Predicting Small Lesion Detectability for a Small Animal TOF PET Scanner

One way to enhance lesion detectability in small animal positron emission tomography (PET) would be to improve the coincidence time resolution (CTR) significantly relative to the current state of the art. This may soon become a real opportunity with the recent launch of the worldwide 10 ps challenge. Shorter crystals may be needed to improve timing performance by reducing the scintillation transit time spread. Although this comes at the cost of sensitivity loss, the time-of-flight (TOF) gain may allow to compromise on sensitivity without degrading imaging performance. As the main advantage of TOF is to enhance contrast-to-noise ratio (CNR), there is a motivation to assess the tradeoffs between crystal length and TOF resolution on the lesion detectability through a quantitative model. To derive such a model, a set of simulations was performed using the LabPET II mouse-version scanner with GATE and NEMA NU4 standards. The images were reconstructed using the CASToR software for tomographic reconstruction. The CNR was evaluated through a factorial design and its outcomes used to derive a model to assess the CNR performance as a function of crystal length, TOF resolution, and lesion size using a polynomial regression with multiple variables. As expected, results showed that CNR performance has positive correlations with lesion size and gain in TOF resolution and a negative correlation with crystal length. This leads to a predictive model that estimates the CNR performance with a mean-squared error of 0.69.