Potential of E-Learning Interventions and Artificial Intelligence-Assisted Contouring Skills in Radiotherapy: The...

Area of Science:

• Radiation Oncology
• Medical Imaging
• Artificial Intelligence

Background:

• Research on AI-assisted contouring predominantly originates from high-income countries.
• The efficacy and safety of AI tools may be influenced by local factors in different healthcare settings.
• Understanding the impact in low- and middle-income countries (LMICs) is crucial for global adoption.

Purpose of the Study:

• To evaluate the impact of AI-assisted contouring on contouring time and quality.
• To assess the effectiveness of AI-assisted contouring and educational interventions among radiation oncologists in LMICs.
• To provide insights into the applicability of AI contouring in resource-limited environments.

Main Methods:

• Ninety-seven radiation oncologists were randomized to manual or AI-assisted contouring groups.
• A teaching session on contouring guidelines was provided between contouring tasks.
• Contour quality was measured using Dice Similarity Coefficient (DSC) against expert contours.
• Contouring time was recorded for various organs-at-risk (OARs).

Main Results:

• AI-assisted contouring significantly improved Dice Similarity Coefficient (DSC) for optic nerve, oral cavity, parotid, spinal cord, and mandible.
• Contouring time was reduced for brain stem, mandible, optic nerve, parotid, and thyroid using AI assistance.
• Teaching alone improved DSC for oral cavity and thyroid but increased contouring time for mandible, oral cavity, and thyroid.

Conclusions:

• AI-assisted contouring is demonstrated to be safe and beneficial for radiation oncologists in LMICs.
• The findings support the integration of AI contouring tools in LMICs.
• Further prospective clinical trials are recommended post-implementation to validate these effects.