Phantom treatment experiments dashboard

The Problem

A radiotherapy device manufacturer needed to collect, validate, and process phantom treatment experiment records from multiple machines before results could reach clinical and engineering teams. After each phantom was treated, the experiment produced a wide range of artifacts that needed to be analyzed together: treatment plans, projected treatment session dose, physical radiation measurements (film, dose chamber, and other detectors), and machine telemetry data. The process depended on manually triggered scripts and one-off status checks, so there was no single place to see where an experiment stood, correlate its artifacts, or catch a failed run early.

Our Approach

We built a Python-based orchestration pipeline that automatically discovers new phantom treatment experiment records per machine, runs them through processing and validation, and archives results to cloud storage. On top of it, we delivered a full-stack dashboard (Angular front end, NestJS backend) so engineers can browse experiments, inspect the associated treatment plans, projected dose, physical measurements, and telemetry side by side, and track pipeline state in one place, with automated Slack alerts replacing manual status checks.

Results

• Eliminated manual script-running for phantom treatment experiments across all machines
• Full visibility into historical data across every experiment and its artifacts
• Significantly simplified debugging by correlating treatment plans, dose, measurements, and telemetry in one view
• Early detection of hardware faults surfaced through telemetry and measurement trends
• Easier cross-release comparisons of treatment performance
• Simplified, centralized data archival for every experiment