) image noise reduction; (c) binarization of the lesion area, (d) application of morphological operations to further decrease the level of artefacts; (e) application of a region growing technique to segment the lesion; and (f) creation of a final 3D lesion model. The algorithm is semi-automatic as the initial selection of the region of the lesion and the seeds for the region growing are done interactively. A software tool, performing all of the required steps, was developed in MATLAB. The method was tested and evaluated by analysing anonymized sets of DBT patient images diagnosed with lesions. Experienced radiologists evaluated the segmentation of the tumours in the slices and the obtained 3D lesion shapes. They concluded for a quite satisfactory delineation of the lesions. In addition, for three DBT cases, a delineation of the tumours was performed independently by the radiologists. In all cases the abnormality volumes segmented by the proposed algorithm were smaller than those outlined by the experts. The calculated Dice similarity coefficients for algorithm-radiologist and radiologist-radiologist showed similar values. Another selected tumour case was introduced into a computational breast model to recursively assess the algorithm. The relative volume difference between the ground-truth tumour volume and the one obtained by applying the algorithm on the synthetic volume from the virtual DBT study is 5% which demonstrates the satisfactory performance of the proposed segmentation algorithm. The software tool we developed was used to create models of different breast abnormalities, which were then stored in a database for use by researchers working in this field.

Suitability of low density materials for 3D printing of physical breast phantoms.

Breast physical phantoms are a basic tool for the assessment and verification of performance standards in daily clinical practice of x-ray breast imaging modalities. They are also invaluable in testing and evaluation of new x-ray breast modalities to be potentially established, e.g. breast computed tomography, dual-energy breast CT and phase-contrast mammography and tomography. Nowadays, there is a lack or there are only a limited number of breast physical phantoms available for this purpose. The aim of this study is to explore a range of 3D printing materials such as resins, PLA, ABS, Nylon etc, to determine their attenuation and refractive properties, and to finally compare them to the properties of the breast tissues: adipose, glandular and skin. To achieve this goal, step-wedge phantoms were computationally modeled and then manufactured using stereolithographic and fused-deposition modeling technologies. X-ray images of the phantoms were acquired, using monochromatic beam at ID17, ESRF, Grenoble for three energies-30 keV, 45 keV and 60 keV. Experimental data were further processed to obtain the linear attenuation coefficients of these materials. Comparison with theoretical data for the linear attenuation coefficients and the refractive indexes for breast tissues was performed. From the studied materials, most of the resins, Nylon, Hybrid, PET-G show absorption properties close to the glandular tissue, while ABS shows absorption characteristics close to these of the adipose tissue. For phase-contrast imaging, it turns out that the ABS combined with resin-based materials to represent the adipose and glandular tissues, respectively may be a good combination for manufacturing of a phantom suitable for these studies. These results can be used for the design and the construction of a new physical anthropomorphic phantom of the breast with improved anatomical and radiological characteristics dedicated for advanced mammography imaging techniques implemented at higher photon energies.

Evaluation of a breast software model for 2D and 3D X-ray imaging studies of the breast.

INTRODUCTION: In X-ray imaging, test objects reproducing breast anatomy characteristics are realized to optimize issues such as image processing or reconstruction, lesion detection performance, image quality and radiation induced detriment. Recently, a physical phantom with a structured background has been introduced for both 2D mammography and breast tomosynthesis. A software version of this phantom and a few related versions are now available and a comparison between these 3D software phantoms and the physical phantom will be presented. METHODS: The software breast phantom simulates a semi-cylindrical container filled with spherical beads of different diameters. Four computational breast phantoms were generated with a dedicated software application and for two of these, physical phantoms are also available and they are used for the side by side comparison. Planar projections in mammography and tomosynthesis were simulated under identical incident air kerma conditions. Tomosynthesis slices were reconstructed with an in-house developed reconstruction software. In addition to a visual comparison, parameters like fractal dimension, power law exponent ß and second order statistics (skewness, kurtosis) of planar projections and tomosynthesis reconstructed images were compared. RESULTS: Visually, an excellent agreement between simulated and real planar and tomosynthesis images is observed. The comparison shows also an overall very good agreement between parameters evaluated from simulated and experimental images. CONCLUSION: The computational breast phantoms showed a close match with their physical versions. The detailed mathematical analysis of the images confirms the agreement between real and simulated 2D mammography and tomosynthesis images. The software phantom is ready for optimization purpose and extrapolation of the phantom to other breast imaging techniques.

High-flexibility combinatorial peptide synthesis with laser-based transfer of monomers in solid matrix material.

