ESTRO 37 Abstract book

ESTRO 37

S576

induces morphological changes in the prostate that cause considerable difficulty in appropriate needle insertion during implant. To clarify the effect on quality of dose delivery, a retrospective analysis was performed using DVH analysis. Material and Methods A total of 580 OCPC patients were treated in our institute between January 2005 and December 2014 with BT alone (243:BT-pts) or BT after ADT (337:BTH-pts). Patients were classified as low/intermediate/high risk (282/255/43, respectively). BT was performed with real- time, intraoperative transrectal US-guided transperineal seed insertion with modified peripheral loading at a prescribed dose of 160Gy. ADT mostly consisted of combined androgen blockade for a median duration of 6 months (range:1 to 88 months). Mean comparative DVH parameters in BT-pts and BTH-pts included prostate volume (pV), the dose covering 90% of the pV (pD 90 ), and the pV covered by 160Gy or 240Gy (pV 100 , pV 150 ) measured during and 30 days after implantation, as determined by US (US-plan) and CT (CT-plan), respectively. These comparisons were also performed in subgroups of BTH-pts according to the duration of ADT (<12 months:BTH <12 vs. >12 months:BTH >12 ). Results The mean and standard deviation of DVH parameters according to US-plan and CT-plan in all patients, BT-pts, and BTH-pts are listed below.

larger dose reduction irrespective of smaller pV in BTH >12 group. With ADT for longer than a year, BT had a disadvantage compared to other irradiation modalities. PO-1026 LDR prostate brachytherapy inverse planning including dose-volume relation and tissue heterogeneity K.A. Mountris 1 , J. Bert 1 , D. Visvikis 1 1 INSERM UMR 1101 - LaTIM, Faculté de Médecine, Brest, France Purpose or Objective Current inverse planning systems for low-dose-rate (LDR) prostate brachytherapy employ meta-heuristic algorithms to predict implants’ position leading to the optimal dosimetric outcome. However, the optimality is compromised, mainly due to the dose overestimation induced by the AAPM TG-43 formalism, and limited control over the dose-volume relation. Our objective was to eliminate these issues, considering tissue heterogeneity during dose calculation using Monte Carlo (MC) dosimetry and evaluating the dose deposition to the volume of the prostate and the critical organs (urethra, rectum) in terms of dose-volume-histogram (DVH) evaluation during inverse planning. Material and Methods MC dose calculation is performed on a heterogeneous phantom of the patient’s anatomy including 4 materials (prostate, urethra, rectum, surrounding tissue) generated from the intraoperative ultrasound image. Needles crossing the prostate but no critical organs are selected and candidate implantation sites are extracted. MC dose kernels are pre-calculated prior to optimization for all the candidate sites. Dose calculations are performed using GGEMS, a graphics processing unit (GPU) MC simulation toolkit. The seed is modeled by a phase space, accounting for the exact seed geometric specifications and intra-seed particle interactions. The total number of seeds to be used is estimated by an empiric formula and an initial plan is generated by accumulating randomly selected seed MC dose kernels. The initial plan is optimized using fast simulated annealing (FSA) method. In each iteration of the FSA, the plan is updated by random single-seed swapping. The DVH of the updated plan’s dose map is calculated on the fly and the variation of the primary and secondary DVH metrics from the AAPM TG-137 report’s recommended values is minimized to obtain the optimal plan. Results For a database of 18 patients, the proposed inverse treatment planning method results in plans satisfying the TG-137 recommendations with 100% success. Compared to clinical plans delivered to the patients in this database, similar prostate V100 (0.2% difference) and reduced V150 by 6.1% are achieved. Furthermore, the urethra D10 and rectum D2cc are reduced by 4.0% and 0.6%, respectively. The implant optimization is performed in 30-45 s (15-20 s during MC dose kernel generation and 15 s during FSA). When tissue heterogeneity is not considered, prostate V100 is overestimated by 4%, similarly to previously published studies.

BT-pts (243)

BTH-pts (337)

all (570)

p

pV US-plan (cc) pV CT-plan (cc)

0.000

28.1±10.6 30.9±11.5 25.5±10.6

28.8±7.0 29.7±5.9

26.4±7.8

0.001

US-plan

pD 90 (Gy) pD 90 (Gy)

192±16.4 193±17.0 191±15.7

0.178

CT-plan

179±23.6 183±23.7 173±22.3

0.000

US-plan

pV 100

96.8±6.6 96.8±6.6

96.7±6.5

0.826

(%)

CT-plan

pV 100

93.2±6.4 93.9±6.4

92.2±6.3

0.001

(%)

US-plan

pV 150

0.638

62.6±14.0 65.6±10.2 63.9±9.5

(%)

CT-plan

pV 150

0.000

57.4±13.4 66.2±13.3 57.4±12.4

(%)

In general, pD 90 (prostate-DVH values) in the CT-plan were smaller than those in the US-plan. In a comparison between BT-pts and BTH-pts, there was no significant difference in prostate-DVH values in the US- plan. However, the results in BTH-pts were significantly less than those in BT-pts in the CT-plan. Change in mean pV in the US-plan and CT-plan was reciprocal in BT-pts and BTH-pts, with statistical significance in both cases (p=0.000). The pV was smaller in the BTH >12 group (23.1 ± 8.9 cc) than in the BTH <12 group (28.6±7.8 cc). The prostate-DVH values were smaller in the BTH >12 group than in the BTH <12 group:pD 90 :168 ± 24.7 vs. 175 ± 21.7Gy; pV 100 :89.3 ± 10.3 vs 92.9 ± 5.0%; pV 150 :52.6±16.1 vs , pV 100 , and pV 150

58.9±12.6%. Conclusion

Antecedent ADT induces additional degradation of prostate dose and coverage. The pV increase in the CT- plan might be partly responsible, but cannot explain the