ESTRO 2024 - Abstract Book

S3489

Physics - Dose prediction, optimisation and applications of photon and electron planning

ESTRO 2024

Hung Hing Man

Pamela Youde Nethersole Eastern Hospital, Clinical Oncology, Chai Wan, Hong Kong

Purpose/Objective:

Hippocampal-sparing-whole-brain radiotherapy (HS-WBRT) refers to the IMRT technique that spares the hippocampal neural stem-cell compartment during whole-brain irradiation. Radiotherapy planning of HS-WBRT can be streamlined through an MR-only approach, where the Dixon-VIBE sequence and T1-weighted MPRAGE sequence can be acquired under a single MR-simulation session, allowing the generation of synthetic-CT (sCT) and contouring of the hippocampi respectively. This study aims to evaluate the dosimetric performance of HS-WBRT plans computed on sCT generated by the commercially-available deep learning-based software syngo.via RT Image Suite (RTiS) VB60A (Siemens Healthcare GmbH. Henkestr 127 91052 Erlangen Germany), and investigate the effect of previous cranial resection on sCT generation and its dosimetric impact. The feasibility of an MR-only workflow was also explored.

Material/Methods:

25 patients who underwent both MR and CT simulation in the brain region were included. Split-arc partial-field HS-WBRT VMAT plans were computed on sCT and recalculated on planning-CT (pCT) with preserved monitor units (MU). DVH parameters (D 2% , D 98% , D max , D min , and D mean of the PTV; D max , D 100% and D mean of the hippocampi; and D max to other OARs), coverage of the PTV encompassed by the prescribed dose (V 30Gy ), quality metrics (conformity index and heterogeneity index) and 3D gamma passing rate were used to evaluate the dosimetric performance of sCT-calculated plans. Difference in these parameters were compared between patients with intact skull bone (n=14) and patients with postoperative bone flaps (n=11).

Results:

All pCT-recalculated plans fulfilled either clinical goals as per protocol (n=17) or with variable acceptance (n=8). Although statistically significant differences were observed in PTV D 2% , D 98% , D min and D mean , hippocampi D max and D mean , deviations were within 1% for all patients. Mean global 3D gamma passing rate (GPR) in the body was 99.77% (3%/3 mm) and 90.68% (1%/1 mm). Exclusion of surface dose (body-3 mm) resulted in higher mean GPR of 99.93% (3%/3 mm) and 94.41% (1%/1 mm). When only the PTV was concerned, mean GPR were 99.97% (3%/3 mm) and 95.71% (1%/1 mm). While significant difference in heterogeneity index of the PTV and hippocampi D max was only observed in postoperative patients with bone flaps, plan quality was not negatively impacted. No significant difference was found in GPR between the two patient groups under all gamma criteria, except in 2%/2 mm global GPR, where patients with postoperative bone flaps had GPR 0.545% higher than patients with intact skull bone.

Conclusion:

This study provides evidence that sCT generated by syngo.via RTiS can be used to compute clinically acceptable HS-WBRT plans. Although the presence of bone flaps did not negatively impact plan quality in sCT-calculated plans, syngo.via RTiS has failed to faithfully model areas of bone flaps in sCT, resulting in appreciable differences between sCT-DRR and pCT-DRR. Prior to the adoption of syngo.via RTiS for an MR-only workflow, further research is required to validate the accuracy of sCT-DRR for pre-treatment positioning verification.