Partialvolume

Partial volume refers to the partial volume effect observed in imaging modalities such as PET, CT, and MRI. It arises when the voxel size of an image is not small enough relative to the spatial structures of interest, so a single voxel contains more than one tissue type or material. In such cases the measured signal is effectively a weighted average of the true signals from the constituent tissues, leading to biased or blurred representations of anatomy or function.

Causes include finite spatial resolution, the point spread function of the imaging system, slice thickness, and sampling geometry. The observed signal in a voxel can be modeled as S = sum w_i T_i, where w_i are the fractional volumes of each tissue within the voxel.

In practice this produces blurred boundaries between tissues, reduced contrast, and systematic errors in quantification. In PET, partial volume effects depress observed activity in small structures and can over- or under-estimate activity in surrounding tissue. In MRI and CT, they can complicate tissue segmentation and morphometric analyses.

Partial volume correction (PVC) methods aim to recover more accurate signals. Strategies include segmentation-based PVC using high-resolution anatomical images, geometric transfer matrix methods, voxel-wise PVC with estimated point spread function, and deconvolution approaches. Higher-resolution imaging, partial volume maps, and super-resolution techniques can also help, as can integrating multiple modalities to constrain tissue fractions.

Partial volume considerations are important in neuroimaging, oncology, and cardiology, where accurate quantification of small regions is critical for measuring function or lesions.