Uncertainties of 4-dimensional computed tomography-based tumor motion measurement for lung stereotactic body radiation therapy

Purpose: To evaluate how well tumor motion measured prior to treatment based on 4-dimensional computer tomography (4DCT) reflects actual tumor motion during beam-on throughout the course of treatment. Methods and Materials: Twenty-three patients who had lung stereotactic body radiation therapy (SBRT) treatments were retrospectively selected. All patients had 4DCT simulation for treatment planning, from which tumor motion ranges were measured (R4DCT). Tumor motion was monitored during treatment using megavoltage (MV) imaging. Tumor motion trajectories were extracted from cine MV images and were used to determine mean and maximum tumor motion range (Mean RMV, Max RMV) throughout entire course of treatment. Comparison and correlations between mean and max RMVand R4DCTwere calculated. Results: On average, an insignificant difference was found between mean RMVand R4DCT(P = .67, mean [±SD] difference = -0.7 [±1.6] mm); meanwhile a significant difference was found between Max RMVand R4DCT(P = .03, mean [± SD] difference = 1.9 [±1.6] mm). The difference between RMVand R4DCTwas found inversely proportional to R4DCT(Y = -0.4X + 0.6, r = 0.76). Max RMVwas greater than R4DCTin all patients; difference between the 2 showed no correlation with R4DCT(Y = -0.02X + 1.9, r = 0.05). Correlation between Mean RMVand R4DCTand between Max RMVand R4DCTcan be expressed as Y = 0.7X (r = 0.88) and Y = 0.8X (r = 0.50), respectively. The same analysis performed on tumors that moved less than 5 mm from 4DCT revealed the following correlations: Y = 1.3X (r = 0.83) and Y = 1.7X (r = 0.49). Conclusions: Tumor motion measured from 4DCT approximates the overall average tumor motion range, but consistently underestimates the overall maximum tumor motion range. These findings may lead to a potential strategy for managing uncertainties of 4DCT in the application of lung SBRT.