Session Item

Intrafraction respiratory motion is a concern for lung tumor radiotherapy, but full evaluation of its impact is hampered by the lack of images representing the full spectrum of motion states. Typically it is estimated with the same 4DCT images used for treatment planning, normally with 10 pseudo-respiratory phases captured over a short time with only a few breathing cycles. It may therefore misrepresent both the motion amplitude and possible irregular breathing patterns. This study presents a novel evaluation using extensive free-breathing image data acquired over realistic clinical treatment times to study the impact of respiratory motion on the robustness of radiotherapy plans with special focus on the planning and prescription approaches employed.

Cine-CT images from 14 lung cancer patients acquired during 8 minutes of free-breathing taken at three different occasions were used. The 300 images per patient were registered to a reference set to enable dose tracking over a time period corresponding to a full treatment, (Figure 1). Photon plans were created using four different methods aiming to deliver a dose of 54 Gy in three fractions to D50% of the target: robust CTV planning based on 10-phase 4DCT, homogeneous fluence to an ITV, an isodose-based prescription to the ITV with a high central dose concurrently normalized to the D50% of the CTV. All treatment plans were recalculated on the extensive 300-image sets and the resulting doses summed for the CTV. Target coverage was evaluated in terms of the dose coverage probability over the whole set of patients. The resulting differences in dose homogeneity and coverage probability between the tested planning methods were evaluated by the Wilcoxon signed-rank test.

The extensive evaluation approach showed that robust planning leads to achieving the intended dose with only a minor risk for underdosage in half of the patients. In contrast, respiratory motion reduces the probability of target coverage for the ITV fluence planning approach (p-value<0.001), as illustrated by the larger variance in dose coverage (Figure 2). No significant difference in dose coverage probability at D50% was found between the robust and the two isodose based plans, however robust planning was found to have a higher probability for a more homogeneous tumor dose (p-value<0.001).

The results show that tumor motion not covered by the 4DCT gives a risk of underdosage but that this effect is mostly minor for robust optimization and isodose-based planning, but with higher variance for the latter. The robustness of any planning method against irregular tumor motion needs to be thoroughly evaluated before clinical implementation. The novel approach to study the impact of irregular motion will also be applied for proton therapy.