The selection of patients for accelerated radiotherapy on the basis of tumor growth kinetics and intrinsic radio sensitivity

Abstract

Mathematical modeling was used to reach qualitative conclusions concerning the relative rate of local tumor control that might be achieved by using accelerated fractionation to treat only the patients with the most rapidly growing tumors, compared with the control rates that could be expected from either conventional or accelerated radiotherapy alone. The results suggest that concomitant boost therapy is equally or more effective than conventional dose fractionation for all tumors, regardless of their growth kinetics. For tumors with very short clonogen doubling times, CHART (continuous hyperfractionated accelerated radiotherapy) may be even more effective than concomitant boost treatment, but CHART is less effective than conventional or concomitant boost therapy for tumors with longer clonogen doubling times. Thus, there is a rationale for using a predictive assay of tumor clonogen doubling times to identify the patients who should be treated with CHART. However, improvements in local tumor control resulting from concomitant boost treatment or the selective use of CHART are not likely to be apparent in the population as a whole, because the overall control rates are largely determined by refractory tumors having little chance of control with any of the treatments and by highly responsive tumors that are likely to be controlled regardless of the treatment choice. Differences in control rates with different treatment strategies are most apparent in the stochastic fraction of the population, which excludes those patients for whom there is either very little chance (e.g. < 1%) or a very high chance (e.g. > 99%) of achieving local control with both treatments. The stochastic fraction can be approximated by excluding those patients with the most radioresistant and the most radiosensitive tumors, since intrinsic tumor radio sensitivity appears to be the single most important factor determining treatment outcome.