Jose Carlos Pichardo
I was able to get in touch with an Elekta physicist who confirmed my suspicions. The issue with MONACO and the poor quality surface and ROIs has to do with the fact that MONACO alters the CT upon import. The CT is revoxelized to allow for faster calculations, so voxels are not equal in size. Since contours in MONACO are generated from the revoxelized CT, all contours match the CT voxel architecture. When the patient surface is exported to AlignRT, the software created triangles to generate the surface, and the triangles are not all the same size. Consequently, the ROIs look jagged and non-continuous.
Unfortunately the only workaround at the present time is to generate the patient contour in a third-party software e.g., Eclipse, and then export the structure into the MONACO plan structure set. When the plan is exported to AlignRT, you can then choose the smooth patient surface, rather than the MONACO one. This results in a very smooth surface and ROIs.
All of that being said, I tested monitoring our pelvis patients with the smooth and non-smooth surfaces, and I did not notice any difference in the accuracy of ROIs or AlignRT’s ability to monitor the patient.