Revising the radiobiological model of chemoradiation in head and neck cancer


Year:

Session type:

Sara Meade1, Paul Sanghera1, Christopher McConkey2, Jack Fowler3, John Glaholm1, Andrew Hartley1, George Fountzilas4
1Hall-Edwards Radiotherapy Research Group, Queen Elizabeth Hospital, Birmingham, UK, 2Clinical Trials Unit, University of Warwick, Warwick, UK, 3Department of Human Oncology and Medical Physics, University of Wisconsin, Wisconsin, USA, 4AHEPA Hospital, Aristotle University of Thessaloniki, Thessaloniki, Greece

Background

Previous studies of synchronous chemoradiotherapy have modelled the additional effect of chemotherapy as additional radiotherapy biologically effective dose (BED). Recent trials of accelerated fractionation chemoradiation versus conventional fractionation chemoradiation have cast doubt on such modelling. The purpose of this study was to identify alternative models.

Method

Nine trials of platinum based chemoradiation were identified. In radiotherapy alone arms, the radiotherapy BED for tumour (tBED) was calculated using standard parameters. In chemoradiation arms three methods were used to calculate tBED; Additional BED: Addition of 9.3Gy BED for tumour to the radiotherapy BED; Zero repopulation: BED calculated with no correction for repopulation; Variable tp: Values of tp 3-10 were used to examine a partial suppression of repopulation. The correlations between the observed and predicted difference in outcome were assessed using the Pearson product moment correlation weighted by the number of patients in each study group.

Results

For local control significant correlations were obtained for all three methods but the correlations were stronger with zero repopulation (p=0.0002) and variable tp (tp =10) (p=0.0005) than additional BED (p=0.02).

Conclusion

Radiobiological models using modified parameters for accelerated repopulation appear to correlate strongly with outcome in chemoradiation studies. The variable tp method shows strong correlation for outcome in local control and is potentially a more suitable model in the chemoradiation setting. However, a lack of trials with an overall treatment time of more than 46 days inhibits further differentiation of the optimal model.