Clinically Useful Observations of Prostate Motion using the Calypso Four-dimensional Localization System

Abstract

Daily localization of the prostate has become a standard of care for external beam treatment (EBRT) of prostate cancer (PC). Also understood is the fact that the prostate gland has random motion in an individual during a EBRT session. No consensus exists as to the use of this information. Our objective was to observe inter- and intra-fraction motion of the prostate using the Calypso four-dimensional localization system (C) to guide CTV to PTV assumptions and adaptive treatment. This is an institutional review board-approved evaluation of 162 sessions of radiotherapy observed in 6 patients treated for PC using C. Each patient was setup to external markings daily and offsets detected by the C were recorded. Localization was confirmed by on-board imaging (OBI). Motion and time data was recorded 10 times each second throughout each treatment session. Nearly 2.5 million data points were used in this analysis. X = lateral motion, Y = anterior-posterior motion, and Z = superior-inferior motion. The paired t test method was used for statistical comparisons. The average daily offsets between external setup and C were 0.18 cm (range, 0.0-0.9), 0.16 cm (0.0-1.5), 0.26 cm (0.0-0.9) in the X, Y, and Z direction. Concordance was found between C and OBI localization (p = NS). Average motion for all sessions was 0.04 cm (range, 0.01-0.2) in X, 0.08 cm (0.01-0.28) in Y, and 0.09 cm (0.02-0.33) in Z directions, respectively. The average maximum motion for all sessions was 0.13 cm (range, 0.02-0.47) in X, 0.39 cm (0.07-5.05) in Y, and 0.42 cm (0.07-6.14) in Z directions, respectively. When considering margin, for all sessions, the time spent within 0.1 cm CTV to PTV expansion (X, Y, Z) is 96%, 70%, 69%; within 0.2cm is 100%, 91%, 91%; within 0.3 cm is 100%, 98%, 98%. For each patient, the average motion in the X, Y, and Z for the first 25% of sessions was not significantly different from the total remaining sessions in their treatment course for each patient. The Calypso four-dimensional localization system allowed the unique opportunity to observe significant, non-uniform intra-fraction motion of the prostate gland. X motion was less in magnitude than Y or Z. Based on 162 EBRT treatment sessions, a 0.1 cm CTV expansion in X dimension, and a 0.3 cm expansion in Y and Z dimensions would confine most random motion to within the prescribed dose envelope. In addition, the first sessions for each patient are predictive of motion for later sessions. This data has given us clinically useful information to help generalize prostate motion for treatment planning and future adaptive EBRT techniques. We have employed this data to design a clinical trial restricting motion and performing imaging when constraints are exceeded.