Room: AAPM ePoster Library
Purpose: To evaluate the dosimetric uncertainty in TRUS-guided HDR prostate brachytherapy due to catheter tracking errors, via Monte Carlo simulation of geometric errors.
Methods: A fast GPU-based TG-43 implementation was written to independently re-calculate the dose distribution for a 17-catheter TRUS-guided HDR prostate boost plan (15 Gy single fraction) following simulated introduction of catheter trajectory errors. For each catheter in the plan, axial/lateral position errors were sampled from Gaussian distributions whose FWHMs matched those of axial/lateral profiles through catheters in the TRUS images. Errors were drawn at both ends of the catheter to generate trajectory errors. Insertion depth errors were also obtained from a Gaussian distribution with 2 mm FWHM.
1000 iterations of the simulation were performed to generate a 95% confidence-interval dose volume histogram. Additionally, the mean and standard deviation of GEC/ESTRO-recommended DVH reporting parameters were calculated.
Results: From simulations, V100% for the prostate CTV was 97.7 ± 1.2 %, D2cc for rectum was 868 ± 15 cGy, and D0.1cc for urethra was 1730 ± 80 cGy. Corresponding clinical values were V100% = 98.7%, D2cc = 855 cGy, and D0.1cc = 1685 cGy. For the CTV, clinical V100% lies near the upper bound of the confidence interval, as tracking errors lead to increased catheter spacing and/or distances between catheter and contour boundary, both reducing V100%. The low rectal and higher urethral dose uncertainties are logical given their location relative to catheters.
Conclusion: simulation tool was developed to study the dosimetric effect of catheter tracking uncertainty in TRUS-guided HDR prostate procedures. The image-based error estimates are conservative; thus the simulations show that the technique is robust. In the future, this tool can be used to quantify the effects of systematic errors in order to determine error tolerance levels, and to evaluate the dosimetric uncertainty of automated catheter reconstruction.