Collapsed cone (CC) superposition method for β-emitter dosimetry: Implementation and validation for 90Y targeted radiotherapy (TRT)

1460 Objectives To speed up the absorbed dose computation (AD) while accounting for heterogeneity of tissue density and composition, a CC superposition algorithm using a Dose Point Kernel (DPK) in water was developed and validated on 90Y-TRT. Methods 90Y-DPK was generated in water medium with MCNP code. The superposition was implemented by scaling the DPK with the radiological distance between the source and destination voxels to account for tissue heterogeneity. To speed up the AD calculation for an extended activity distribution, a CC algorithm was implemented (Ahnesjo Phys Med 1989). Validation was performed using semi-infinite sources corresponding to 6 interface combinations involving soft tissue, lung and bone. Comparison to MC was done in terms of relative AD difference (ΔAD) in low gradient regions (LGR), distance to agreement (DTA) in high gradient regions (HGR) and the combined Γ criterion with a tolerance of 1% in dose and 1 mm in distance. A clinical validation was performed in case of an intravascular liver treatment with a lung shunt. The comparison to MC was expressed in terms of DVHs. Results For the semi-infinite sources, ΔAD in LGR was below 1.0% between the computation using CC and MC. DTA was below 0.6 mm. All profiles verified the Γ(1%,1mm) criterion. For the clinical case, differed by -0.8%, -0.2% and 0.3% in lung, non-tumoral liver and the tumor. DVHs departed by less than 0.6%, 0.4% and 0.3% for the same tissues respectively. Calculation time was below 4 min on a single processor for CC superposition and 40 h on a 40 nodes cluster for MC (100M histories). Conclusions Our results show that the CC superposition seems to be a very promising alternative of MC for 90Y TRT dosimetry, while greatly reducing computation time. Research Support INSERM grant n°PC201227