Impact of electrometer sensitivity variation on absorbed dose-to-water determination: a multicenter uncertainty study

Technical Summary: Impact of Electrometer Sensitivity Variation on Absorbed Dose-to-Water Determination

Problem Statement
Accurate determination of absorbed dose to water in radiotherapy relies on a traceability chain involving primary standards laboratories (PSL). While protocols such as AAPM TG-51 and IAEA TRS-398 mandate the use of PSL-calibrated electrometer coefficients ($k_{elec}$) to ensure accuracy, periodic verification of electrometer performance is also critical. Traditionally, this verification requires either irradiation in a beam (subject to output variations) or the use of expensive, dedicated precision current sources.

Recent work by Tsuno et al. suggested that electrometers of the same model exhibit sufficiently similar sensitivity characteristics to allow the use of a common, user-derived calibration coefficient ($k_{elec,user} \approx 1.000$) for quality control, utilizing a low-cost, battery-powered current source (SCG002). However, while the variability of this user-derived coefficient was previously quantified at approximately 0.20%, its specific impact on the overall uncertainty budget of absorbed dose-to-water determination had not been quantitatively evaluated. The central problem addressed is whether substituting institution-specific PSL coefficients with a common user-derived coefficient introduces a meaningful uncertainty contribution that could compromise reference dosimetry.

Methodology
This study employed a multicenter approach involving 12 radiotherapy facilities across Japan, encompassing 24 unique combinations of linear accelerators, ionization chambers, and electrometers.

• Systems: The study utilized 6-MV and 10-MV photon beams under reference conditions following the AAPM TG-51 protocol. All facilities used a TN30013 Farmer-type ionization chamber (PTW) paired with various electrometers (PTW UNIDOS series and Toyo Medic RAMTEC series).
• Data Source: The study did not experimentally derive new $k_{elec,user}$ values. Instead, it incorporated previously published data from Tsuno et al. regarding the variability of user-derived coefficients obtained via the SCG002 current source.
• Comparative Analysis: Absorbed dose-to-water values were calculated using two distinct models:
  1. Conventional Model: Using the PSL-provided calibration coefficient ($k_{elec}$).
  2. Common Coefficient Model: Using a generic user-derived coefficient ($k_{elec,user} \approx 1.000$).
• Output Metrics: The calculated absorbed doses were converted to Dose Monitor Unit (DMU) values to facilitate intercomparison.
• Uncertainty Analysis: Uncertainty budgets were constructed in accordance with the Guide to the Expression of Uncertainty in Measurement (GUM). The analysis compared the combined standard uncertainty and expanded uncertainty of both models, specifically isolating the contribution of the electrometer calibration coefficient.

Key Contributions and Results
The study provides a quantitative assessment of the uncertainty introduced by replacing PSL-calibrated coefficients with a common user-derived value.

• Dose Comparison: The comparison of DMU values revealed negligible differences between the two methods. For 6-MV beams, the mean DMU values were $1.003 \pm 0.0044$ (using $k_{elec}$) and $1.002 \pm 0.0046$ (using $k_{elec,user}$). For 10-MV beams, the means were $1.002 \pm 0.0044$ and $1.002 \pm 0.0045$, respectively. The coefficients of variation (CV) remained below 0.5% for both conditions, and all values stayed within a $\pm 1.5\%$ range.
• Uncertainty Budget Impact: The substitution of the coefficient resulted in a minor increase in the user-dependent uncertainty component, rising from 0.17% to 0.23% (at $k=1$). This represents an increase of approximately 0.06 percentage points.
• Overall Significance: Crucially, the study found that this increase remained substantially smaller than the dominant uncertainty sources in reference dosimetry (such as the beam quality conversion coefficient, $k_{Q,Q0}$). Consequently, the combined standard uncertainty and the expanded uncertainty of the absorbed dose-to-water determination remained effectively unchanged when using the common coefficient.

Significance and Claims
The paper concludes that the uncertainty contribution associated with using a common electrometer calibration coefficient is small and has no meaningful influence on the overall uncertainty budget of absorbed dose-to-water determination.

The authors assert that these findings support the use of a common coefficient for electrometer sensitivity verification and quality-control applications within a specific electrometer type. The study positions this approach not as a replacement for traceability, but as a practical, complementary method for routine quality assurance. While periodic PSL calibration remains essential for maintaining the primary traceability chain, the common-coefficient-based verification offers a viable, efficient alternative for routine checks, particularly in settings where dedicated precision current sources are unavailable. The results provide quantitative evidence that the variability of user-derived coefficients does not compromise the accuracy of reference dosimetry.