INTRODUCTION: The purpose of this study is to assess the impact of a vacuum immobilisation system on reproducibility of patient set-up, interfraction stability and tumour motion amplitude. METHODS: From February 2010 to February 2012 as part of a prospective clinical trial 12 patients with solitary pulmonary metastases had consecutive four-dimensional computed tomography (4DCT) scans performed with and without vacuum immobilisation. The displacement of the tumour centroid position was recorded in each of the 10 phases of the 4DCT reconstruction. A further six patients with seven metastases underwent single fraction stereotactic ablative body radiotherapy (SABR) during this period (a total of 19 targets) and were included in an analysis of positional reproducibility and intrafraction immobilisation. Couch shifts recorded in the medio-lateral (X), cranio-caudal (Y) and ventero-dorsal (Z) planes. RESULTS: For the 19 treatments delivered, the median (0-90% range) shift required immediately pretreatment was 1 mm (0-3) in the X-plane, 2 mm (0-6) in the Y-plane and 4 mm (0-8) in the Z-plane, respectively. The mean (+/- standard deviation) of mid-treatment shifts were 0.3 mm (+/- 0.7), 1.1 mm (+/- 2) and 0.8 mm (+/- 1.5) in the X, Y and Z planes, respectively. Mid-treatment shifts were <2 mm in all directions (P = < 0.001). The length of treatment time correlated to the required shifts in the Z plane (r(2) = 0.377, P = 0.005), but not in the X or Y planes (P = 0.198 and P = 0.653, respectively). In the subset of 12 patients who had two 4DCTs, the median (range) amplitude of tumour displacements in the X, Y and Z planes when immobilised were 0.9 mm (0.3-2.9), 2.6 mm (0.2-10.6) and 1.6 mm (0.5-5.5), respectively. Immobilisation reduced the volume of tumour displacement during respiration by a median of 52.6% (P = 0.021). CONCLUSIONS: Vacuum immobilisation reduces total tumour excursion, facilitates reproducible positioning and provides robust intrafractional immobilisation during SABR treatments for pulmonary metastases.