Experimental Results of an Edgeless PET System for Breast Imaging: DeepBreast

Breast cancer causes the greatest number of cancer-related deaths among women. Screening programs using X-ray mammography have made possible to increase survival rates, but this technique presents some limitations. Positron Emission Tomography (PET) has demonstrated to be a potential way to overcome these drawbacks, offering more accurate diagnosis at initial staging and the possibility of monitoring the therapy treatment. However, further improvements in PET technology are still required. The proposed PET system, named DeepBreast, has been specifically designed to efficiently scan the breast in prone position. The system is based on a novel edgeless detector ring composed of curved monolithic scintillators glued together using a high-refractive index compound. This design minimizes the physical gaps between modules boosting the system sensitivity. It also improves the image quality since edge effects are reduced. For the x- and y-photon impact estimation, a Neural Network technique has been successfully implemented. The depth of interaction and energy were calibrated using a method based on Voronoi diagrams. The DeepBreast has been fully constructed and calibrated. Experimental data are being acquired following a modified version of the NEMA protocol for this dedicated system. A spatial resolution of 1.8, 2.0 and 1.5 mm FWHM was achieved at the center of the field of view (cFOV) along the radial, tangential and axial directions, respectively. These values remain almost constant over the entire FOV. A sensitivity of 2.3% was measured at the cFOV. Preliminary reconstructed images of phantoms show the ability of the DeepBreast system to provide high-quality images.