Laser writing is used to structure surfaces in many different ways in materials and life sciences. However, combinatorial patterning applications are still limited. Here we present a method for cost-efficient combinatorial synthesis of very-high-density peptide arrays with natural and synthetic monomers. A laser automatically transfers nanometre-thin solid material spots from different donor slides to an acceptor. Each donor bears a thin polymer film, embedding one type of monomer. Coupling occurs in a separate heating step, where the matrix becomes viscous and building blocks diffuse and couple to the acceptor surface. Furthermore, we can consecutively deposit two material layers of activation reagents and amino acids. Subsequent heat-induced mixing facilitates an in situ activation and coupling of the monomers. This allows us to incorporate building blocks with click chemistry compatibility or a large variety of commercially available non-activated, for example, posttranslationally modified building blocks into the array's peptides with >17,000 spots per cm(2).

Computer aided preoperative evaluation of the residual liver volume using computed tomography images.

Major hepatectomy causes a risk of postoperative liver dysfunction, failure, and infections like surgical site infection. Preoperative assessment of the liver volume and function of the remnant liver is a mandatory prerequisite before performing such surgery. The aim of this work is to develop and test a software application for evaluation of the residual function of the liver prior to the intervention of the surgeons. For this purpose, a technique for evaluation of liver volume from computed tomography (CT) images has been developed. Furthermore, the methodology algorithms were implemented and incorporated within a software tool with three basic functionalities: volume determination based on segmentation of liver from CT images, virtual tumour resection and estimation of the residual liver function and 3D visualisation. Forty-one sets of abdominal CT images consisting of different number of tomographic slice images were used to test and evaluate the proposed approach. Volumes that were obtained after manual tracing by two surgeon experts showed a relative difference of 3.5 %. The suggested methodology was encapsulated within an application with user-friendly interface that allows surgeons interactively to perform virtual tumour resection, to evaluate the relative residual liver and render the final result. Thereby, it is a tool in the surgeons' hands that significantly facilitates their duties, saves time, and allows them to objectively evaluate the situation and take the right decisions. At the same time, the tool appears to be appropriate educational instrument for virtual training of young surgeon specialists.

Studies on circular isocentric cone-beam trajectories for 3D image reconstructions using FDK algorithm.

Image reconstruction from X-ray cone-beam projections collected along a single-circular source trajectory is commonly done using the Feldkamp (FDK) algorithm which performs well only with a small cone-angle. Although this method does not provide an "exact" reconstruction, the approximation is considered adequate for many purposes. In FDK reconstruction the degree of inaccuracy is highly object-dependent, and the largest errors are to be expected for planes parallel to and remote from the midplane. In this study we investigated the possibility to accurately reconstruct these regions by applying FDK algorithm along three-orthogonal to each other circular scanning trajectories. After appropriate weighting, based on the expected errors for each individual reconstruction, the final 3D volume contains the most precisely recovered values. By comparing the quality of 3D reconstructed images using FDK algorithm on projections acquired along classical single-circular and two- and three-orthogonal circular trajectories, we show that using three-orthogonal circular isocentric orbits with an error-based weighted averaging, image quality of reconstructed slices significantly improves, reconstruction error due to circular scanning is reduced and becomes almost independent of the slice position even for relatively large cone-angles.

Estimation of the heart respiratory motion with applications for cone beam computed tomography imaging: a simulation study.

Computed tomography (CT) reconstruction methods assume imaging of static objects; object movement during projection data acquisition causes tomogram artifacts. The continuously moving heart, therefore, represents a complicated imaging case. The associated problems due to the heart beating can be overcome either by using very short projection acquisition times, during which the heart may be considered static, or by ECG-gated acquisition. In the latter case, however, the acquisition of a large number of projections may not be completed in a single breath hold, thus heart displacement occurs as an additional problem. This problem has been addressed by applying heart motion models in various respiratory motion compensation algorithms. Our paper focuses on cone beam computed tomography (CBCT), performed in conjunction with isocentric, fluoroscopic equipment, and continuous ECG and respiratory monitoring. Such equipment is used primarily for in-theater three-dimensional (3-D) imaging and benefits particularly from the recent developments in flat panel detector technologies. The objectives of this paper are: i) to develop a model for the motion of the heart due to respiration during the respiratory cycle; ii) to apply this model to the tomographic reconstruction algorithm, in order to account for heart movement due to respiration in the reconstruction; and iii) to initially evaluate this method by means of simulation studies. Based on simulation studies, we were able to demonstrate that heart displacement due to respiration can be estimated from the same projection data, required for a CBCT reconstruction. Our paper includes semiautomatic segmentation of the heart on the X-ray projections and reconstruction of a convex 3-D-heart object that performs the same motion as the heart during respiration, and use of this information into the CBCT reconstruction algorithm. The results reveal significant image quality improvements in cardiac image reconstruction.