Purpose: Hypoxia is a big challenge for tumor control in head and neck cancers. With advanced imaging techniques, individual patient hypoxic regions can be identified. Several clinical trials perform dose-escalation based on pre-treatment hypoxia, typically boosting hypoxic tumor areas. However, longitudinal imaging studies showed that the location of hypoxic cells changes throughout treatment. The purpose of this work is to quantify the potential loss of biological effect due to changes in hypoxia during treatment in dose-escalation studies based on initial hypoxia only and compare this strategy to one based on patient imaging at weeks 0 and 2 of treatment.
Methods: A previously published reoxygenation model based on LQ survival formulas with reoxygenation parameter Δ is extended to allow for dose-escalation with dose d(H) per fraction for initial hypoxic cells. The new model allows for some hypoxic cells to miss the escalated dose depending on parameter “m” (m=0 no missed cells, m=1 all cells missed). Using published radiobiological parameters and hypoxia reduction factor (HRF) of 1.5-2.5, we calculate the EQD2 when hypoxia is present for 70Gy delivered in 35 fractions with d(H)(2.2-2.6Gy) to the hypoxic region.
Results: For an HRF of 1.5, the biological effect with escalated dose per fraction up to 2.6Gy cannot sufficiently compensate hypoxia if there is no reoxygenation (Δ=0). As Δ is increased to 0.2 and beyond (20% or more hypoxic cells re-oxygenate after each fraction), the EQD2 becomes comparable to the non-hypoxic values. Based on previous results, the dose-escalation of the whole tumor after 20Gy based on hypoxia imaging at 2 weeks yields comparable or higher EQD2 values for more modest escalated doses to the whole tumor.
Conclusion: The new model allows to evaluate and compare strategies for dose- escalation to compensate for hypoxia during treatment and quantifies limitations of dose-escalation based on initial hypoxia.