The Thin-TOF PET (TT-PET) project aims at the construction of a small-animal PET scanner based on monolithic silicon pixel sensors. The iterative image reconstruction was developed using a maximum a posteriori probability estimation algorithm. Reconstructed images from Monte-Carlo data show high signal-to-noise ratio with unprecedented spatial resolution. The detector ASICs were designed with the support of a Technology CAD simulation framework. A measurement setup based on a 22Na radioactive source showed that the single layer detection efficiency of the demonstrator chip is approximately 0.19%, which is in good agreement with the simulations. The same setup was used to measure the time resolution, that resulted to be 70ps (RMS). This result, although excellent for a silicon sensor, is a factor two below the specification of the TT-PET project. It is created by a systematic error in the measurement of the charge that will be corrected in future chips.