Classification of three prognostically different groups of head and neck cancer patients based on their metabolic response to induction chemotherapy (IC-1)

Objectives: There exist no uniform decision criteria for conservative organ preservation treatments in head and neck cancer patients. Even with F-18-FDG-PET/CT after induction chemotherapy patient selection is challenging. This study correlated metabolic tumor response with treatment types and recurrence patterns. Materials and methods: Decrease in SUVmax in F-18-FDG-PET/CT was measured 21-28 days after IC-1 in 102 patients and correlated to cancer-specific endpoints. Results: Residual SUVmax (resSUVmax) values were uniformly distributed across five cut-off levels (0-0.2 vs. > 0.2-0.4 vs. > 0.4-0.6 vs. > 0.6-0.8 vs. > 0.8) containing 20%, 25% 25%, 15% and 15% of patients. Patients were stratified into three response categories according to residual SUVmax (Group A: 0-0.4 = high response Group B: > 0.4-0.8 = moderate response, Group C > 0.8 = non-response), 5-year local control rates were 90.5% (Group A) vs. 78.9% (Group B; univariate p = 0.07, multivariate: HR: 3.6, p = 0.03) vs. 49.4% (Group C vs. B; univariate p = 0.04, multivariate: HR 5.5, p < 0.01). After IC-1, Group A received chemoradiotherapy (CRT) only. Group B received surgery plus either (chemo)radiotherapy (B_S + RT/CRT) or chemoradiotherapy (B_CRT), yielding local control rates of 100% and 74.2% (p = 0.11). Group C received surgery plus CRT or CRT alone; both achieved equally poor local control (p = 0.71). Group C had significantly worse distant metastasis-free survival and overall survival than Groups A and B (p < 0.05). Conclusion: Metabolic response after IC-1 differentiates HNC patients into three subgroups predicting local tumor control. Non-response was associated with a poor outcome.