Except the isodose lines aren't real. For a 2-arc Stage III lung I currently have on beam (treating with VMAT):
The MU calcs prior to starting:
View attachment 353744
The portal dosimetry after starting:
View attachment 353743
The QA for the other arc is a little worse but within acceptable limits. I'm sure there's day-to-day variation, but to what degree is anyone's guess. Tolerance of gamma pass rates vary by institution, but current TG recommendation is that pass rates should be ≥95% using 3%/2mm tolerance and 10% threshold
The final treatment plan I approved look great. The isodose lines fell where I wanted them to fall. However, those isodose lines were born from an algorithm using voxels from the AVG sequence of a 4D CTSIM. In order for those to be "real":
Patient needs to be set up exactly as they were at time of sim
Breathing cycle/pattern needs to be the exact same at time of sim
Linac needs to spit out radiation precisely as the algorithm calculated
The radiation needs to interact with matter within the patient precisely as the algorithm calculated
The probability of all that happening simultaneously is very low, which is why the PTV margin exists. The volume within the PTV margin represents the probability that your CTV receives the prescribed dose. Shaving the PTV margin to make the treatment plan look better on review is only reducing the probability that your CTV receives the desired dose.
Again, is it
clinically significant? Most of the time, probably not. But the isodose lines you're approving aren't real.