What Neoadjuvant Radiation Volumes Should be Covered in Pancreas Cancer? An Analysis of Recurrence Patterns among Tumors with Arterial vs. Venous Involvement

Purpose/Objective(s)

In Pancreatic ductal adenocarcinoma (PDAC) candidates for neoadjuvant therapy followed by surgical resection, it is not clear whether there is difference in failure outcome between arterial and venous involvement subgroups, and which group benefits from neoadjuvant stereotactic ablative radiation (NA SABR). We hypothesize that: 1-loco-regional failure (LRF) is higher for arterial vs venous involved PDAC, 2- NA SABR decreases LRF in arterial involved PDAC, 3- LRFs in patients who received neoadjuvant chemotherapy (NA CTH) only can be used to propose a NA SABR radiation planning volume.

Materials/Methods

We reviewed our institutional database for patients with the following criteria: 1- histopathological diagnosis of PDAC, 2- tumors with vessel involvement at baseline, and 3- received neoadjuvant therapy followed by surgical resection. Vessel involvement was stratified on baseline CT scan with IV contrast to either arterial (Celiac artery (CA), superior mesenteric artery (SMA), and common hepatic artery) or venous involvement (Portal and Superior mesenteric vein). LRF was defined as new lesion(s) following surgical resection in remnant pancreas, surgical bed, pancreatojejunostomy site, CA, SMA, and regional nodes. SABR was defined as 5 fraction radiation under image guidance. To address the hypothesis: 1- in NA CTH only cohort, we compared LRF events in arterial vs venous involvement using log rank test, 2- in patients with arterial involvement, we used log rank test to compare LRF events in NA SABR+NA CTH versus NA CTH only, 3- in patients who received NA CTH only, with arterial involvement and developed LRF, we did a soft tissue imaging fusion using CA and SMA of all failure scans to baseline CT scans. The distance along the SMA that was involved by LRFs was measured along the y axis.

Results

Fifty-seven patients met the inclusion criteria with a median follow up of 31 months. Thirty-one patients received NA CTH only, while 26 patients received NA SABR plus chemotherapy. Among the NA CTH only cohort (n=31), 7 and 24 patients presented with arterial and venous involvement respectively. Among the NA SABR cohort (n=26), 22 and 4 patients presented with arterial and venous involvement respectively. 1- among NA CTH only cohort (n=31), one-year LRF free was 35.7% in arterial involvement vs 75.5 % in venous involvement (HR= 7.7, 95 % CI 1.8-32.8, p value=0.0059), 2- among patients with arterial involvement (n=29), the one-year LRF was 73.8% in SABR treated compared to 35.7 % in non SABR treated (HR= 5.3, 95 % CI 1.3 to 21, p value=0.0187), 3- we propose low and high dose volumes from the NA CTH that we aim to validate independently. The volume of combined LRFs was 102.2 cc. The distance along the SMA to be covered during planning was 5.7 cm along the y axis. Median distance from tumor artery contact to the LRFs was 1.2 and 1.7 cm in the superior and inferior directions respectively.

Conclusion

NA SABR may decrease LRF in arterial involved PDAC. LRFs in non SABR treated cohort can guide radiation planning volumes.