Predicting late normal tissue responses.

One goal of radiation research is to predict, in an individual patient, the response of the cancer and relevant normal tissues to a course of radiotherapy and to use such information in adjusting the individual's risk-benefit ratio. There are four potential types of predictors of radiation response: physical dose, biological dose, risk factors and realtime assays. Biological dose, which is derived from physical dose and tissue radiobiology, is likely to be the most useful predictor for late normal tissue injury, although risk factors are also important. For any real-time assay to be useful in predicting ultimate response to a conventional regimen, for example, of 30 fractions, it must be both accurate and precise, especially if used after only one dose fraction. No sufficiently accurate such assays exist and, in any event, they are less critical for normal tissues than for tumors, because normal tissues are less heterogeneous in their cellular composition and radiation responses than tumors. In order to apply recent concepts in biological dosimetry to late normal tissue responses, it is important to determine "tolerance" doses of human tissues more accurately; to understand and quantify the volume effect; and to determine isoeffect relationships for all tissues at doses per fraction less than 2 Gy.