Machine learning-based radiotherapy time prediction and treatment scheduling management.

Area of Science:

• Medical Physics
• Artificial Intelligence in Healthcare
• Radiotherapy Technology

Background:

• Efficient utilization of medical devices, particularly radiotherapy equipment (linacs), is critical due to high costs and patient loads.
• China faces challenges with a large population and limited healthcare resources, necessitating improved medical device efficiency.
• Radiotherapy devices are specialized, valuable, and handle significant treatment volumes, making their efficient operation paramount.

Purpose of the Study:

• To propose and evaluate a novel method for enhancing the operational efficiency of radiotherapy devices (linacs).
• To improve scheduling management by accurately predicting total treatment times for each appointment.
• To enable flexible assignment of treatment and non-treatment tasks based on predicted availability.

Main Methods:

• Collected data from 1665 patients, including patient positioning time (PT) and treatment time (TT).
• Extracted features related to PT and TT to train a machine learning model for predicting these times.
• Applied the prediction results to a minute-based scheduling tool for practical implementation.

Main Results:

• Developed a machine learning model capable of accurately predicting patient positioning time and treatment time.
• Demonstrated encouraging prediction outcomes for effective radiotherapy scheduling management.
• Showcased the potential to significantly improve the efficiency of linac operations through AI-driven predictions.

Conclusions:

• Artificial intelligence offers a promising solution for complex problems in specialized fields like radiotherapy.
• The study's findings indicate that AI-based prediction can enhance linac efficiency and scheduling.
• This research expands the application of medical data and suggests future research avenues in radiotherapy optimization.