A critical appraisal of the usefulness of some biological parameters in predicting tumour radiation response of human head and neck cancer.

Tumour stage is considered to be a valuable prognostic parameter. However, in our experience with head and neck tumours, previous reduction of the tumour stage and/or size by chemotherapy does not affect the response to radiotherapy. Furthermore, we have found no clear correlation between response to treatment and kinetic parameters such as growth rate, generation time, growth fraction and cell loss. Recently, using cytometric analysis of biopsies, we concluded that the presence in the tumour of a single dividing "diploid" population vs subpopulations with more than one "ploidy" and different growth characteristics has little predictive value. At present, the possibility of selecting radioresistant tumours for a particular modality of treatment can only rely on empirical grounds. Holstsi et al. (1978) have proposed the use of (an) initial large fractional radiation dose(s) in tan attempt to exploit the possible increase in cell killing and/or reoxygenation. Following this approach, 34 patients with advanced or recurrent tumours have been irradiated with an initial single dose of 8-10 Gy. After 10 days of rest, the tumour shrinkage was estimated and the tumours classified as responders or nonresponders. When the patients underwent the remaining part of treatment according to a conventional fractionation, 2/3 to 3/4 of the responders exhibited a complete tumour shrinkage while none of the nonresponders exhibited a complete response. We feel that this approach could be an interim method of empirically identifying radioresistant tumours.