Phase II Trial of Neoadjuvant Chemotherapy and Stereotactic Body Radiation Therapy (SBRT) for Resectable and Borderline Resectable (BLR) Pancreas Cancer

Abstract

There is emerging interest in the role of SBRT in locally advanced pancreas cancer, however little prospective data exists examining the safety, efficacy, and optimal target volumes for SBRT in the neoadjuvant setting for resectable or BLR pancreatic cancer. The durability of local control and failure patterns following SBRT in the neoadjuvant setting are not well described. A phase II prospective trial was conducted in resectable and BLR pancreatic patients. Eighteen patients were enrolled from 11/2014-6/2017. All patients received 3 cycles of chemotherapy and restaging prior to radiation. SBRT was delivered to the tumor and abutting vessel with fiducials/compression and a 3 mm PTV margin to 33 Gy (6.6 Gyx5fxn) with an optional elective PTV to 25 Gy (5 Gyx5fxn) customized to the nodal space and mesenteric vessels. Patients without progression underwent surgery 4-6 weeks following SBRT. The primary endpoint is ≥ Grade 3 acute and late GI toxicity. Secondary endpoints included overall survival (OS), progression-free survival (PFS), and cumulative incidence of local failure (LF). LF is defined as recurrence within conventional RT volumes from the time of resection to local failure or last CT with no progression. Local failures were fused to planning CTs for dose quantification. Thirteen patients had BLR tumors due to arterial abutment (n = 7) or SMV encasement (n = 6); 3 patients had resectable tumors. All patients received 4 months of gemcitabine/nab-paclitaxel (n = 13) or FOLFIRNOX (n = 5) prior to SBRT. There were no ≥ Grade 3 acute or late GI events. Following completion of SBRT, 6 patients progressed (33%) at restaging or surgery with liver metastases (n=5, 28%) or local disease (n=1, 5%). Surgery was performed in 12 patients (67%) with 11 (92%) R0 resections (median margin distance 1.75 mm [0-9mm]). Lymph nodes were involved in 4 (33%) patients. Median OS was 21 months for all patients. In patients undergoing resection median OS was 31 months compared to 9 months without resection (p=0.003). The median PFS was 11 months. Progression occurred in 75% (9/12) of resected patients with first site of failure as distant (n = 4, 44%), local only (n = 4, 44%), and local and distant (n = 1,11%). The cumulative incidence of LF at 12 months from resection was 56% with one patient surviving 15 months without LF. All LF were outside to the PTV33 with median D90 of 11.5 Gy (4-25 Gy), V25 Gy of 51% (0-90%), and V33 Gy of 45% (0-52%). In resected patients, LF occurred in 4/8(50%) patients treated with a 25 Gy elective PTV and in 4/5(80%) of patients treated to the primary only (p=0.56). SBRT as a component of neoadjuvant therapy was well tolerated. However, local failures were predominantly observed outside the PTV33 volume within conventional RT volumes. Therefore, the durability of local control after SBRT in the neoadjuvant setting using target volumes and dose as described merits close examination relative to chemoradiation prior to incorporation into routine practice.