Author: Fatemeh Akbari, Deidre Batchelar, Luc Beaulieu, Nathan E. Becker, Juanita Crook, Dakota Mckeown, Matthew Jonathan Muscat, Rowan M. Thomson 👨🔬
Affiliation: Département de physique, de génie physique et d'optique, Université Laval, BC Cancer Agency, UBC, Carleton University, BC Cancer - Kelowna, BC Cancer 🌍
Purpose:
To study the effects of inter-seed attenuation and tissue inhomogeneities on dose-volume metrics of critical structures in prostate low-dose-rate (LDR) permanent seed implant brachytherapy.
Methods:
Dose distributions for 245 patients were simulated using the egs_brachy Monte Carlo dose calculation algorithm. Patients had received either 144Gy (monotherapy, 218 patients) or 110Gy (boost, 27 patients) to the clinical target volume (CTV, prostate) and all implants were performed using I-125 with source strengths between [0.328U, 0.493U]. Contours for missing bladder or rectal structures were supplemented by Limbus AI. Commonly used dose-volume metrics to the CTV, urethra, rectum and bladder were calculated under the assumptions of TG43 (all water) and TG186 (full tissue) for each patient and compared. Correlation was then investigated between these differences and: volume of prostate micro-calcifications, prostate volume and number of seeds.
Results:
P-values for all dose-volume metric differences for every structure were significant using the paired Wilcoxon signed rank test (p<0.05), except for bladder D1cc. The effect sizes and frequencies in which the TG43 dose-volume metric was greater than its TG186 pair were largest for the CTV ranging from: 58% (CTV V200%, mean difference 2.4Gy) to 77% (CTV D99.9%, mean difference 18.8Gy). Urethra D0.1cc was within this range with a frequency of 62% (mean difference 12.5Gy). The investigated correlates (median, range) were: calcification volume (0.29, [0.05,1.88])cc, prostate volume (36.8, [13.8, 88.0])cc, and number of seeds (94, [62, 151]). Linear regression indicated that the calcification volume was a statistically significant predictor (p<0.05) of the dose-volume metric paired differences (TG43-TG186) to the CTV, urethra and bladder, but not the rectum.
Conclusion:
This analysis demonstrates that the assumptions of the TG43 formalism, even when compared using the same underlying model architecture (Monte Carlo), may be overestimating commonly used dose metrics to the target volume and organs at risk.