An International Inter-Consortium Validation of Knowledge-Based Plan Prediction Modeling for Whole Breast...

Area of Science:

• Radiation Oncology
• Medical Physics
• Artificial Intelligence in Medicine

Background:

• Knowledge-based (KB) planning leverages prior experience to optimize radiotherapy.
• Evaluating model transferability across institutions is crucial for sharing and broader implementation.
• This study validates KB prediction models developed in one consortium using data from another.

Purpose of the Study:

• To assess the transferability of right whole breast tangential field (RWB-TF) KB prediction models.
• To validate models developed by a national consortium using data from an international consortium.
• To determine if KB models can be effectively applied beyond their development environment.

Main Methods:

• Ten RWB-TF KB models were developed within a national consortium.
• A cohort of 20 patients from an external, multi-institutional consortium was used for testing.
• Transferability was defined by ipsilateral (IPSI) lung first principal component (PC1) within the 10th-90th percentile of the training set; predicted dose-volume parameters were compared to clinical DVHs.

Main Results:

• Planning target volume (PTV) coverage was comparable despite significant volume differences in PTV and contralateral breast (p=0.002, p=0.02).
• IPSI lung mean dose/V20 Gy prediction showed a standard deviation of 1.13 Gy/2.9% in the external cohort vs. 0.55 Gy/1.6% in the training cohort.
• Differences between predicted and clinical IPSI lung mean dose and V20 Gy were <2 Gy and <5% in 88.7% and 92.3% of cases, respectively; PC1 values were within the 10th-90th percentile for ≥90% of patients in 6/10 models.

Conclusions:

• RWB-TF KB models demonstrate feasibility for application beyond their development consortium.
• These findings support the broader clinical implementation of KB planning in radiotherapy.
• The study highlights the potential for cross-institutional collaboration and knowledge sharing in radiation oncology.