Segmentation of dynamic PET images using cluster analysis

Quantitative PET studies can provide in-vivo measurements of dynamic physiological and biochemical processes in humans. A limitation of PET is its inability to provide precise anatomic localisation due to relatively poor spatial resolution when compared to MR imaging. Manual placement of regions of interest (ROIs) is commonly used in the clinical and research settings in analysis of PET datasets. However, this approach is operator dependent and time-consuming. Semi- or fully-automated ROI delineation (or segmentation) methods offer advantages by reducing operator error and subjectivity and thereby improving reproducibility. In this work, the authors describe an approach to automatically segment dynamic PET images using cluster analysis, and they validate their approach with a simulated phantom study and asses its performance in segmentation of dynamic lung data. The authors' preliminary results suggest that cluster analysis can be used to automatically segment tissues in dynamic PET studies and has the potential to replace manual ROI delineation