RESUMEN
PURPOSE: To assess the 3D geometric sampling accuracy of a new PET-guided system for breast cancer biopsy (BCB) from areas within the tumour with high 18F-FDG uptake. MATERIALS AND METHODS: In the context of the European Union project MammoCare, a prototype semi-robotic stereotactic prototype BCB-device was incorporated into a dedicated high resolution PET-detector for breast imaging. The system consists of 2 stacked rings, each containing 12 plane detectors, forming a dodecagon with a 186mm aperture for 3D reconstruction (1mm3 voxel). A vacuum-assisted biopsy needle attached to a robot-controlled arm was used. To test the accuracy of needle placement, the needle tip was labelled with 18F-FDG and positioned at 78 target coordinates distributed over a 35mm×24mm×28mm volume within the PET-detector field-of-view. At each position images were acquired from which the needle positioning accuracy was calculated. Additionally, phantom-based biopsy proofs, as well as MammoCare images of 5 breast cancer patients, were evaluated for the 3D automated locating of 18F-FDG uptake areas within the tumour. RESULTS: Needle positioning tests revealed an average accuracy of 0.5mm (range 0-1mm), 0.6mm (range 0-2mm), and 0.4mm (range 0-2mm) for the x/y/z-axes, respectively. Furthermore, the MammoCare system was able to visualize and locate small (<10mm) regions with high 18F-FDG uptake within the tumour suitable for PET-guided biopsy after being located by the 3D automated application. CONCLUSIONS: Accuracy testing demonstrated high-precision of this semi-automatic 3D PET-guided system for breast cancer core needle biopsy. Its clinical feasibility evaluation in breast cancer patients scheduled for neo-adjuvant chemotherapy will follow.