Click here to

Session: Automation in Treatment Planning [Return to Session]

The Impact of Advanced ESAPI-Based Automated Total Body Irradiation Treatment Planning On Plan Standardization

R Frederick1,2*, L Van Dyke2, J Lovis2, A Hudson1,2, G Pierce1,2, (1) University of Calgary, Calgary, AB, CA, (2) Tom Baker Cancer Centre, Calgary, AB, CA


MO-H345-IePD-F7-5 (Monday, 7/11/2022) 3:45 PM - 4:15 PM [Eastern Time (GMT-4)]

Exhibit Hall | Forum 7

Purpose: To determine if implementing automated treatment planning for volumetric modulated arc therapy total body irradiation (VMAT-TBI) improves plan standardization.

Methods: Our VMAT-TBI technique uses extended source-to-skin distance anteroposterior-posteroanterior (AP/PA) arc delivery. Manual beam placement for treatment planning requires import of Python-modified arcs and setting other beam parameters by hand. In March 2021, we introduced automated beam placement through the Varian Eclipse Scripting Application Programming Interface (ESAPI). Our script sets beam parameters and calculates arc length and field isocentre placement depending on body contour height and AP width. Dosimetrists generate plans from Varian’s Eclipse using a graphical user interface. We performed a retrospective review of 84 automated plans and 84 manual plans to determine plan standardization changes. Automated data querying was completed using ESAPI. Body-5mm D98% and D2% and mean lung dose (MLD) were compared between the manual and automated planning groups. Plan parameters were compared to an external standard to determine plan compliance.

Results: All results are listed for manual and automated cohorts respectively. Median Body-5mm homogeneity indexes (HI = D2/D98) are both 1.18, and median MLD are 103.6% and 104.0%. Medians are not significantly different, but variances are. Less variation is seen in the automated plans; HI: 1.5 and 0.8 (p=0.015), and MLD: 0.004 and 0.001 (p=0.007). For parameter checks, the total plan failure rates are 2.1% (202/9457 checks) and 0.5% (51/10111 checks). Median passing rates are 97.6% and 99.2% (p<<0.05). The decrease in failure rate for automated plans is attributable to course and arc ID standardization. Notable residual failures in automated plans occurred due to human error.

Conclusion: Automation in TBI treatment planning improves standardization, particularly for IDs that will aid in future data querying. The correlation of decreased variation in both HI and MLD and automated planning points towards a potential dosimetric benefit.

Funding Support, Disclosures, and Conflict of Interest: This work was supported by the Natural Sciences and Engineering Research Council of Canada and the University of Calgary Eyes High Program. There are no relevant financial disclosures or conflicts of interest to declare.


TBI, Treatment Planning


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

Contact Email