Final Report Summary - PHASETOMO (Development of a three-dimensional Reconstruction Algorithm for Phase Contrast Breast Tomosynthesis)
X-ray Mammography is considered the best practice for screening of breast cancer. However, due to insufficient soft tissue contrast and the effect of overlying structures in the planar imaging, a significant number of the cancers still remain undetected. New techniques are continuously developed to improve the cancer detection – Phase Contrast (PhC) setups enhance the edges of the structures in the x-ray projections, while Digital Tomosynthesis (DTS) is adding three-dimensional information to the results of the examinations. An approach that combines the advantages of both X-ray Phase Contrast (PhC) imaging and digital breast tomosynthesis would result in high contrast tomograms with well outlined breast cancer margins. The overall goal of the PHASETOMO project has been to develop, test and validate a three-dimensional reconstruction algorithm for PhC breast tomosynthesis. Specific project objectives were: (i) development of simulation tool for generation of PhC images over a limited angular range; (ii) development of reconstruction algorithm for PhC tomosynthesis; and (iii) validation and optimisation of the PhC platform. The project implementation was accomplished in three phases that correspond to the three specific objectives.
A significant part of the work was dedicated to the development and validation of a computer-based simulation platform for planar PhC projection images and images for three-dimensional PhC breast imaging research. The software platform has been based on a previously developed software tool for to x-ray imaging simulation with modules for object creation and x-ray image formation. To simulate PhC effects, those modules have been updated to take into account the refractive index of the simulated materials and and further developed to implement the Fresnel-Kirchhoff diffraction theory for x-ray waves propagation. Projection images can now be generated in an in-line PhC mode for several acquisition geometry configurations – planar imaging, tomosynthesis and computed tomography. PhC simulations are available for simple and complex (anthropomorphic) phantoms.
The platform has been validated to correctly and realistically generate PhC images mainly by comparing to experimental PhC images. Simple in shape and content, as well as, more complex in-house developed physical phantoms were used for the validation. Measurements were performed at beamline ID17 at the European Synchrotron Radiation Facility (ESRF), Grenoble. The results show very good correlation between simulated and experimental images. PhC images of an anthropomorphic breast phantom were generated using the platform and reported for first time. The improved visibility of mammographic structures preliminary suggests further investigation and optimisation of the PhC technique dedicated to breast imaging, especially when abnormalities are present.
Another part of the work was dedicated to the development of a reconstruction algorithm for PhC breast tomosynthesis. Just as with x-ray projection mammography, PhC projection imaging lacks depth-of-focus information. The use of tomosynthesis overcomes this limitation. Two basic back-projection type of algorithms have been adopted for reconstruction of PhC tomograms. The software simulation platform was further extended to include the reconstruction and was thoroughly validated, considering a breast tomosynthesis setup. The validation was accomplished again using experimental data, acquired at beamline ID17, ESRF, Grenoble. The developed platform was used to show an improvement of the detection task characterising precisely the location and dimensions of the objects within the phantom, while preserving the edge enhancement observed in the planar PhC projection images.
Besides the scientific work, the fellow was involved in the organisation and participated in a number of seminars, workshops during the integration grant period. A special panel during an IEEE conference at the Host institution was initiated, giving focus on new techniques for screening and diagnosing the breast cancer. The research and the organisational activities of the fellow at the Host institution – Technical University of Varna, Bulgaria, resulted in the actual establishment of a team, working in the wider field of Biomedical engineering. A new Horizon2020 project “Three dimensional breast cancer models for X-ray Imaging research – MaXIMA”, supporting the research initiatives from PHASETOMO, began in January 2016. A new research Laboratory on Computer Simulations in Medicine, has been established and led by the fellow. Since 2015, the fellow has undertaken teaching activities at the Host institution. Outcomes from the research have been already included in the educational process within the M.Sc. program in the field of Medical Electronics, and in the module “Anthropomorphic Phantoms” from the European Course EUTEMPE-RX: European Training and Education for Medical Physics Experts in Radiology. In June 2016, the fellow applied for a position of the Associate Professor and is currently awaiting nomination. That will be an important additional part of her entire integration at the Host institution.
A significant part of the work was dedicated to the development and validation of a computer-based simulation platform for planar PhC projection images and images for three-dimensional PhC breast imaging research. The software platform has been based on a previously developed software tool for to x-ray imaging simulation with modules for object creation and x-ray image formation. To simulate PhC effects, those modules have been updated to take into account the refractive index of the simulated materials and and further developed to implement the Fresnel-Kirchhoff diffraction theory for x-ray waves propagation. Projection images can now be generated in an in-line PhC mode for several acquisition geometry configurations – planar imaging, tomosynthesis and computed tomography. PhC simulations are available for simple and complex (anthropomorphic) phantoms.
The platform has been validated to correctly and realistically generate PhC images mainly by comparing to experimental PhC images. Simple in shape and content, as well as, more complex in-house developed physical phantoms were used for the validation. Measurements were performed at beamline ID17 at the European Synchrotron Radiation Facility (ESRF), Grenoble. The results show very good correlation between simulated and experimental images. PhC images of an anthropomorphic breast phantom were generated using the platform and reported for first time. The improved visibility of mammographic structures preliminary suggests further investigation and optimisation of the PhC technique dedicated to breast imaging, especially when abnormalities are present.
Another part of the work was dedicated to the development of a reconstruction algorithm for PhC breast tomosynthesis. Just as with x-ray projection mammography, PhC projection imaging lacks depth-of-focus information. The use of tomosynthesis overcomes this limitation. Two basic back-projection type of algorithms have been adopted for reconstruction of PhC tomograms. The software simulation platform was further extended to include the reconstruction and was thoroughly validated, considering a breast tomosynthesis setup. The validation was accomplished again using experimental data, acquired at beamline ID17, ESRF, Grenoble. The developed platform was used to show an improvement of the detection task characterising precisely the location and dimensions of the objects within the phantom, while preserving the edge enhancement observed in the planar PhC projection images.
Besides the scientific work, the fellow was involved in the organisation and participated in a number of seminars, workshops during the integration grant period. A special panel during an IEEE conference at the Host institution was initiated, giving focus on new techniques for screening and diagnosing the breast cancer. The research and the organisational activities of the fellow at the Host institution – Technical University of Varna, Bulgaria, resulted in the actual establishment of a team, working in the wider field of Biomedical engineering. A new Horizon2020 project “Three dimensional breast cancer models for X-ray Imaging research – MaXIMA”, supporting the research initiatives from PHASETOMO, began in January 2016. A new research Laboratory on Computer Simulations in Medicine, has been established and led by the fellow. Since 2015, the fellow has undertaken teaching activities at the Host institution. Outcomes from the research have been already included in the educational process within the M.Sc. program in the field of Medical Electronics, and in the module “Anthropomorphic Phantoms” from the European Course EUTEMPE-RX: European Training and Education for Medical Physics Experts in Radiology. In June 2016, the fellow applied for a position of the Associate Professor and is currently awaiting nomination. That will be an important additional part of her entire integration at the Host institution.