Deep learning synthetic computed tomography supported head-and-neck adaptive radiotherapy dose recalculation
Synthetic computed tomography from cone-beam computed tomography achieved a 98.7% gamma pass rate for head-and-neck dose recalculation.
Synthetic computed tomography from cone-beam computed tomography achieved a 98.7% gamma pass rate for head-and-neck dose recalculation.
A physics-constrained network reconstructed pelvic cone-beam computed tomography from two simulated radiographs with substantially lower error than generative baselines.
Dynamic collimation improved target-region contrast and signal-to-noise ratio while preserving full-field information without increasing the total photon budget.