Click here to

Session: Automation in Treatment Planning [Return to Session]

Development and Implementation of An Automated Planning Algorithm for Multi-Metastatic Stereotactic Radiosurgery Using Eclipse Scripting

T Mann1,2*, N Ploquin1,2, K Thind1,3, (1) University of Calgary, Calgary, Alberta, CA, (2) Tom Baker Cancer Centre, Calgary, AB, CA, (3) Henry Ford Health Systems, Detroit, Michigan


MO-H345-IePD-F7-6 (Monday, 7/11/2022) 3:45 PM - 4:15 PM [Eastern Time (GMT-4)]

Exhibit Hall | Forum 7

Purpose: Treatment planning for multi-metastatic Stereotactic Radiosurgery (SRS) is a complex and lengthy process. This study investigates the implementation of a novel automated planning solution for multi-metastatic SRS to determine its impact on clinical efficiency.

Methods: During the 3 month accrual period 17 patients were prospectively planned using the Stereotactic Optimized Automated Radiotherapy (SOAR) planning solution (1 to 9 brain metastases). As described in our previous published work, this planning solution used patient-specific geometry, collision prediction, beam angle optimization with multiple geometric heuristics, and automation with Eclipse Scripting to design and build single or multi-isocenter treatment plans with minimal planner guidance, up to the first round of VMAT optimization. Initial plan quality immediately following the SOAR process was compared to the final clinical plan quality. Target coverage and dose to relevant OARs was compared using double-sided Wilcoxon signed-rank tests (α=0.05, no adjustment for multiple comparisons). Planning times for a manually planned retrospective cohort of 374 patients treated with SRS during the past two years (1 to 19 brain metastases) and prospective auto-planned patients were determined using patient Care Path analytics and compared.

Results: The average time for the SOAR planning process was 14 minutes (range: 6-30 minutes). Preferred target coverage (>99%) was achieved for 71% of targets in the initial plans compared to 98% in the final clinical plans. Differences in V10Gy and V12Gy volumes were not statistically significant. The average time for simple manual plans (1-3 metastases) was 5.9 hours compared to 4.8 hours for the automated planning cohort. The average time for complex manual plans (4+ metastases) was 10.1 hours compared to 8.5 hours for SOAR.

Conclusion: The novel SOAR automated planning solution improved clinical timing efficiency for SRS planning in this initial cohort by 16-19%. SOAR promotes consistency and quality in treatment planning for multi-metastatic SRS patients.

Funding Support, Disclosures, and Conflict of Interest: The authors thank Varian Medical Systems for in kind support.


Stereotactic Radiosurgery, Treatment Planning, Software


TH- External Beam- Photons: Treatment planning using machine learning/Knowledge Based Planning/automation

Contact Email