Dosimetric and motion analysis of margin-intensive therapy by stereotactic ablative radiotherapy for resectable pancreatic cancer

BACKGROUND The retroperitoneal margin is a common site of positive surgical margins in patients with resectable pancreatic cancer. Preoperative margin-intensive therapy (MIT) involves delivery of a single high dose of ablative radiotherapy (30 Gy) focused on this surgically inaccessible margin, utilizing stereotactic techniques in an effort to reduce local failure following surgery. In this study, we investigated the motion of regional organs at risk (OAR) utilizing 4DCT, evaluated the dosimetric effects of abdominal compression (AC) to reduce regional motion, and compared various planning techniques to optimize MIT. METHODS 10 patients were evaluated with 4DCT scans. All 10 patients had scans using AC and seven of the 10 patients had scans both with and without AC. The peak respiratory abdominal organ and major vessel centroid excursion was measured. A "sub-GTV" region was defined by a radiation oncologist and surgical oncologist encompassing the retroperitoneal margin typically lateral and posterior to the superior mesenteric artery (SMA), and a 3-5 mm margin was added to constitute the PTV. Identical 3D non-coplanar SABR (3DSABR) plans were designed for the average compression and non-compression scans. Compression scans were planned with 3DSABR, coplanar IMRT (IMRT), and Cyberknife (CK) planning techniques. Dose volume analysis was undertaken for various endpoints, comparing OAR doses with and without AC and for different planning methods. RESULTS The mean PTV size was 20.2 cm3. Regional vessel motion of the SMA, celiac trunk, and renal vessels was small (< 5 mm) and not significantly impacted by AC. Mean pancreatic motion was > 5 mm, so AC has been used in all patients enrolled thus far. AC did not significantly increase OAR dose including the stomach and traverse colon. There were several statistically significant differences in the doses to OARs as a function of the type of planning modality used. CONCLUSIONS AC does not significantly reduce the limited motion of structures in close proximity to the MIT target and does not significantly increase the dose to OARs that can be displaced by the compression plate. The treatment planning techniques evaluated in this study have different advantages with no clearly superior method in our analysis. Dose to adjacent vessels may be reduced with 3DSABR or IMRT techniques, while conformality is increased with IMRT or CK.