2.2

ROI precision

To determine the appropriate CT values for the Upper and Lower Autocontouring thresholds, select the Profile tool and drag the cursor across the structure to be autocontoured. A plot of CT values for each point in the profile is generated.

As you can see from the plot, the CT values for the lung fall between 90 and 450 and the CT values for the tissue around the lung are above 800. To autocontour the lung, you can select a lower threshold value between 450 and 800 and leave the upper threshold at its default value.

2.2

ROI precision ROIs have the same resolution as the data set to which they are assigned (typically the primary CT data set); thus, the ROI may have a certain coarseness. Because the ROI cannot be more precise than the data set, you can draw a tiny contour within a single data set voxel and the Colorwash display is just a square. The smallest possible true size for an ROI is 1 data set voxel. The software recognizes no difference among 1, 10, or any number of tiny contours drawn within a single voxel.

Colorwash and Poly modes are more accurate representations of the ROI volume than Contour display. Colorwash and Poly modes display the same thing, but in Poly mode the corners of the contours are rounded to display a smooth surface.

12

Regions of Interest

Pinnacle 3

Release 9

Philips Medical Systems

The Contour mode reflects what you draw in the software. In drawing a contour you may touch only a portion of a voxel. In Colorwash and Poly modes the whole voxel is considered part of the ROI, so the ROI may appear larger. The Colorwash and Poly modes reflect the internal representation of the ROI (how the software “sees” it), which is used for ROI statistics, volume calculations, blocks, and MLCs.

ROI expansion/contraction

2.3

The line is the ROI as it appears in Contour mode. The gray area is the ROI as it appears in Colorwash or Poly mode. The voxels with an “E” in them represent the edge voxels (which are treated differently than the inner voxels when determining ROI volume).

ROI volume and statistics are calculated using the internal representation of the ROI (represented by the Colorwash and Poly display modes). The formula is: ROI volume = voxel size x [number of inner voxels + (0.5 x number of edge voxels)] The edge voxels are multiplied by 0.5 because voxels at the edge of a ROI typically do not have their entire volume intersected by the contour. The average volume intersected at the edge of an ROI is 0.5 voxels.

2.3

ROI expansion/contraction This section provides information to supplement Expand and contract ROIs in the Regions of Interest chapter of the Pinnacle3 Planning, Classic Pinnacle3 Planning, and AcQSim3 Instructions for Use.

2.3.1

Evaluate ROI expansion/contraction results Certain expansions and contractions may produce shapes that are not what you expect. Remember that ROI expansion/contraction is • a 3D operation • applied over the entire source ROI, and • affected by image set resolution.

4535 604 45201 A

The following sections discuss each of these points. Expansion/contraction is a 3D operation

Philips Medical Systems

Use 3D views or a combination of transverse, sagittal, and coronal 2D views (with the ROIs in Poly or Colorwash mode) to determine the correctness of an expansion or contraction. There are two reasons why a single 2D contour view is insufficient: • Since the operation is 3D, parts of the source ROI that fall on other

slices affect the destination ROI. This can lead to misunderstandings about what you see in 2D views. For instance, when you expand or contract an ROI in a specific direction, it may seem as though the software has contracted or expanded the ROI in additional non-specified directions. In the Pinnacle 3

Release 9

Regions of Interest

13