Both conformal and intensity-modulated radiation therapy have great potential to further increase tumor control rates and decrease morbidity. A homogeneous escalation of 'biological' dose within a tumor should increase the likelihood of local cure, especially within the mid-range (e.g. 15% to 80%) of tumor control rates, and conversely, a lower control rate should follow a homogeneously reduced dose. However, when the dose to critical normal tissues is tightly contrained, the dose distributions within the treatment volume may necessarily be heterogeneous. and the effect on tumor control probability will depend upon the magnitude of over- or underdosage, and on the proportions of the tumor clonogen population receiving higher or lower than the nominal dose. Dose-volume histograms provide a measure of heterogeneity of dose within the planned treatment volume, but tumor control probability is also influenced by other variables, e.g. inherent tumor clonogen radiosensitivity and growth rates during a course of treatment, alpha/beta ratios, oxygenation and clonogen density throughout the target volume. Heterogeneity in these factors introduces heterogeneity in tumor responses and a less steep change in tumor control probability with change in dose, reducing the gains or losses that would be predicted to result from heterogeneity of dose. Similarly, modeling the effect of inhomogeneous dose distributions on estimates of probability of complications in normal tissues is hindered by uncertainty of estimates for alpha/beta ratios, especially for late-responding tissues, and lack of data on volume effects. Although the effects of dose inhomogeneity cannot be presented with sufficiently reliable quantitation to be directly applicable to dose prescriptions in radiation therapy, the relative influences of heterogeneities in dose and volume can be modeled to provide a Framework for clinical decision-making. The magnitude of a dose reduction is the major determinant of decline in tumor control probability. A large dose reduction to even a small volume of tumor can profoundly decrease tumor control probability. Conversely, the most rapid improvement in tumor control probability occurs the closer to 100% the amount of tumor exposed to an increased dose. Escalation of dose is of little value unless it is distributed through most of the tumor: even very large increases in dose to small volumes are of little benefit.
