Reproducibility of tumor motion probability distribution function in stereotactic body radiation therapy of lung cancer.
PURPOSE: To evaluate the reproducibility of tumor motion probability distribution function (PDF) in stereotactic body radiation therapy (SBRT) of lung cancer using cine megavoltage (MV) images. METHODS AND MATERIALS: Cine MV images of 20 patients acquired during three-dimensional conformal (6-11 beams) SBRT treatments were retrospectively analyzed to extract tumor motion trajectories. For each patient, tumor motion PDFs were generated per fraction (PDF(n)) using three selected "usable" beams. Patients without at least three usable beams were excluded from the study. Fractional PDF reproducibility (R(n)) was calculated as the Dice similarity coefficient between PDF(n) to a "ground-truth" PDF (PDF(g)), which was generated using the selected beams of all fractions. The mean of R(n), labeled as R(m), was calculated for each patient and correlated to the patient's mean tumor motion rang (A(m)). Change of R(m) during the course of SBRT treatments was also evaluated. Intra- and intersubject coefficient of variation (CV) of R(m) and A(m) were determined. RESULTS: Thirteen patients had at least three usable beams and were analyzed. The mean of R(m) was 0.87 (range, 0.84-0.95). The mean of A(m) was 3.18 mm (range, 0.46-7.80 mm). R(m) was found to decrease as A(m) increases following an equation of R(m) = 0.17e(-0.9Am) + 0.84. R(m) also decreased slightly throughout the course of treatments. Intersubject CV of R(m) (0.05) was comparable to intrasubject CV of R(m) (range, 0.02-0.09); intersubject CV of A(m) (0.73) was significantly greater than intrasubject CV of A(m) (range, 0.09-0.24). CONCLUSIONS: Tumor motion PDF can be determined using cine MV images acquired during the treatments. The reproducibility of lung tumor motion PDF decreased exponentially as the tumor motion range increased and decreased slightly throughout the course of the treatments.
Zhang, F; Hu, J; Kelsey, CR; Yoo, D; Yin, F-F; Cai, J
Volume / Issue
Start / End Page
Pubmed Central ID
Electronic International Standard Serial Number (EISSN)
Digital Object Identifier (DOI)