OBJECTIVES: Optimal radiation target volumes for neoadjuvant therapy in patients with borderline resectable pancreatic cancer (BRPCa) are undefined. Most local recurrences are near the celiac axis and superior mesenteric artery. Methods for generating radiation target volumes include symmetric expansion around the tumor or a customized vascular based approach. We investigated 3 current prospective trials' coverage of vascular regions at increased risk of recurrence by comparing them to 2 reference standards. MATERIALS AND METHODS: Fourteen computed tomography simulation scans from an institutional prospective trial on BRPCa were used to replicate distinct volumes corresponding to each of 3 contemporary BRPCa trials. Trial volumes were compared with 2 reference volumes (vascular planning target volume and Hopkins planning target volume), which were both based on vascular regions at increased risk of recurrence. Boolean operators and DICE analyses were performed to evaluate trial volume coverage of reference standards. RESULTS: A total of 42 target volumes and 28 reference volumes were created using the 14-patient data set. DICE coefficients were highly variable ranging from 0.11 to 0.99. Mean % coverage of reference volumes ranged from 5.8% to 98.6%. CONCLUSIONS: The wide range of DICE coefficients and coverage indicate heterogeneity in high risk vascular target coverage using symmetric Boolean expansions from the primary tumor. This approach may inadequately cover regions at high risk of local recurrence in BRPCa. A customized clinical target volume that specifically includes the superior mesenteric artery and celiac axis will improve coverage to this region and will account for individual and tumor variability.
OBJECTIVES: Optimal radiation target volumes for neoadjuvant therapy in patients with borderline resectable pancreatic cancer (BRPCa) are undefined. Most local recurrences are near the celiac axis and superior mesenteric artery. Methods for generating radiation target volumes include symmetric expansion around the tumor or a customized vascular based approach. We investigated 3 current prospective trials' coverage of vascular regions at increased risk of recurrence by comparing them to 2 reference standards. MATERIALS AND METHODS: Fourteen computed tomography simulation scans from an institutional prospective trial on BRPCa were used to replicate distinct volumes corresponding to each of 3 contemporary BRPCa trials. Trial volumes were compared with 2 reference volumes (vascular planning target volume and Hopkins planning target volume), which were both based on vascular regions at increased risk of recurrence. Boolean operators and DICE analyses were performed to evaluate trial volume coverage of reference standards. RESULTS: A total of 42 target volumes and 28 reference volumes were created using the 14-patient data set. DICE coefficients were highly variable ranging from 0.11 to 0.99. Mean % coverage of reference volumes ranged from 5.8% to 98.6%. CONCLUSIONS: The wide range of DICE coefficients and coverage indicate heterogeneity in high risk vascular target coverage using symmetric Boolean expansions from the primary tumor. This approach may inadequately cover regions at high risk of local recurrence in BRPCa. A customized clinical target volume that specifically includes the superior mesenteric artery and celiac axis will improve coverage to this region and will account for individual and tumor variability.