Phantom-Based Evaluation of a Novel Tomographic Ultra-Fast Molecular Breast Imaging Prototype

A novel stationary tomographic Ultra-Fast Molecular Breast Imaging (UFMBI) technology was developed to reduce acquisition time and radiation dose of Molecular Breast Imaging (MBI), enabling its wider acceptance for supplemental dense breast screening. The system combines a stationary dual-head gamma camera made of CZT detectors with densely packed multi-pinhole collimators and novel de-multiplexing reconstruction methods, exploiting the ultra-high 3D position resolution of CZT detectors and their excellent energy resolution. A small field-of-view prototype (132 mm × 88 mm) was designed, built and calibrated. Raw detector data were processed using a machine-learning model providing 10×10 subpixelisation of 2 mm pixels in 5 mm thick CZT detectors and depth-of-interaction (DOI) estimation from anode signals alone. Tomographic images were reconstructed using an OSEM algorithm, including de-multiplexing and corrections for attenuation and collimator response. The prototype was evaluated using a 99mTc-filled phantom with lesions of different sizes and uniform background acquired separately. Prototype performance was compared with a clinical planar MBI system by imaging 99mTc-infused gel phantoms, with lesion detectability assessed via contrast-to-noise ratio (CNR). The results from the spheres' phantom indicated that 5-7 mm diameter lesions with tumour-to-background ratios of 10-20:1 were detectable even with a dose/time reduction factor (DRF) of 8 relative to current MBI clinical practice. The gel phantom experiments showed that, with equal activity concentration and scanning time (DRF=1), the UFMBI prototype provided 2.5 to 3.9-fold CNR of the clinical planar system depending on the lesion activity and position within the field-of-view; with DRF of 6, UFMBI provided 1.6 to 2.4-fold CNR of the clinical system, improving to 2.5 to 3.4-fold with application of non-local means noise filtering. These results demonstrate the feasibility of UFMBI as a low-dose, high throughput approach and support its further development toward cancer screening of dense breasts and image-guided biopsy applications.