SU‐FF‐J‐37: Comparison Between CT‐Based and Ultrasound‐Based Localization for Prostate Patients

Purpose: Prostate cancer is often treated with intensity‐modulated radiation therapy (IMRT) that provides very steep dose fall‐off outside of the target volume and therefore requires precise alignment. In this report, we compare ultrasound (U/S) and CT localization, and evaluate the uncertainties of the two modalities. Method and Materials: A total of 19 prostate patients (275 alignments) were included in the study. The prostate was localized with the BAT U/S system (Nomos, Cranberry, PA). CT scans were performed with the Primatom CT‐on‐Rails system (Siemens, Concord, CA). The two alignment techniques were identical: the simulation contours of the prostate, proximal seminal vesicles, bladder, and rectum were aligned with a pre‐treatment image. The random uncertainty of the CT system was estimated based on intra‐user variability, and a simple mathematical model of the motion of radio‐opaque markers. Then the random error of the ultrasound alone was calculated. Results: The systematic differences between (U/S) and CT alignments were (in mm): 0.3 (lateral), 0.4 (AP) and 0.3 (longitudinal). The random differences between the two modalities (one standard deviation) were (in mm): 2.2 (lateral), 2.2 (AP) and 2.4 (longitudinal). The estimated ranges of random uncertainties of the CT alignments (one standard deviation) were (in mm): 0.9 to 1.2 (lateral), 0.7 to 1 (AP), and 1.2 to 1.5 (longitudinal). Based on these results, the calculated ranges of random uncertainties of the U/S alignments were (in mm): 1.8 to 2.0 (lateral), 2.0 to 2.1 (AP), and 1.9 to 2.1 (longitudinal). Conclusion: There was a strong correlation between the extent of prostate inter‐fraction alignments using ultrasound and CT. The localization of the prostate had a total uncertainty (two standard deviations) of 2 to 3 mm when using the Primatom CT‐on‐Rails and around 4 mm when using the BAT system.