Room: AAPM ePoster Library
To quantitatively evaluate the differences between planned and accumulated doses and the impact these variations have on the correlation with radiographic-based toxicity measurements.
150 locally-advanced non-small cell lung cancer cases treated on a prospective clinical trial with passive scattering proton therapy (PSPT) or intensity-modulated radiation therapy (IMRT) were planned on the average-intensity image (3D planned dose), and received weekly 4DCTs. Hybrid intensity-based deformable image registration was performed, through which the 4D planned dose (on planning 4DCT) and 5D accumulated dose (over all weekly 4DCTs) were calculated. Clinical metrics were compared between the 3D/4D-planned/5D-accumulated doses for each modality, including mean lung dose (MLD), lung V20, mean heart dose (MHD), heart V70, mean esophagus dose (MED), and esophagus D33. The overlap of isodose intervals with 5Gy increments that are inside lung and outside GTV were compared between 3D-planned/5D-accumulated cases using Dice similarity coefficient (DSC). Isodose intervals for 3D-planned/5D-accumulated were also mapped onto the follow-up CT (~6-month post-treatment) to evaluate the dosimetric impact on toxicity assessment (percent change in mean HU).
11 cases (6 photon/5 proton) have been analyzed to date. No substantial difference was seen in clinical metrics for IMRT. However, PSPT cases that exceeded ~10Gy difference in MLD was seen for one case, ~5Gy difference in MED for 4 cases and 20Gy difference in esophagus D33 for 3 cases. Average DSC of isodose intervals between 3D-planned/5D-accumulated doses was 0.58 and 0.21 for IMRT and PSPT, respectively. Differences of up to 26% were observed in mean HU-change from planning to follow-up between 3D-planned/5D-accumulated doses.
Although the average differences in clinical metrics were small for photons, individual cases had substantial differences for protons. Substantial differences between the percent HU-change from planning to follow-up for dose intervals between 3D-planned/5D-accumulated indicate the potential impact of dose accumulation on toxicity correlations.
Funding Support, Disclosures, and Conflict of Interest: This research is funded in part by RaySearch Laboratories and the Helen Black Image-Guided Fund.