Daily stereotactic ultrasound prostate targeting: inter-user variability

We analyzed the inter-user variability of patient setup for prostate radiotherapy using a stereotactic ultrasound-targeting device. Setup variations in 20 prostate cancer patients were analyzed. Users were a radiation oncologist, a medical physicist, four radiation technologists (RTT) and a radiologist. The radiation oncologist, radiologist, physicist and two RTTs were experienced users of the system (>18 months of experience); two RTTs were users new to the system. Gold standard for this analysis was a control CT acquired immediately following ultrasound targeting. For inter-user variability assessments, the radiation oncologist provided a set of axial and sagittal freeze-frames (standard freeze-frames) for virtual targeting by all users. Additionally each user acquired individual freeze-frames for target alignments. We analyzed the range of virtual setups in each patient along the principal room axes based on standard and individual freeze-frames. The magnitude of residual setup error and percentage of setup change for each user was assessed by control CT/planning CT comparison with individual virtual shifts. A total of 184 alignments were analyzed. The range of virtual shifts between users was 2.7+/-1.4, 3.6+/-1.1, and 4.4+/-1.4 mm (mean+/-SD) in x, y and z-direction for setups based on standard freeze-frames and 3.9+/-2.6, 6.0+/-4.7, and 5.4+/-2.7 mm for setups based on individual freeze-frames. When only virtual shifts of experienced users were analyzed, the mean ranges were reduced by up to 2.4 mm. Average magnitude of initial setup error before ultrasound targeting was 14.3 mm. Average improvement of prostate setup was 63.1+/-23.4% in experienced and 35.14+/-37.7% in inexperienced users, respectively (p<0.0001). Only 5 of 184 (2.7%) virtual alignments would have introduced new larger setup errors (mean 3.2 mm, range 0.2 to 9.5 mm) than the magnitude of the initial setup error. We conclude that ultrasound guided treatment setup for patients treated for prostate cancer can be performed with high inter-user consistency and does lead to improved treatment setup in more than 97% of attempted setups. Experienced use is correlated with a reduced range of setups between users and higher degree of setup improvement when compared with users new to the system