Factors Associated With In-field Recurrence in Patients Treated With Liver SBRT

Abstract

Stereotactic body radiation therapy (SBRT) is an emerging technique for treatment of primary and secondary liver malignancies. Reports in the literature of the impact of set up uncertainties and other factors influencing in-field recurrences are lacking. The aim of this study was to examine the magnitude and consistency of interfractional shifts, their association, and association of other factors with in-field recurrences in patients treated with liver SBRT. All patients treated with liver SBRT at a single institution between 2014 and 2016 with doses 40 – 50 Gy in 5 fractions were included. Tumor histologies included hepatocellular carcinoma (50%), cholangiocarcinoma (4%), liver metastases from colorectal (39%) and other cancers (7%). Median follow-up was 15 months. A bony match was performed using orthogonal kV images. A soft-tissue match was then performed using a cone beam CT (CBCT). The shift data included in this study are the shifts made after the CBCT. Respiratory motion management techniques included abdominal compression, respiratory gating or free breathing, with the target respiratory motion below 1 cm. No dedicated internal fiducial markers were used. Retrospective analysis of prospectively collected data was performed as a part of an institutional Quality Assurance process. A total of 46 patients met the eligibility criteria. Data on interfractional shifts were available for 22 patients. One- and two-year in-field control rates were 64% and 30%, respectively. Mean vector shift length was 0.37 cm (systematic error 0.14 cm, random error 0.29 cm), 86% of patients had no shifts greater than 1 cm, 96% of all the shifts performed were less than 1 cm. Inferior tumor location had lower interfractional shift values compared with other tumor locations (one sided t-test=1.81, p-value=0.047). Despite that, inferior tumor location was weakly associated with in-field recurrence on univariate analysis (HR=2.63, p-value=0.093; 95% CI 0.86, 8.06), with median time to in-field recurrence 11 months as opposed to 21 months in patients with other tumor locations. Liver metastasis histology was also weakly associated with in-field recurrence, compared with primary liver histologies (HR=1.63, p-value=0.087; 95% CI 0.93, 2.87). Radiation dose, motion management technique, planning target volume and interfractional shift magnitude were not associated with in-field recurrence. Set up uncertainty and reproducibility was acceptable and did not impact in-field recurrences. Cone beam CT with soft tissue matching appears to be an appropriate set-up verification method. Inferior tumor location was associated with higher in-field recurrences in this patient cohort, likely due to proximity of critical structures, such as small bowel, resulting in target volume coverage compromises. Detailed evaluation of dose distribution in patients with in-field recurrences is underway.