OC-0378 HELICAL TOMOTHERAPY DYNAMICS ARE NOT AN ISSUE TO TREAT LUNG TUMORS FOR PATIENTS COACHED TO ENSURE REGULAR BREATHING

Abstract

Listen

Translate article

Purpose/objective: To evaluate helical tomotherapy (HT) treatment delivery and breathing motion interplay effect on dose distributions using a comprehensive 4D Monte Carlo (MC) dose engine including non-rigid deformation of dose maps computed on a 4D CT scan. Material/methods: Treatment planning is usually performed on a 3DCT image, even though lung tumors may move during delivery. To mitigate this issue, 4DCT-based delineation and appropriate margin definition are used to ensure good tumor coverage. However, treatment beam and breathing-induced motions interplay may lead to clinically unacceptable delivered doses, irrespective of the margin definition technique used. The particular delivery scheme of HT, with synchronous motion of both the gantry and the treatment couch, raise concerns specific to this modality for treating lung tumors. To evaluate the effect of motion interplay on dose distributions, a 4D MC dose computation engine was devised: 1) the MC core was TomoPen, a previously validated MC model of tomotherapy; 2) dose maps were computed on a 4D CT scan (10 phases); 3) resulting dose maps were accumulated through non-rigid deformation. Regular breathing was ensured with patient audio coaching. The treatment sinogram was correlated to the measured breathing period. 4D dose distributions (“interplay simulated”) were computed for 7 patients with a large range of motion amplitudes (up to 11.4 mm in the superior-inferior direction). Those were compared to MC dose distributions calculated on the 3DCT (“planned” dose distributions) and also on the 4DCT with the entire sinogram computed on every single phase, thus accounting only for breathing motion assuming an infinitely slow treatment (“no interplay”). Generalized equivalent uniform doses (gEUDs) and typical DVHs metrics (D95, D2, V95, Dmean, …) were computed for the tumor volumes (CTV and GTV) and compared for all simulation modalities. Results: For all modalities and all 7 patients, dose distributions complied with local clinical requirements (reported as in ICRU 83, RTOG 0618 and 0236 depending on the cases studied). Between “interplay simulated” and “no interplay”, D95 and D2 for tumor volumes were within 2.2% whereas Dmean and gEUDs were within 1% . The maximum difference between planned dose and “interplay simulated” was a 3% increase in Dmean and gEUD (see figure (b)). Since the “no interplay” effect was in agreement within 0.1 % with “interplay simulated”, the difference was attributed to the different quality of images of the 3D and 4D CT data sets. Conclusions: For the patients included in this study, coached to ensure regular breathing, HT delivery and breathing motion interplay was found clinically negligible, confirming in practical clinical routine previous theoretical analysis. Thus, as with 3D conformal static delivery, the problem of covering moving lung tumors may be restricted to the sole definition of margins if gating or tracking technologies are not available.