Clinically oriented inverse planning implementation

Inverse treatment planning has been implemented in a more clinically oriented approach. The clinical implementation required the use and development of an optimization algorithm that was flexible, easy to use, fast, accurate, and reliable. The authors used a variant of the dynamic penalized likelihood (DPL) inverse planning algorithm based on the maximum likelihood estimator (MLE) used for image reconstruction in conjunction with a penalization term. The implementation was achieved by focusing on 3 key areas of the inverse planning process, particularly, the number and type of preoptimization parameters, the control held by the optimization algorithm over the planning process, and the type of plan evaluation tools available in inverse planning. The inverse planning algorithm was developed to accept few and only clinically relevant preoptimization parameters and to calculate several alternative optimized solutions for the same clinical case. Requiring few parameters as well as not having to decide on nonclinically relevant parameters reduces the uncertainty in the definition of the problem. Several optimized solutions for the same case are calculated by both relaxing and strengthening the dose volume constraints. By providing multiple optimized solutions, the authors´ clinical implementation of inverse planning assures the planner/physician that specifying a precise amount of OAR sparing is not critical. The planner/physician regains control of the planning process by reviewing several solutions instead of just one solution. The authors´ added dosimetric, conformity and uniformity index, and radiobiological, normal tissue complication probability and equivalent uniform dose, evaluation tools to help decide which plan is best for treating the patient