Weekly dose-volume parameters of mucosa and constrictor muscles predict the use of percutaneous endoscopic gastrostomy during exclusive intensity-modulated radiotherapy for oropharyngeal cancer

Purpose: To define predictors of percutaneous endoscopic gastrostomy (PEG) use during intensity-modulated radiotherapy (IMRT) for oropharyngeal cancer. Methods and Materials: Data for 59 consecutive patients treated with exclusive IMRT at a single institution were recovered. Of 59 patients, 25 were treated with hyperfractionation (78 Gy, 1.3 Gy per fraction, twice daily; "HYPER"); and 34 of 59 were treated with a once-daily fractionation schedule (66 Gy, 2.2 Gy per fraction, or 70 Gy, 2 Gy per fraction; "no-HYPER"). On the basis of symptoms during treatment, a PEG tube could have been placed as appropriate. A number of clinical/dosimetric factors, including the weekly dose-volume histogram of oral mucosa (OM DVHw) and weekly mean dose to constrictors and larynx, were considered. The OM DVHw of patients with and without PEG were compared to assess the most predictive dose-volume combinations. Results: Of 59 patients, 22 needed a PEG tube during treatment (for 15 of 22, ≥3 months). The best cutoff values for OM DVHw were V9.5 Gy/week <64 cm ³ and V10 Gy/week <54 cm ³. At univariate analysis, fractionation, mean weekly dose to OM and superior and middle constrictors, and OM DVHw were strongly correlated with the risk of PEG use. In a stepwise multivariate logistic analysis, OM V9.5 Gy/week (≥64 vs. <64 cm ³) was the most predictive parameter (odds ratio 30.8, 95% confidence interval 3.7-254.2, p = 0.0015), confirmed even in the no-HYPER subgroup (odds ratio 21, 95% CI 2.1 confidence interval 210.1, p = 0.01). Conclusions: The risk of PEG use is drastically reduced when OM V9.5-V10 Gy/week is <50-60 cm ³. These data warrant prospective validation.