Developing a web-based dashboard for adaptive radiotherapy workflows

Area of Science:

• Medical Physics
• Radiation Oncology
• Radiotherapy Technology

Background:

• Adaptive radiotherapy (ART) requires reliable thresholds for plan adaptation.
• Current ART decision-making lacks unified metrics for geometric, dosimetric, and radiobiological data.

Purpose of the Study:

• To develop and verify a web-based dashboard for ART decision support.
• To integrate geometric, dosimetric, and radiobiological metrics into a unified platform.

Main Methods:

• A dashboard was developed using Python/Dash with a PostgreSQL backend.
• It processed RTStruct/RTDose data to compute DVHs and five radar-chart metrics.
• Metrics included organ displacement, Dice scores, objective scores, and radiobiology scores (TCP/NTCP).
• Verification involved comparing dashboard calculations against MIM Maestro for ten patients (Head and Neck, prostate MR-linac).

Main Results:

• Dashboard metrics showed negligible to small residuals compared to MIM Maestro.
• Geometric residuals (central point displacement, Dice scores) were within acceptable limits.
• Dose and volume endpoint residuals were small, with radiobiology residuals being modest.
• Three metrics reached statistical significance, but effect sizes were negligible to small.

Conclusions:

• The dashboard accurately reproduces geometric, DVH, objective, and radiobiology metrics, providing a credible foundation for ART decision support.
• Limitations include DVH projection as visualization and radiobiology sensitivity to DVH sampling in steep gradients.