Strict Constraints on the Most Important OAR
Dealing with OARs very close to the PTV (or even overlapping the PTV) is a challenging planning task. Often it is not possible to fulfill both the desired PTV coverage volume wishes and the OAR constraints. Instead, the best compromise between these two contradicting goals has to be found.
Sometimes the OAR constraints have to be relaxed (with side effects that might be tolerable in certain situations). Within the software, this can for example be achieved by setting the OAR constraint to Off or Smart instead of Strict.
The alternative is to relax the PTV coverage volume. Under most circumstances, the optimizer will guarantee a certain coverage (given by the Desired Coverage Volume ①). However, as described above, the optimizer will slowly reduce the coverage down to a minimum of Tolerated Coverage Volume ② if any constraints on the Most Important OAR are set to Strict and nevertheless cannot be fulfilled. How much the coverage is reduced depends on the amount of OAR constraint violation and on the position of the Weighting slider (Target side: no reduction in coverage, OAR side: maximum reduction, i.e., to Tolerated Coverage Volume if OAR constraints cannot be fulfilled otherwise).
The coverage is not simply reduced for the whole PTV. Instead, location based information is used to ensure the coverage is “sacrificed” at the right location (e.g., close to the OAR). This is done by utilizing a series of so called “halo” objects.
A “halo” is a virtual planning object (invisible in the software) that is automatically created by expanding the Most Important OAR ③ uniformly into all directions. For example a “halo” could be created for a radius of 1, 2 or 3 mm ④. This means that the “halo” object is 1, 2 or 3 mm larger than the OAR.
The actual halo is deduced from the amount of “sacrificed” coverage in the following way. First, the amount of PTV volume ① “sacrificed” is calculated. For example, if only 96% instead of 98% coverage shall be achieved, effectively 2% of the volume is “sacrificed”. Next, a halo object is created that covers exactly the amount of “sacrificed” PTV volume (e.g., 2%). This means the halo will be bigger for larger PTV volume reductions. The PTV voxels covered by the halo ② will then be marked as “to be sacrificed” for the optimization. These voxels do not have to fulfill the desired PTV dose.