Validation of the mid-position strategy for lung tumors in helical tomotherapy

Purpose: To validate the “mid-position” approach for lung tumor motion management in helical tomotherapy with 4D Monte Carlo planning simulation, in comparison with conventional ITV. Materials & Methods: 8 patients with stage I non-small cell lung cancer (NSCLC) treated by SBRT were included, as well as 6 patients with stage II-III NSCLC treated by Simultaneous Integrated Boost (SIB) and participating in a dose escalation protocol. Prior to treatment, a contrast-enhanced CT (CE-CT) and a 4DCT (for SBRT) or a combined 4D FDG-PET-CT (for SIB) were acquired. The GTV, CTV, and OARs were delineated on the CE-CT according to our clinical protocol. Next, 4D data were used to generate first the ITV and then the MidP volume in its exact time-weighted mean position of the respiratory motion, using a validated Morphon non-rigid registration algorithm. The PTVs were finally drawn according to the margin formula for geometric uncertainties developed by Van Herk et al. and adapted to the specific features of lung tumor tomotherapy. For each patient, two treatments were planned based on margins derived from the ITV and MidP volume. Volumetric and dosimetric parameters, as well as conformity indexes were compared with both strategies. Moreover, dose distributions were computed using a 4D Monte Carlo (MC) model, in order to assess the impact of intra-fraction tumor motion on tumor coverage (quantified by D95), with and without the interplay effect. Results: For SBRT and SIB patients, the PTVs defined with the ITV approach were on average 1.2 times larger than those derived from MidP. Consequently, the dose to all the OARs was on average lower when using the MidP. Nonetheless, the planned dose conformity to TVs was identical between both strategies (0,92 ± 0,03 and 0,84 ± 0,05 for DICE and Paddick indexes, respectively). For all SBRT patients, D95 to the GTV computed from 4D MC dose distributions complied within 1% of the planning recommendations when using the ITV approach. In contrast, MidP failed to ensure adequate GTV coverage in 3 patients. For one patient, the simulated interplay effect lowered the D95 to the GTV by 4.35% compared to the planned dose distribution (Fig a). Although the interplay effect did not affect the two other patients, simulated MC calculations demonstrated significant GTV underdosages, with D95 to GTV reduced by 2.16% and 2.61% compared to the planned doses (Fig b). 4D MC computations are ongoing for the SIB group. Conclusion: Compared to the ITV, the MidP strategy significantly reduced the PTV and irradiated volumes in all patients. However, if MidP could safely be applied in helical tomotherapy in most cases, it might lead to insufficient tumor coverage for very small tumors ( 10mm). Those particular cases are indeed questioning the fundamental hypothesis underlying the framework of the MidP concept. As SIB treatments are usually delivered to large and centrally located tumors, MidP might allow for a safe reduction of the PTVs and irradiated tissues, although this still needs to be further confirmed