A numerical observer study of MAP with anatomical and functional priors for lesion detection

In a previous study, we described the use of a functional prior in addition to an anatomical prior to improve the contrast and quantification of lesions in PET imaging. The functional prior is specified by applying a threshold to a preliminary reconstructed PET image. The segmented regions are quarantined from the regularization during reconstruction, resulting in a form of adaptive smoothing. One additional parameter was added to control the weighting of the functional prior. We call this dual-prior technique quarantine maximum a posteriori MAP reconstruction (QMAP). In this study, we used the channelized Hotelling observer (CHO) to investigate the effect of QMAP on lesion detectability. The results show that the inclusion of a functional prior in addition to an anatomical prior improves the performance of MAP reconstruction in a lesion detection task compared with the use of anatomical or smoothing priors on their own.