Cohort-based T-SSIM Visual Computing for Radiation Therapy Prediction and Exploration.

Area of Science:

• Medical Physics
• Computer Science
• Oncology

Background:

• Radiation therapy (RT) planning for head and neck cancer faces challenges with toxicity to surrounding organs at risk.
• Clinicians seek to predict RT outcomes by analyzing data from previously treated similar patients.

Purpose of the Study:

• To develop a visual computing approach for RT planning using spatial similarity.
• To introduce a novel topology-based spatial similarity measure (T-SSIM) and a predictive algorithm.

Main Methods:

• Developed a topology-based spatial similarity measure (T-SSIM).
• Created a predictive algorithm utilizing T-SSIM.
• Integrated the algorithm with a visual steering interface featuring parallel-marker encoding.

Main Results:

• Quantitative results obtained from a cohort of 165 patients.
• Qualitative evaluation conducted with radiation oncology, data management, biostatistics, and medical imaging experts.

Conclusions:

• The developed visual computing approach and T-SSIM offer a novel method for RT planning.
• The system facilitates understanding and prediction of RT outcomes based on patient cohort similarity.