When Can Log‑File QA Replace Measurement‑Based QA? a Multi‑Technique Correlation Study across Multiple Criteria
Abstract
Purpose
This study assesses whether log-file based patient-specific QA can reliably substitute for measurement-based QA by evaluating correlations between detector-based and log-file based gamma pass rates across multiple plan types and gamma criteria.
Methods
30 clinical plans were created in RayStation treatment planning software across different disease sites and techniques. Each plan was delivered on an Elekta VersaHD linear accelerator and measured using a 2.5 D ArcCHECK diode array detector. The resulting machine delivery log files were used in a web based independent log-file dose recalculation system (SunCHECK). Composite ArcCHECK gamma pass rates were compared against 3D SunCHECK gamma results with different criteria. Spearman correlation coefficients (ρ) were calculated and categorized as weak (ρ0.7). Different techniques were analyzed separately for VMAT, SBRT, and the overall dataset to determine the performance level for different techniques.
Results
Across all plans (n = 30), log-file and measurement-based QA demonstrated weak correlations with ρ ranging from 0.002-0.116. However, when analyzed by technique, VMAT plans (n = 15) demonstrated a moderate to strong correlation for common gamma criteria, with ρ ranging from 0.593 (3%/3mm), 0.557 (3%/2mm), and 0.589 (2%/2mm), and decreasing with stricter criteria. In contrast, SBRT plans (n = 15) demonstrated a weak correlation with ρ ranging from 0.122 (3%/3mm) to 0.436 (2%/1mm).
Conclusion
Overall, this study indicates that log-file QA demonstrates moderate reliability as a substitute for measurement-based QA for VMAT plans under standard clinical gamma criteria. Conversely, SBRT plans demonstrate a poor correlation between the two QA methods, suggesting that the appropriateness of replacing measurement-based QA with log-file based QA may be technique and device dependent.