A web-based application for automated quality assurance in medical imaging and radiotherapy using open-source modules

Abstract

Purpose: 

Rigorous quality assurance (QA) protocols are fundamental to the safety and efficacy of radiotherapy and medical imaging. However, manual QA evaluation is frequently labor-intensive and highly susceptible to subjective interpretation. While commercial automated solutions address these computational inefficiencies, they are often cost-prohibitive for resource-constrained clinical facilities. Conversely, powerful open-source libraries require programming proficiency, complex local environment setups, and ongoing software maintenance, limiting their clinical adoption. This work aims to develop a user-friendly, ready-to-use web-based application for automated QA using these open-source libraries.

Methods:

The web-based application was developed using the Python framework Streamlit to construct the frontend interface and the pylinac library to execute backend QA analysis. It automates core linear accelerator QA procedures: Winston-Lutz tests, light and radiation field coincidence, epid-based beam flatness and symmetry, spoke-shot analysis, CT and CBCT imaging evaluation, and Picket-Fence MLC tracking. The workflow optimizes clinical efficiency by allowing users to drag and drop exported DICOM files directly into the web browser. The application automatically analyzes the datasets against defined baseline tolerances to generate a visual pass/fail report.

Results:

Successfully deployed as a fully functional, cloud-hosted platform, the application demonstrated robust performance across standard web browsers without requiring localized software installation or specialized computational hardware. Preliminary deployment confirms a seamless workflow; the platform successfully parses uploaded DICOM datasets, executes complex QA algorithms, and renders visually intuitive pass/fail reports. The graphical interface effectively abstracts the underlying code, providing a zero-barrier entry point for clinical analysis.

Conclusion: 

This application bridges the gap between powerful open-source programming libraries and accessible clinical utility. By transforming complex coding environments into an intuitive, zero-installation interface, it provides a highly cost-effective, scalable solution for automated linac QA. The platform streamlines routine workflows for resource-constrained radiotherapy departments and establishes the essential framework for future clinical validation.