SRS/SBRT commissioning process with a hybrid beam data model in new preconfigured CyberKnife system

A new preconfigured CyberKnife system was introduced to improve the SRS/SBRT commissioning efficiency. The vendor provides a reference beam data set, termed "beam data model (BDM)", and recommends its validation and combination with measurement data to develop a hybrid beam data model (h-BDM). However, the validation method by the user and the clinical acceptability with the h-BDM has not been established.

This study reports the first clinical validation of the BDM and aims to demonstrate the SRS/SBRT commissioning process with the h-BDM is clinically acceptable.

The BDM was validated by comparing with the measurements of tissue phantom ratios (TPRs), off-center ratios (OCRs), and output factors (OFs) for fixed and Iris collimators and MLC in CyberKnife S7 system. Dose difference (DD) and distance-to-agreement (DTA) between BDM and measurement data were assessed against predefined thresholds: DDs of 1.0% for TPRs; DDs of 1.0% at d_15mm and d_100mm and 1.5% at d_300mm for central regions in OCRs; and DTAs of 0.3-1.0 mm for penumbra regions. The h-BDM was constructed by supplementing the BDM with the measurement beam data for all OFs, and for the TPRs and OCRs of the three smallest field sizes with each collimator. Additional replacements were also made when discrepancies exceeded the proposed thresholds based on clinical usability considerations. The conventional measurement beam data model (m-BDM) was also developed with only measurement data. Single-beam validation was performed in homogeneous water-equivalent phantom and heterogeneous phantoms, including lung-equivalent slabs, by comparing between the measured and calculated point doses with h-BDM and m-BDM at multiple depths. Finally, patient-specific quality assurance (PSQA) was performed for SRS/SBRT plans. In phantom case, SRS/SBRT plans were created for a single spherical target on the CT images of the StereoPHAN phantom and anthropomorphic lung phantom for brain and lung sites, respectively. For the SBRT plans of spine and prostate sites, the C-shape and test-prostate structure sets, provided by the AAPM TG 119, were employed. In clinical case, six SRS/SBRT plans for each patients' CT images in brain, lung, spine, and prostate sites, were employed and the PSQA plans were generated using the CT images of the StereoPHAN phantom, anthropomorphic lung phantom, and the I'mRT phantom. The point dose and γ pass rate were evaluated using micro ionization chamber and radiochromic film, respectively.

Differences between the BDM and measurement exceeded the specific thresholds for the OCRs with three largest MLC field sizes. In the single-beam validation, dose differences between measurements and calculations with h-BDM were within 2.0% for field sizes of > 10.0 mm, except up to 3.0% at 140 mm depth with MLC, and within 5.0% for field sizes of ≤10.0 mm. In PSQA with the h-BDMs, the point dose differences were within 3.0% and the γ pass rates (2%/2 mm and 3%/1 mm) were > 95.0% for all SRS/SBRT plans. The single-beam validation and PSQA results of the h-BDM were comparable to those of the conventional m-BDM.

The SRS/SBRT commissioning process with the proposed h-BDM provides a new clinically acceptable framework.