Transabdominal ultrasonography, computed tomography and electronic portal imaging for 3-dimensional conformal radiotherapy for prostate cancer.

PURPOSE: To evaluate the feasibility and accuracy of daily B-mode acquisition and targeting ultrasound-based prostate localization (BAT) and to compare it with computed tomography (CT) and electronic portal imaging (EPI) in 3-dimensional conformal radiotherapy (3-D CRT) for prostate cancer. PATIENTS AND METHODS: Ten patients were treated with 3-D CRT (72 Gy/30 fractions, 2.4 Gy/fraction, equivalent to 80 Gy/40 fractions, for alpha/beta ratio of 1.5 Gy) and daily BAT-based prostate localization. For the first 5 fractions, CT and EPI were also performed in order to compare organ-motion and set-up error, respectively.
RESULTS: 287 BAT-, 50 CT- and 46 EPI-alignments were performed. The average BAT-determined misalignments in latero-lateral, antero-posterior and cranio-caudal directions were -0.9 mm+/-3.3 mm, 1.0 mm+/-4.0 mm and -0.9 mm+/-3.8 mm, respectively. The differences between BAT- and CT-determined organ-motion in latero-lateral, antero-posterior and cranio-caudal directions were 2.7 mm+/-1.9 mm, 3.9+/-2.8 mm and 3.4 +/- 3.0 mm, respectively. Weak correlation was found between BAT- and CT-determined misalignments in antero-posterior direction, while no correlation was observed in latero-lateral and cranio-caudal directions. The correlation was more significant when only data of good image-quality patients were analyzed (8 patients).
CONCLUSION: BAT ensures the relative positions of target are the same during treatment and in treatment plan, however, the reliability of alignment is patient-dependent. The average BAT-determined misalignments were small, confirming the prevalence of random errors in 3-D CRT. Further study is warranted in order to establish the clinical value of BAT.