Correlative Analyses between Tissue-Based Hypoxia Biomarkers and Hypoxia PET Imaging in Head and Neck Cancer Patients during Radiochemotherapy: Results from a Prospective Trial

Abstract

Tumor hypoxia influences the response of head and neck cancer (HNSCC) patients to radiotherapy and is associated with reduced outcomes. While both tissue-based hypoxia biomarkers and hypoxia PET imaging have previously been correlated with patient outcomes, the goal of this analysis was to evaluate potential correlations between biological markers and longitudinal biological imaging in a prospective trial of HNSCC patients. Forty-nine patients with locally advanced HNSCCs scheduled for definitive chemoradiation were enrolled in this prospective trial. They underwent baseline biopsies and [18F]FMISO PET imaging at baseline and at weeks 2 and 5 during treatment as well as [18F]FDG PET imaging at baseline. Immunohistochemical analyses for p16, Ki67, CD34-dependent microvascular density, hypoxia-inducible factor 1 α (HIF1α), carbonic anhydrase 9 (CAIX), Ku80 and CD44 were performed, and HPV status was assessed from tumor biopsies. Biomarker expression was quantified by the H-score and correlated with biological imaging information (Pearson product-moment correlation coefficient) and patient outcome data (Cox regression analysis). High HIF1α H-scores in the tumor were found to significantly correlate with increased tumor hypoxia during the first weeks of treatment as assessed by the difference in the FMISO tumor-to-background ratios (TBR) between baseline and week 2 (r=0.35; p<0.05) and deferred decrease in FMISO uptake between weeks 2 and 5 (r=0.35; p<0.05). Similarly, high CAIX H-scores were associated with a delayed reduction in FMISO TBR between weeks 2 and 5 (r=-0.39; p<0.05). Clinically, patients with high expression of CAIX exhibited significantly increased loco-regional recurrence rates after chemoradiation (hazard ratio 2.80 [1.08 – 7.31]; p<0.05). Additionally, loco-regional recurrence was associated with high expression of the DNA repair factor Ku80 in the tumor, although statistical significance was not reached. HPV status did not correlate with any of the tested biomarkers, and HPV-positive patients showed lower loco-regional recurrence rates (hazard ratio 0.13 [0.02 – 0.99]; p<0.01) and overall survival (hazard ratio 0.03 [0.001 – 1.73]; p<0.01) independent of their hypoxia dynamics and expression of tissue-based hypoxia biomarkers. In this trial, high expression of the tissue-based hypoxia biomarkers HIF1α and CAIX at baseline correlated with adverse hypoxia dynamics in HNSCCs during chemoradiation as assessed by PET imaging, and high CAIX levels were associated with increased loco-regional recurrence rates. Hence, tissue-based hypoxia biomarkers may be useful as predictors of hypoxia-associated radiation resistance and response of HNSCCs and warrant further investigations.