Infographics / CT Physics

Partial volume effect

Left: an enlarged voxel split diagonally between tissue A (40 HU) and tissue B (240 HU); an arrow leads to a single voxel showing 140 HU, the volume-weighted average. Right: a line profile across a tissue boundary where the true sharp step is replaced by a smoothed ramp, showing edge blurring. Bottom: two side-view voxels showing that a fixed structure fills a larger fraction of a thin slice than of a thick slice, so it is less diluted with thinner collimation.

Partial volume effect in CT: a voxel containing two tissues is displayed as their averaged CT number Left: an enlarged voxel split diagonally between tissue A (40 HU) and tissue B (240 HU); an arrow leads to a single voxel showing 140 HU, the volume-weighted average. Right: a line profile across a tissue boundary where the true sharp step is replaced by a smoothed ramp, showing edge blurring. Bottom: two side-view voxels showing that a fixed structure fills a larger fraction of a thin slice than of a thick slice, so it is less diluted with thinner collimation. Partial volume effect: each voxel shows the average of the tissues inside it one voxel straddling two tissues tissue A 40 HU tissue B 240 HU average 140 HU one value HU = f₁·40 + f₂·240 = ½(40)+½(240) = 140 a sharp boundary is blurred: edge HU ramps instead of stepping 240 40 HU position across boundary blurred edge true object measured (CT) Why slice thickness matters (z-axis depth of the voxel) a fixed structure fills a larger fraction of a thin voxel, so its averaged HU is less diluted thick slice structure fills ~¼ of voxel voxel HU ≈ 90 (heavily diluted) thin slice structure fills ~⅗ of voxel voxel HU ≈ 160 (closer to true 240 → sharper) Thicker slices → more averaging (worse low-contrast); thinner collimation → less averaging, sharper edges.

Practice the Physics paper

FRCRX is a Part 1 Physics question bank: true/false items with explanations, timed mocks and topic analytics.