MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

Automated Proton Beam Model Validation

A Egan1*, D Maes1, R Regmi1, C Bloch1,2, J Saini1, P Taddei1,2, S Bowen1,2, T Wong1, (1) Seattle Proton Therapy Center, Seattle, WA, (2) University of Washington School of Medicine, Seattle, WA

Presentations

(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose:
Manual beam model validation can be an error-prone and time-consuming process. To verify clinical viability, the model must first be tested across a wide range of beam and geometry parameters. For each individual calculation, dose-distribution data must be manually exported and then manually matched and compared to measured data, often in varied software. At the Seattle Proton Therapy Center (SPTC) an in-house suite of Python-based automation scripts have been developed to automate this process.


Methods:
These tools were initially developed to assist in the commissioning of IBA uniform scanning [Ion Beam Applications, Louvain-Neuve, Belgium] for the RayStation treatment planning system [RaySearch Laboratories, Stockholm, Sweden]. The RayStation scripting interface was used in conjunction with Python code to automate beam model parameter variation, dose calculation, and data export. Parameter set lists can either be scraped from previously measured data file names or specified manually. Comparison data searching/matching, device-specific data file reading (e.g. multi-layer ion-chamber or planar array), data-type-specific reformatting (e.g. noise smoothing or curve normalization), and finally analysis (e.g. gamma comparison) is all fully automated requiring just minutes of user-specific configuration. Validation results are documented in pdf printing of comparison plots/tables, and analysis metrics compiled into a single csv file for easy review in spreadsheet software.

Results:
For this specific model validation instance, plan calculation and export required 30 minutes of user configuration and approximately 8 hours of unattended calculation time. Analysis of this data set required 30 minutes of configuration and approximately 1 hour of unattended calculation time.

Conclusion:
At SPTC, significant time savings have been achieved with the automation of an otherwise time-consuming standard medical physics task. Current automation modules are being extended to other modalities (e.g. proton pencil beam scanning) for future commissioning projects, and are also in current use for automated monthly and annual QA processes.

Keywords

Protons, Commissioning, Quality Assurance

Taxonomy

TH- External Beam- Particle/high LET therapy: Proton therapy – Development (new technology and techniques)

Contact Email