Phantom for evaluation of dedicated nuclear medicine breast imaging systems

Objectives Nuclear breast imaging requires high sensitivity and resolution close to the camera face for detection of small lesions with low target to background count ratios (T/B). Manufacturers have developed planar imaging systems with pixelated detectors and optimized collimation. Most quality control phantoms are oversized, SPECT specific and utilize high-contrast cold defects. A new phantom has been designed to better emulate the breast-imaging scenario by producing small areas of increased counts at varying T/B.

Methods The phantom was constructed by layering eight acrylic sheets. Six 2.4 mm thick interior sheets were cut to create an easily filled 16x23 cm chamber divided into sections of 1/6, 2/6, 3/6 or 4/6 layers of acrylic/water. Target activity is an array of 1.6 to 16 mm diameter holes in each section. Images appear to have structures of different sizes and sectional T/B of 1.2, 1.5, 2.0 and 3.0. The phantom was filled with ^99mTc solution and imaged with a CZT molecular breast imaging system (D750b), a pixelated scintillation camera with LEHS (2020HS) and LEHR (2020HR) collimation, and conventional LEHR scintillation cameras. The phantom was placed directly on and 3.5 cm above (separated by a water bath) the collimator for high (optimal performance) and patient equivalent count images.

Results The table shows the smallest structure visualized with the phantom directly on the collimator. When a scattering medium was introduced, image quality degraded by varying degrees for all systems. The 16 mm structure provided a measure of contrast without significant partial volume effects.

Conclusions A phantom has been designed for assessment of dedicated nuclear medicine breast imaging systems. Images of the phantom demonstrate that it can be used to evaluate the systems in a breast-specific setting. Images also demonstrate the advantage of dedicated breast imaging systems over those designed for general imaging.

Research Support This work was supported in part by GE Healthcare.