(Tuesday, 7/16/2019) 10:30 AM - 11:00 AM
Room: Exhibit Hall | Forum 7
Purpose: The Halcyon (Varian) has a novel two-layer MLC design. We assessed the clinical impact of MLC positioning errors and determined the ability of portal dosimetry to detect such errors.
Methods: Positioning errors were simulated by adding random shifts within ±3mm, ±5mm, ±7mm, or ±10mm, or systematic shifts of 3mm, 5mm, 7mm, or 10mm, to MLC control points in 11 physician-approved head-and-neck treatment plans (99 total plans). Plans were recalculated and considered to contain clinically-significant errors if >5% changes to DVH metrics of normal tissues or PTVs occurred. To assess the ability of portal dosimetry to detect these errors, we performed VMAT portal dosimetry QA for the original and simulated error plans. The delivered fluence from the simulated error plans was compared to calculated fluence from the original clinical plans using global gamma analysis at 3%/3mm, 3%/2mm, 2%/2mm, and 2%/1mm, with a 95%-agreement passing criteria.
Results: The following percentage of systematic error plans experienced clinically significant errors to normal structures at the indicated amount of MLC error: 36%, 64%, 64%, and 100% at 3mm, 5mm, 7mm, and 10mm, respectively. For PTVs, these values were 18%, 91%, 100%, and 100%. Fewer plans with random MLC errors experienced clinically significant errors to normal structures: 0%, 9%, 45%, and 73% at 3mm, 5mm, 7mm, and 10mm. For PTVs, only 9% of plans had clinically significant errors at any magnitude of random error.Gamma analysis using 3%/2mm identified and failed all plans with systematic error of ≥3mm, while allowing most plans with 3mm (100%) and 5mm (82%) random error to pass. Other evaluated criteria were less successful.
Conclusion: Systematic errors have greater dosimetric impact than do random errors but are reliably detected by portal dosimetry. Small random errors are less reliably detected but can be identified as they reach magnitudes that impact patient treatment.
Funding Support, Disclosures, and Conflict of Interest: This work was partially funded by Varian Medical Systems.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email
