ESTRO 2024 - Abstract Book

S5859

RTT - Service evaluation, quality assurance and risk management

ESTRO 2024

34

Digital Poster

Analysis of delivery uncertainty for breast VMAT using transit dosimetry

Sangjun Son

Seoul National University Hospital, Radiation Oncology, Seoul, Korea, Republic of

Purpose/Objective:

This study was conducted to quantitatively assess the accuracy of dose delivery during VMAT (Volumetric Modulated Arc Therapy) treatment for breast cancer using the EPID (Electronic Portal Imaging Device) in real patients. The research aimed to explore the factors that influence dose delivery uncertainties.

Material/Methods:

A total of 40 breast cancer patients, including those with SCL (Superior Clavicular Lymph nodes), Axillary LN, and IMN (Internal Mammary Node) involvement, were selected for this study. They received treatment on the Varian Halcyon 2.0 treatment machine, with a daily dose of 2.7 Gy delivered over 15 or 16 fractions. The study comprised twenty left sided breast cancer patients and twenty right-sided breast cancer patients. Transit dosimetry images obtained through the EPID were analyzed every other day, starting from the first treatment image. The research focused on measuring two key parameters: the gamma index average, referred to as Accuracy (indicating the precision of dose delivery during treatment), and the standard deviation, referred to as Deviation (indicating the variation in dose delivery between treatments). To capture the combined effect of these parameters, a coefficient of variation (Deviation / Accuracy) was calculated. The study also investigated the correlations between these values and patient specific factors, such as breast volume, body mass index, age, and breast density (determined using the absolute value of breast HU SD / HU mean ). Furthermore, a multivariate logistic regression model was developed to predict whether the treatment met the acceptable pass/fail criteria with a 90% accuracy tolerance level.

Results:

The average Accuracy for the 40 patients was found to be 92.74%±7.04% (with a 2 mm / 2% threshold and absolute dose criteria). The correlation analysis indicated a statistically significant relationship between the CV and left breast volume (r=0.65, p=0.002) and breast density (r=-0.264, p=0.021). However, this significance was not observed for the right breast. Nine out of the 40 patients did not meet the 90% Accuracy level. The logistic regression model, which predicted the acceptability of treatment based on breast volume, mean and standard deviation of HU in the breast, breast density, body mass index, and age, exhibited an AUC (Area Under the Curve) of 0.89 (AUC p=0.0004, and Hosmer-Lemeshow p=0.2333).

Conclusion: