Comparison of ultrasound and implanted seed marker prostate localization methods: Implications for image-guided radiotherapy

Purpose: To analyze two methods of image-guided radiotherapy (IGRT) for external beam radiotherapy of prostate cancer. Methods and Materials: The prostate was localized by ultrasound (US) in lateral (left/right), vertical (anteroposterior), and longitudinal (superior/inferior) dimensions and then by fiducial seed marker (SM) kV X-ray. Assuming initial setup to skin marks as the origin, the mean suggested shifts (for all dimensions) were hypothesized to be similar and within 1 mm of the origin. The three-dimensional distance discrepancy between suggested SM and US shift points was calculated. We hypothesized a mean discrepancy >5 mm as clinically significant. Results: From 40 patients, 1019 US/SM measurements were obtained. Lateral, vertical, and longitudinal dimensional comparisons reveal statistically significant differences in mean shifts (p < 0.0001 for all). US dimensional shifts reveal significantly greater variability. The US three-dimensional vector is greater and more variable than the SM vector (p < 0.0001). The mean US/SM three-dimensional distance discrepancy is 8.8 mm (significantly >5 mm, p < 0.0001). Conclusions: Ultrasound and SM methods suggest different shifts. US data reveal greater systematic/random error vs. SM data. The US data suggest larger PTV expansion margins (∼9 mm) are necessary for US IGRT vs. SM IGRT (∼3 mm). The hypotheses that US and SM methods suggest similar shifts and that the mean US/SM three-dimensional distance discrepancy is ≤5 mm are rejected.