Factors Affecting Long-Term Survival in Patients with Recurrent Head and Neck Cancer May Help Define the Role of Post-Treatment Surveillance

Objectives/Hypothesis: Assess impact of clinical variables upon survival in patients diagnosed with recurrent head and neck carcinoma and better delineate role of post-treatment surveillance. Methods: Variables including previous tumor site, stage, prior therapy, location of recurrence, patient-identified symptoms, compliance with recommended surveillance, definitive salvage therapy, and clinical outcome were recorded for patients diagnosed with recurrent head and neck cancer from which estimates of survival were generated and compared. Results: One hundred five patients were identified as having been diagnosed with recurrent disease. The majority (81%) were deemed compliant with recommended surveillance. Eighty-nine patients (85%) identified new symptoms/findings prior to diagnosis. Survival was not significantly different between compliant and noncompliant patients (P = .20). For patients who underwent salvage therapy, no significant differences in survival were seen according to original primary site or mode of treatment for recurrence. Better survival was seen in patients with original prior early stage disease (P = .0001) and in patients with local-only site of disease recurrence (P = .0001). Conclusions: In patients with recurrent disease, survival may relate more to variables associated with prior disease or recurrence location than those associated with follow-up surveillance. Surveillance and salvage therapy for recurrent disease may, however, preferentially benefit certain subgroups: 1) patients with prior early stage disease, and 2) those in whom site of recurrence is local-only. Despite high compliance with recommended surveillance, survival remains extremely poor for patients with recurrent disease previously treated for advanced stage disease or with those with regional recurrence. As such, the impact of routine surveillance upon survival in these patients remains unclear.