2602 - Validation of a Treatment Contingency Application

Purpose/Objective(s): Radiotherapy workflows depend on the continuous availability of OIS and TPS platforms. Unplanned server outages or system failures can significantly disrupt clinical operations and delay/prevent patient treatments. This study validates a Radiation Oncology Treatment Continuity Application (RadOncTCA) – a contingency tool designed for safe treatment delivery and decision-making during OIS downtime.

Materials/Methods: The RadOncTCA framework was developed to maintain clinical throughput during an OIS outage, defined institutionally as any OIS failure that severely impairs patient care and prevents treatment on two or more treatment units. The system automatically manages an external cache in real-time of the necessary DICOM files and treatment documentation, while also creating a web-based contingency report compiling the necessary data needed to confirm data integrity prior to treatment delivery. During an OIS outage, treatment schedules are verified via web-based RadOnCTCA, and associated treatment approved plans are loaded on the C-arm LINACs from file mode containing all the externally stored plan data. Independent verification of plan parameters, prescriptions, and imaging instructions against the web-based report is conducted prior to treatment delivery. Following OIS restoration, treatment delivery logs are imported back to OIS, and the delivered fractions are verified. For validation, an independent automated application was developed to perform a comprehensive comparison between the exported plan data, the contingency report, and the clinical system. We present the results of the RadOncTCA validation.

Results: Validation included an audit of 298 plans from 257 patients across multiple machines. Results demonstrated 100% accuracy for the appointment schedule, plan details, and plan reports. However, 2% of plans were identified as non-treatable due to missing or inaccurate DICOM data. Major discrepancies included 6 cases of missing DICOM data, specifically structure set files required for patient alignment and one instance of an incorrect plan version.

Conclusion: RadOncTCA provides reliable scheduling and reporting, but the 2% failure rate in DICOM data integrity presents a significant clinical risk. These errors would prevent treatment or potentially lead to medical events if SOPs are not strictly followed. This highlights the necessity of independent machine verification during outages. All major deviations were successfully detected via the independent validation tool, and RadOncTCA updates to the software to mitigate these risks are being implemented currently. The department’s recommendations are to implement a mandatory validation workflow to verify data integrity before treatment delivery during an outage.