Author: Jeffrey Andrews, Nathan E. Becker, Juanita Crook, Andrew Jirasek, Matthew Jonathan Muscat 👨🔬
Affiliation: UBC, BC Cancer Agency, BC Cancer 🌍
Purpose: To map dosimetry and imaging information to ultrasound and multi-parametric magnetic resonance (mpMR) guided trans-perineal core-needle biopsies, performed during two-fraction prostate high-dose-rate (HDR) brachytherapy.
Methods: Patients with intermediate risk prostate cancer underwent HDR brachytherapy of the prostate in two equal fractions of 13.5Gy. MpMR and ultrasound fusion guided trans-perineal core-needle biopsies targeting DILs were collected in both treatment fractions, prior to radiation delivery of the respective fraction. Imaging of the biopsy needle extraction site were collected from approximately 40 patients, of which 7 patients have been assessed to-date. The biopsy extraction sites were estimated, segmented and registered to the patients’ planning structure set, delivered dosimetry and mpMR images. Positional uncertainty of the biopsy extraction site was propagated to uncertainty in dosimetry and imaging information by randomly sampling in the vicinity of the estimated biopsy extraction site. Simulated cores positioned at optimal locations within the prostate were compared to excision site locations of delivered biopsies, to determine if accuracy may be significantly improved.
Results: The mean, 25th and 75th quantiles of the distribution of probability of successful lesion excision, provided by all sampled biopsy voxels are 25%, 2%, and 75% respectively. Simulated optimized biopsy segmentations exhibit significantly (p < 0.05, Wilcoxon) greater successful lesion excision probabilities, with a mean difference of -37% and a common language effect size of 4% (paired, delivered–simulated). The distribution of dosimetry across all sampled biopsy voxels provided 25th, 50th and 75th quantiles: 18.12Gy, 21.36Gy and 30.30Gy respectively. The distribution of apparent diffusion coefficients across all voxels provided 25th, 50th and 75th quantiles: 1170 mm2/s, 1338 mm2/s, 1626 mm2/s.
Conclusion: This research demonstrates a novel methodology to map dosimetry and imaging information along the length of the extracted tissue core. This may be applied to retrospective biopsy assessment, and for correlative research.