Dedicated breast PET scanners have been intensively investigated in recent years as their increased sensitivity and resolution are hypothesized to result in improvements in the early detection and determination of treatment response for breast cancer. However, there is currently no well-defined protocol for accurately evaluating their performance. This work aims to determine parameters for a set of phantoms for estimating count rate performances in pendant breast PET scanners.

A collection of 216 dedicated pendant breast CT images were divided into quartiles of overall volume (v), average radius ( r), and total length (l) of the breast. The 25th, 50th, and 75th quartile values of r and l were the parameters used for a set of 3 cylindrical phantoms. The cylindrical shape was chosen to remain consistent with the current NEMA whole body and small animal protocols. These cylinders, made out of high density polyethylene, are the small, medium, and large breast phantoms. For each phantom, the line source is placed at a radial offset of 4/5 of the phantom radius. Out-of-field of view activity is modeled using the NEMA-NU2 2007 scatter phantom. The parameters for each cylindrical phantom are as follows: Small, r = 4.2 cm, l = 7.1 cm; Medium, r = 4.8 cm, l = 9.0 cm; Large, r = 5.4 cm, l = 10.3 cm.

The activity ratios were determined by matching the patient prompts minus randoms (P − R) rate in each breast phantom and the remaining patient randoms (R) rate in the whole body phantom. The ratio of activity in the whole body phantom to activity in the breast phantoms is as follows: Small, 105:1; Medium, 40:1; Large, 22:1.