ESTRO 2024 - Abstract Book

S4536

Physics - Machine learning models and clinical applications

ESTRO 2024

TB0:

- Abdomen

- Mean Gap between MLC leaves

- 20 failed

- Bone

- Q1 gap between MLC leaves

- 1153 passed

- Brain

- Mean TGi

TB1:

- Breast

- TGi ratio

- 41 failed

TB0:

- Extremities

- Mean MLC Speed

- 501 passed

VMAT

- 151 failed - 151 passed

- Head and Neck

- MLC speed modulation

TB3:

- Lung

- Mean Gantry Speed

- 149 failed

- Pelvis

- Gantry Speed modulation

- 1008 passed

- Prostate

- Mean RepRate

Clinac:

- Thorax

- RepRate modulation

- 9 failed - 13 passed

TB0:

- 5 failed

- 456 passed

- Brain

- Mean Gap between MLC leaves

TB0:

TB1:

- Breast

- Q1 gap between MLC leaves

- 41 failed

- 5 failed

- Extremities

- Mean TGi

- 41 passed

- 544 passed

IMRT

- Head and Neck

- TGi ratio

TB1:

TB3:

- Lung

- Mean MLC Speed

- 21 failed - 21 passed

- 106 failed

- Pelvis - Thorax

- MLC speed modulation

- 834 passed

Clinac:

- 412 failed - 1306 passed

Results:

The most important features for classification are the mean gap between MLC leaves followed by MLC Speed modulation for VMAT and the TGi Ratio, followed by the mean TGi and the mean MLC speed for IMRT.