Feasibility and Clinical Acceptability of Automation-Assisted 3D Conformal Radiotherapy Planning for Patients With...

Area of Science:

• Radiation Oncology
• Medical Physics
• Cancer Treatment

Background:

• The Ocean Road Cancer Institute (ORCI) in Tanzania introduced 3D conformal radiation therapy (3DCRT) in 2018.
• Challenges including steep learning curves, high patient volume, and a limited workforce led to extended radiation therapy (RT) planning workflows.

Purpose of the Study:

• To assess the feasibility of implementing an automation-assisted cervical cancer 3DCRT planning system at ORCI.
• To evaluate the efficiency and accuracy of automated contouring and planning compared to manual methods.

Main Methods:

• Chart abstractions were performed on 30 cervical cancer patients treated with 3DCRT.
• The Radiation Planning Assistant (RPA) generated automated contours and plans from anonymized CT images.
• Physician investigators assessed edit times, dose-volume safety metrics, and clinical acceptability. Dice Similarity Coefficient (DSC) analyzed contour agreement.

Main Results:

• Automated planning was feasible for 97% of patients, with same-day contour and plan generation.
• Automated contours required minimal edits (<2 minutes), and plans required <4 minutes.
• High agreement was observed for normal structures (e.g., bladder DSC 0.90), with lower agreement for target volumes (CTVp DSC 0.69, CTVn DSC 0.63).
• RPA plans demonstrated excellent target coverage.

Conclusions:

• Automation-assisted 3DCRT contouring shows high agreement for normal structures and meets safety and acceptability metrics.
• The RPA system has the potential to significantly reduce RT planning timelines.
• This technology can enhance high-quality RT delivery in resource-constrained environments.