Abstract
BACKGROUND AND PURPOSE: A recent study of NSCLC patients showed small residual shifts of the high dose region towards/away from the heart after image-guidance were significantly related to overall survival. This study investigates whether the effect is observed in a SABR cohort, who have significantly different baseline outlook and are treated using an imaging protocol matching on the tumour rather than bony-anatomy alone. MATERIALS AND METHODS: 136 NSCLC patients treated with SABR were studied. The mean baseline shift of the tumour in the direction of the heart over the course of treatment was determined for each patient and used to categorise patients into risk groups. Kaplan-Meier survival curves were plotted and multivariable analysis performed to assess significance of the vector shift to the heart alongside common clinical variables. RESULTS: The vector shift to the heart was independent of all tested clinical variables. A significant difference was seen in patient survival, with patients with shifts towards the heart having significantly worse prognosis as compared to patients with shifts away. Multivariable analysis found a hazard ratio of 1.262 per mm (p = 0.013) for the vector shift to the heart, i.e. for every 1 mm shift of the high dose region towards the heart there is a 1.262 higher chance of death. CONCLUSIONS: Baseline shifts towards the heart significantly correlate with overall survival in a cohort of NSCLC SABR patients, with increased risk with increasing shifts towards the heart. These results provide further evidence for the use of stricter heart dose planning constraints for thoracic radiotherapy and suggest a heart planning organ at risk volume may be required for SABR treatments to account for baseline shifts.