ABSTRACT
OBJECTIVE: To evaluate interfractional and intrafractional movement of patients with rectal cancer during radiotherapy with electronic portal imaging device (EPID) and surface infrared (IR) markers. METHODS: 20 patients undergoing radiotherapy for rectal cancer with body mass index ranging from 18.5 to 30 were enrolled. Patients were placed in the prone position on a couch with a leg pillow. Three IR markers were put on the surface of each patient and traced by two stereo cameras during radiotherapy on a twice-weekly basis. Interfractional isocentre movement was obtained with EPID images on a weekly basis. Movement of the IR markers was analysed in correlation with the isocentre movement obtained from the EPID images. RESULTS: The maximum right-to-left (R-L) movement of the laterally located markers in the horizontal isocentre plane was correlated with isocentre translocation with statistical significance (p = 0.018 and 0.015, respectively). Movement of the surface markers was cyclical. For centrally located markers, the 95% confidence intervals for the average amplitude in the R-L, cranial-to-caudal (C-C) and anterior-to-posterior (A-P) directions were 0.86, 2.25 and 3.48 mm, respectively. In 10 patients, intrafractional movement exceeding 5 mm in at least one direction was observed. Time-dependent systematic movement of surface markers during treatment, which consisted of continuous movement towards the cranial direction and a sail back motion in the A-P direction, was also observed. CONCLUSION: Intrafractional movement of surface markers has both cyclic components and time-dependent systematic components. Marker deviations exceeding 5 mm were mainly seen in the A-P direction. Pre- or post-treatment EPID images may not provide adequate information regarding intrafractional movement because of systematic movement in the A-P direction during radiotherapy. ADVANCES IN KNOWLEDGE: This work uncovered a sail back motion of patients in the A-P direction during radiotherapy. Pre- or post-treatment EPID images may not provide accurate positioning of patients in the A-P direction because of this time-dependent intrafractional motion.
