Dummy run quality assurance study in the Korean Radiation Oncology Group 19-09 multi-institutional prospective cohort study of breast cancer

Background The Korean Radiation Oncology Group (KROG) 19 - 09 prospective cohort study aims to determine the effect of regional nodal irradiation on regional recurrence rates in ypN0 breast cancer patients. Dosimetric variations between radiotherapy (RT) plans of participating institutions may affect the clinical outcome of the study. We performed this study to assess inter-institutional dosimetric variations by dummy run. Methods Twelve participating institutions created RT plans for four clinical scenarios using computed tomography images of two dummy cases. Based on a reference structure set, we analyzed dose-volume histograms after collecting the RT plans. Results We found variations in dose distribution between institutions, especially in the regional nodal areas. Whole breast and regional nodal irradiation (WBI + RNI) plans had lower inter-institutional agreement and similarity for 95% isodose lines than WBI plans. Fleiss's kappa values, which were used to measure inter-institutional agreement for the 95% isodose lines, were 0.830 and 0.767 for the large and medium breast WBI plans, respectively, and 0.731 and 0.679 for the large and medium breast WBI + RNI plans, respectively. There were outliers in minimum dose delivered to 95% of the structure (D95%) of axillary level 1 among WBI plans and in D95% of the interpectoral region and axillary level 4 among WBI + RNI plans. Conclusion We found inter-institutional and inter-case variations in radiation dose delivered to target volumes and organs at risk. As KROG 19 - 09 is a prospective cohort study, we accepted the dosimetric variation among the different institutions. Actual patient RT plan data should be collected to achieve reliable KROG 19 - 09 study results.

Automated summary

This study assessed inter-institutional dosimetric variations in radiation dose plans for breast cancer patients. Variations were found, particularly in the regional nodal areas. The study highlights the importance of collecting actual patient data for reliable research results.