ESTRO 2024 - Abstract Book

S4421

Physics - Machine learning models and clinical applications

ESTRO 2024

sec, and prediction took on average 1.14 sec per scenario. Strategy (3) has the potential to scale up to larger scenario numbers.

Figure 1: Boxplots showing differences in plan parameters between the Erasmus-iCycle ground truth (GT) and predictions (pred) with the three investigated DL strategies, over all scenarios for the 78 test patients.