Type of fiducial marker and clinical outcomes in stereotactic body radiation therapy for pancreatic cancer.

414 Background: Stereotactic body radiation therapy (SBRT) for pancreatic cancers has been shown to improve local control, and is an important option for treatment, especially for unresectable disease. Verifying the tumor location prior to delivery of SBRT is challenging, so fiducial markers are used to track tumor location. There is currently no standard of which fiducials to use in treatment. This study would like to compare outcomes of patients treated with SBRT using different fiducial markers. Methods: Records of patients diagnosed with primary pancreas cancer who were treated with chemotherapy and SBRT were reviewed from 2006-2019. Patients were excluded if they were treated with Cyberknife, were metastatic at presentation, recurrence /persistent disease after Whipple/radiation therapy, were secondary metastatic disease (from another primary), and if they were resected after SBRT. Patients were categorized according to the fiducial used for tumor tracking during SBRT treatment: gold seeds, intrabiliary stent, or both. Cumulative incidence of local recurrence (CIR) was analyzed with death as competing event, and time to over-all survival was estimated using Kaplan-Meier curves. Results: A total of 129 patients with available fiducial information were included in this study, of which 64 (49.6%) were treated with SBRT using gold seeds, 23 (17.8%) using intrabiliary stent, and 42 (32.6%) using both the seeds and stent. There were no difference between groups in terms of baseline characteristics such as age (p = 0.169), sex (p = 0.293), and stage grouping (p = 0.293). Median follow-up time was 15 months (range: 0.3-37.3 months). The 6- and 12-month CIR were 1.5% (95%CI, 0.1%-7.4%) and 11% (95%CI, 4.8%-20.2%) for patients treated with seeds, 4.3% (95%CI, 0.2%-18.6%) and 30.4% (95%CI, 13.1%-49.8%) for patients treated with stent, and 4.8% (95%CI, 0.8%-14.6%) and 19.5% (95%CI, 9.0%-32.9%) for patients treated with both (p = 0.007). Median time to overall survival was 15.3 months (95%CI, 13-17.8 months) for patients treated with seeds, 21.3 months (95%CI, 14.7-29.6 months) for patients treated with stent, and 15.7 months (95%CI, 11.5-19.7 months) for patients treated with both (p = 0.307). Univariate analysis for predictors of local failure did not show significance for age (p = 0.812), or advanced stage (p = 0.483), but was significant for the presence of seeds (p = 0.006). Conclusions: The type of fiducial marker used for tracking during pancreas SBRT treatment was associated with local failure but no difference in overall survival. Further analysis is warranted to see which clinical factors contribute to this difference.