An Investigation of PET Scan Parameters for Lesion Detection With GE VISTA Small-Animal Scanner Using Monte Carlo Simulation and Multivariate Analysis

In vivo detection of animal xenograft is a major application of positron emission tomography (PET) imaging. This study investigated the relationship among the scan time, radioactivity, and radiation dose to achieve optimal lesion detectability while minimizing physiologic and pharmacologic effects of imaging procedures in small animal PET studies. A small-animal PET system was modeled based on Monte Carlo simulation to generate the emission image and dose distribution. A multivariate approach was used to investigate the simultaneous effects of tumor size, target-to-background ratio (TBR), scan duration, and injected radioactivity on the contrast-to-noise ratio (CNR) and recovery coefficient (RC). Among the four predictors, TBR and scan time were the most relevant contributors of CNR and RC variations, respectively. In 1.86 10⁵ Bq/ml injected activity, the absorbed doses for a body and tumor with TBR = 2 were 2.46 and 5.39 cGy, respectively. A substantial improvement in CNR or RC was not observed in images acquired with radiotracer activity larger than 9.3 10⁴ Bq/ml and scan duration longer than 30 min. The coefficient of determination was greater than 0.93 for both regression models. Although the improvement of counting statistics by increasing scan duration and injected activity can reduce statistical noise and improve apparent spatial resolution, it is crucial to maintain the radiation exposure and anesthetic dose received by animals as low as possible to reduce biological damage.