Author: CheukKai Becket Hui, Goldie M. Klein, Yildirim D. Mutaf, Amir Pourmoghaddas 👨🔬
Affiliation: UC Health, Kaiser Permanente 🌍
Purpose:
In this study, we investigated the effect of SGRT (Align-RT) implementation on patient setup accuracy for DIBH breast radiotherapy treatments.
Methods:
We extracted vertical post-imaging shifts and couch coordinates from the R&V (Varian) database for two groups of patients: 188 treated with conventional setup before SGRT implementation, and 196 who were set up with surface guidance. Only whole-breast RT cases were considered. The pre-treatment positioning protocol for all DIBH breast RT cases in our clinic involves daily 2D imaging for alignment. We compared the vertical post-imaging shift per fraction for all patients before and after the implementation of SGRT to demonstrate its effect on setup accuracy. Additionally, the standard deviation of vertical couch position during the course of treatment was also calculated to evaluate SGRT’s effect on positioning consistency.
Results:
The mean vertical couch shift decreased from 0.32 cm to 0.19 cm after implementation of SGRT (p < 0.001) and the mean standard deviations of vertical couch positions decreased from 0.34 to 0.30 cm (p<0.001), showing that implementation of SGRT led to more accurate initial positioning as confirmed by imaging. Similar trends were noted when separately analyzing right sided and left sided cases (from 0.36 to 0.23 cm and from 0.31 to 0.18, respectfully; p<0.001). The average of maximum vertical couch shifts (calculated for each patient) also decreased from 0.91 to 0.54 cm (p<0.001). The mean vertical couch shift decreased for patients in each fractionation cohort, differences were significant in all cases (p<0.01). Although the variability in couch position also decreased for each fractionation cohort, the differences were not significant.
Conclusion:
SGRT proves to be an effective technology for patient setup, resulting in a more accurate initial positioning compared to conventional setup and positioning.