Clinical prognostic factors for patients with recurrent head and neck cancer: implications for randomized trials

The purpose of our study was to determine the clinical prognostic factors for the duration of the overall survival from recurrence (OSR) in patients with recurrent head and neck squamous-cell carcinoma. We performed a retrospective analysis on 496 patients treated between 1982 and 1993 at the Antoine Lacassagne Center. The significant favorable prognostic factors for the OSR were: initial T(1-2) (p = 0.008), no initial nodal involvement (p = 0.002), no initial chemotherapy exposure (p = 0.002), induction chemotherapy response (p = 0.001), duration of disease-free interval (DFI; p = 0.0001), performance status (PS) 0-1 (p = 0.004) and local-regional recurrence (p = 0.001). In the multivariate analysis, the apparent nonsignificance of all factors apart from the DFI suggested that relevant prognostic factors could be embedded in the DFI. Multivariate analysis was performed after excluding the DFI. The results indicated that local-regional recurrence, PS 0-1 and no initial chemotherapy exposure remained significant favorable prognostic factors for the OSR. The advantages of taking into account such prognostic factors are to eliminate the patient selection bias and to ensure a fairer comparative evaluation of new or already existing agents in recurrent head and neck cancer.