Repeat imaging of tumor hypoxia during radiochemotherapy with FMISO PET in head and neck cancer. Free

3501

Purpose: Tumor hypoxia is associated with treatment resistance and increased malignant potential. Positron Emission Tomography (PET) with [18F]-fluoromisonidazole (FMISO) has the ability to non-invasively quantify hypoxia in tumors. We investigated the value of repeat FMISO PET imaging as a predictor of treatment outcome in head and neck cancer (HNC).

Methods: Thirteen patients with locally advanced HNC, all treated with concomitant chemoradiotherapy, were enrolled in an ongoing PET protocol. Both FMISO and FDG scans were performed at baseline and an FMISO PET was repeated during treatment (week 5). All examinations were performed on a Biograph (Siemens) PET-CT scanner; FDG and FMISO were injected i.v. at 0.1 mCi/kg. FDG scans were acquired 60 minutes after injection; this was 3 hours for FMISO. All pixels showing a higher intensity than 50% of the maximum FDG uptake in the region of interest were included in the FDG avid volume (primary tumor and lymph nodes). The hypoxic volume (HV) was defined as the number of pixels with a tissue to blood (T/B) FMISO ratio, calculated using venous blood obtained during the emission part of the scan, of ≥ 1.4, indicating significant hypoxia.

Results: Median follow-up was 32 months; actuarial 2-year loco-regional control (LRC) was 42.5%. The mean FDG avid volume was 27.17 cm3 and the mean of the maximal FDG standard uptake value (SUVmax) was 8.5. The FDG avid volume was significantly correlated with LRC (p = 0.02) in univariate analysis, while the maximum FDG uptake had no prognostic significance. Significant hypoxia was identified in all baseline FMISO scans: the mean HV was 6.03 cm3 and the mean T/Bmax was 1.63. No statistically significant correlation between increased glucose metabolism and hypoxia was observed. In the second FMISO scan, significant hypoxia was observed in 9 patients: the mean HV was 0.62 cm3 and the T/Bmax was 1.55. LRC was significantly correlated with the HV before (p = 0.04) and during (0.01) treatment in univariate analysis. The median decrease in HV between the two time-points was 88.96% (range: 53.96% to 100%). When the patients were stratified according to the median decrease in HV, patients with less decline in hypoxia had a significantly worse 2-year LRC rate (0% vs. 62.5%, p = 0.026). In multivariate analysis, including clinical characteristics, decrease in HV (p = 0.04) and T classification (p = 0.03) were found to be the independent predictors of loco-regional control. When the FMISO scans at the different time-points were co-registered to compare the location of the hypoxic sub-volumes, there was only partial overlap between the two volumes (mean overlap 42.5%).

Conclusion: Both FDG and FMISO provide prognostic information on the examined tumor. Hypoxia is an important predictor of treatment outcome in HNC and the decrease of hypoxia during RT seems to be prognostically meaningful. Further validation of these findings in a larger patient group is ongoing.