Interfraction variation in lung tumor position with abdominal compression during stereotactic body radiotherapy

To assess the effect of abdominal compression on the interfraction variation in tumor position in lung stereotactic body radiotherapy (SBRT) using cone-beam computed tomography (CBCT) in a larger series of patients with large tumor motion amplitude.

Thirty patients with lung tumor motion exceeding 8 mm who underwent SBRT were included in this study. After translational and rotational initial setup error was corrected based on bone anatomy, CBCT images were acquired for each fraction. The residual interfraction variation was defined as the difference between the centroid position of the visualized target in three dimensions derived from CBCT scans and those derived from averaged intensity projection images. The authors compared the magnitude of the interfraction variation in tumor position between patients treated with [n = 16 (76 fractions)] and without [n = 14 (76 fractions)] abdominal compression.

The mean ± standard deviation (SD) of the motion amplitude in the longitudinal direction before abdominal compression was 19.9 ± 7.3 (range, 10-40) mm and was significantly (p < 0.01) reduced to 12.4 ± 5.8 (range, 5-30) mm with compression. The greatest variance of the interfraction variation with abdominal compression was observed in the longitudinal direction, with a mean ± SD of 0.79 ± 3.05 mm, compared to -0.60 ± 2.10 mm without abdominal compression. The absolute values of the 95th percentile of the interfraction variation for one side in each direction were 3.97∕6.21 mm (posterior∕anterior), 4.16∕3.76 mm (caudal∕cranial), and 2.90∕2.32 mm (right∕left) without abdominal compression, and 2.14∕5.03 mm (posterior∕anterior), 3.93∕9.23 mm (caudal∕cranial), and 2.37∕5.45 mm (right∕left) with abdominal compression. An absolute interfraction variation greater than 5 mm was observed in six (9.2%) fractions without and 13 (17.1%) fractions with abdominal compression.

Abdominal compression was effective for reducing the amplitude of tumor motion. However, in most of the authors' patients, the use of abdominal compression seemed to increase the interfraction variation in tumor position, despite reducing lung tumor motion. The daily tumor position deviated more systematically from the tumor position in the planning CT scan in the lateral and longitudinal directions in patients treated with abdominal compression compared to those treated without compression. Therefore, target matching is required to correct or minimize the interfraction variation.