Room: AAPM ePoster Library
Purpose: Planning dose-constraints generally are derived from old dose-effect data, when treatment accuracy was less. We developed a model that corrects old dose constraints for application in new radiotherapy techniques.
Methods: Planning data used in dose-effect studies are estimates of the true OAR doses; the real DVH curves are spread around the planned curves, depending on the treatment accuracy at the time. We assume that complications in these dose-effect studies came from those patients with the worst DVHs. Now, assume a dose constraint D(Vcrit)
with F?¹(x) the probit function. In case SD(Dcrit,new)=0, we get the “true” complication threshold dose.
We applied this to Quantec dose-constraints. To get an estimate of SD(Dcrit) values, we simulated (Monte-Carlo) systematic geometric errors (S) up to 4mm (1SD) in current treatment plans and calculated DVHs. The spread (1SD) in various Dcrit (Dmax, Dmean, D50%, D25%, D10%, D5%, D1%) was calculated for relevant OARs (i.e. Dmax>70%*Dprescribed). Accuracy during Quantec (20yrs ago) was assumed to be S=4mm vs S=2mm nowadays.
Results: Included were 71 treatment plans. At S=4mm, the average SD for the various Dcrit were respectively (see above) 4.0%, 4.7%, 4.5%, 6.4%, 7.6%, 7.3%, and 6.5%. For S=2mm SD these values are 2.8%, 2.5%, 2.3%, 3.5%, 4.4%, 4.4%, and 4.0%, so on average 2.4% less. Using this on 22 Quantec planning dose constraints, we found that True constraints are on average 2.4Gy higher, while New constraints are on average 1.1Gy higher.
Conclusion: Our new method determines true and new dose planning constraints from existing older constraints. It shows that with increasing radiotherapy accuracy, higher dose constraints yield the same NTCP. The method can also be applied when evaluating measured DVHs.
Funding Support, Disclosures, and Conflict of Interest: This work has been supported by a grant from Varian Medical Systems
Not Applicable / None Entered.