Related Experiment Videos
An institutional experience of a scripting-driven automation of the emergent palliative radiotherapy workflow
Shadab Momin1, Eduard Schreibmann1, Justin Roper1
1Department of Radiation Oncology, Emory University, Atlanta, Georgia, USA.
Background:
Emergency palliative radiotherapy is often delivered under time-constrained and high-stress conditions, frequently bypassing conventional computed tomography (CT) simulation and relying on nonstandard workflows. These factors increase variability and the potential for error.
Purpose:
This work describes the development, implementation, and early clinical evaluation of a simulation-free, cone-beam CT (CBCT) based emergency radiotherapy workflow that leverages treatment planning system (TPS) scripting to standardize and automate deterministic steps while preserving clinical oversight.
Methods:
An ESAPI scripting-driven emergency workflow was developed within the TPS to automate prescription handling, plan generation, beam configuration, dose calculation, and setup preparation using CBCT images. Safety checks were implemented in the form of prompts with the appropriate actions required. Alternatively, manual input was retained for image verification and treatment field borders definition. The workflow was implemented across multiple clinical sites within a single healthcare system and evaluated through dosimetric comparison of CT- and CBCT-based planning. Metric parameters such as workflow time, reductions in manual interactions, and pre- and post-training survey outcomes of automated 3D workflow were compared to the same parameters of the conventional 2D workflow.
Results:
The automated workflow successfully generated clinically acceptable emergency treatment plans with minimal manual input. CBCT-based dose calculations demonstrated agreement with CT-based planning within 2.5% when heterogeneity corrections were enabled and within 4.7% when disabled. Automation eliminated 33 manual data entry steps present in the 2D workflow and enabled treatment delivery with average on-table time of approximately an hour. Survey results demonstrated strong staff support for automation, with over 90% anticipating improvement prior to training and 100% reporting perceived improvement following training. Staff comfort and confidence improved post-training, and concerns regarding workflow clarity and safety decreased.
Conclusions:
A scripting-based, simulation-free emergency radiotherapy workflow was developed for safe and standardized palliative treatments while maintaining established clinical safeguards. By automating deterministic planning tasks and preserving human oversight for clinical decision-making, the proposed approach improves consistency and staff confidence and has potential to improve safety and accuracy of palliative treatments.