NoloSight has developed ProPET, an advanced probabilistic reconstruction and image analysis methodology that enhances lesion detection in PET/CT imaging. By combining informed prior models with machine learning, ProPET improves the ability to identify small lesions while maintaining clinical accuracy.
ProPET is a probabilistic reconstruction algorithm designed to improve PET image analysis. Unlike conventional reconstruction methods that provide a single optimized image, ProPET generates a posterior distribution of plausible activity maps. This approach integrates prior knowledge of tracer distribution with an explicit model of scanner physics—including the point spread function (PSF) and noise characteristics.
The result is enhanced lesion visibility, reduced noise, and improved spatial resolution, supporting earlier and more confident clinical decision-making.
The ProPET Framework:
ProPET applies a Bayesian approach to PET image analysis. The observed PET image is modeled as a convolution of the true activity distribution with the scanner’s PSF, plus noise. By incorporating:
This probabilistic framework enables real-time clinical application while maintaining theoretical rigor.