1610-1611 / 2953
„Automated planning“
Basic idea/aim:
Automatically find a patient-specific set of CF-prescriptions/constraints as
starting point of an IMRT optimization that will result in the
clinically
most acceptable plan
from the solution space of this patient.
How does it work?
•
Finding: In a population of patients, a similar compromise of PTV
coverage vs. dose constraints will be
realizable
for patients with
Automatic choice of the
„optimal“ pareto-optimal plan!
ESTRO IMRT course 2016
IMRT optimization: algorithms&cost functions
London April 4th, 2016
‚similar geometry‘
•
Use of prior knowledge to predict realizable DVHs/dose presciptions
for a new patient
•
Algorithm based on database of patients/correlation model/machine-
learning/...
To some extent institution-specific; the larger the ‚input‘ patient-
database, the better it works...