Dosimetric consequences of misalignment and realignment in prostate 3DCRT using intramodality ultrasound image guidance

Purpose: It is common practice to correct for interfraction motion by shifting the patient from reference skin marks to better align the internal target at the linear accelerator's isocenter. Shifting the patient away from skin mark alignment causes the radiation beams to pass through a patient geometry different from that planned. Yet, dose calculations on the new geometry are not commonly performed. The intention of this work was to compare the dosimetric consequences of treating the patient with and without setup correction for the common clinical scenario of prostate interfraction motion. Methods: In order to account for prostate motion, 32 patients initially aligned to the room lasers via skin marks were realigned under the treatment beams by shifting the treatment couch based on ultrasound image guidance. An intramodality 3D ultrasound image guidance system was used to determine the setup correction, so that errors stemming from different tissue representations on different imaging modalities were eliminated. Two scenarios were compared to the reference static treatment plan: (1) Uncorrected patient alignment and (2) corrected patient alignment. Prostate displacement statistics and the dose to the clinical target volume (CTV), bladder, and rectum are reported. Monte Carlo dose calculation methods were employed. Results: Comparing the uncorrected and corrected scenarios using the static treatment plan as the reference, the average percent difference in D-95 for the CTV improved from -5.1% (range -40%, 1.3%) to 0.0% (-3.5%, 2.0%) and the average percent difference in V-90 for the bladder and rectum changed from -11% (-84%, 232%) to -8.3% (-61%, 5.2%) and from -47% (-100%, 108%) to 0.9% (-62%, 102%), respectively. There was no simple correlation between displacement and dose discrepancy before correction. After patient realignment, the prescribed dose to the CTV was achieved within 1% for 75% (24/32) of the patients. After patient realignment, 50% of the patients had doses that differed from the static treatment plan by 25% for the bladder and 8% for the rectum. Conclusions: The dose degradation due to prostate motion (before correction) is not accurately predicted from the average trends for all patients. Outliers included smaller displacements that lead to larger dosimetric differences in the corrected scenario, especially for the bladder and rectum, which exhibited doses substantially different from that planned.