Room: AAPM ePoster Library
Purpose: To utilize cloud-based infrastructure towards improving best practice in treatment planning.
Methods: Both multi- and single-institutional planning studies were undertaken for a cohort of patients undergoing external beam radiotherapy for synchronous bilateral breast cancer treated in deep-inspiration breath-hold. The cohort included all bilateral breast cancer patients treated between 2015 and 2018. For both studies, dosimetric constraints were defined for normal tissue and target volumes. A patient from this cohort with representative anatomy was selected for the multi-institutional planning study. The multi-institutional study was conducted using the cloud-based ProKnow plan study quality system (Elekta AB, Stockholm, Sweden). This enabled users to create a treatment plan for the representative patient using any delivery or planning technique. The single-institutional planning study was performed by creating retrospective treatment plans for the entire cohort. In both studies, all plans that met target coverage goals were included and compared for normal tissue doses.
Results: Of the 36 cloud-based submissions, 28 plans used VMAT as the treatment modality and were included in the analysis. For these 28 plans, a mean (± standard deviation) heart dose of 6.6 (± 1.2) Gy was achieved and the mean lung V20Gy was 28.1 (± 2.8) %. The single-institutional study included all six bilateral breast patients treated within the specified time frame. For these six plans, a mean (± standard deviation) heart dose of 6.7 (± 0.9)Gy was achieved and the mean lung V20Gy was 26.4 (± 2.2) %. The normal tissue dose variability between the two studies was comparable.
Conclusion: The results of the multi-institutional study - even when limited to a single treatment technique (i.e. VMAT) - show variation similar to the variation reported in the single-institutional planning study with different patient anatomies. This demonstrates the capability of cloud-based infrastructure and multi-institutional studies to contribute towards best practice guidelines.
Funding Support, Disclosures, and Conflict of Interest: B. Nelms is a co-founder of ProKnow and has a commercial interest in the technology used in this study.