Determination and use of radiobiological response parameters in radiation therapy optimization

Abstract

Optimization of radiation therapy is critically dependent on the use of patient related information. For this reason, radiobiological models describing the dependence of tumour and normal tissue responses on the irradiated volume and the dose-time-fractionation schedule should be introduced. In addition, the heterogeneity of the delivered dose distribution and tumour or normal tissue sensitivity variations have to be taken into account clinically. In the present study, a treatment optimization procedure is used that considers the shape and the structure of the target tissues and healthy organs at risk, their relative position and their dose-response relations for the individual patient.

Mathematical models largely based on the Poisson statistics and the linear-quadratic model of cell kill, have been used to quantify the radiobiological response of normal human tissues and tumours to radiation therapy. The presented models predict a decreasing probability of achieving complication free tumour control with increasing tumour size and increasing volume of normal tissues irradiated. The radiobiological parameters D50, gamma, sigma and Vref of the Poisson and relative seriality models have been estimated for certain normal tissues and targets. The process for determining these dose-response relations was based on clinical materials where the treatment information and follow-up results of the individual patient were available. The statistical methods used, estimated and verified the parameters and their uncertainties. The clinical range of variability of the dose-response relations is important for their correct use in the clinical routine.

The clinical use of the derived dose-response relations is demonstrated using radiobiological parameters for different tumours and normal tissues that were also calculated based on data clinical trials. A biological evaluation procedure is introduced and applied on clinical cases. This procedure uses the biological models and dose-response data of the involved organs and optimizes the dose level of the treatment technique under study. This is done by evaluating the plan using the P+, objective, which estimates the probability to achieve tumour cure without having severe complications to the healthy tissues. The clinical value of biologically based treatment planning was compared with alternative physical criteria (e.g. tolerance doses) and with the judgment of personnel on particular clinical cases.

It is demonstrated that the radiobiological objective functions allow a much higher conformity and a more clinically relevant scoring of the treatment outcome. The probability of achieving tumour control without fatal complications in normal tissues is increased and the dose delivery optimized. Recent developments can reduce or even eliminate the need for intracavitary treatment by delivering more conformal dose distributions using intensity modulated external dose delivery. In these cases the reliability of the patient setup becomes critical for the effectiveness of the treatment. It is realized that accurate information concerning the response of different organs to fractionated intensity modulated radiation therapy is the key to true optimization of the delivered dose distribution.