This study includes 155 patients with T1–4N0 carcinoma of the postcricoid region seen between 1963 and 1993. Sixty-seven were treated by primary surgery, 50 by primary irradiation therapy, 36 were unsuitable for curative treatment and two patients were lost to follow-up. Reasons for deciding against curative therapy were: advanced age, poor general condition and advanced disease at the primary site. This study included only those patients who had no neck node metastases at presentation.
Patients receiving surgery tended to be in better general physical condition and tended to have more advanced disease than those treated by irradiation in this series. The tumour-specific five-year survival rate for those treated by surgery was 43 per cent (95 per cent confidence interval (CI) 23–60 per cent). For those patients treated by irradiation the five-year survival rate was 48 per cent (95 per cent CI 27–66 per cent) and for those receiving no treatment the median survival rate was three months (95 per cent CI two-six months). The observed survival for the surgery group was only 18 per cent and for the radiotherapy group 25 per cent at five years.
Multiple logistic regression showed no significant difference in proportions of host and tumour factors between the group receiving radiotherapy and the group receiving surgery. Recurrence at the primary site and the appearance of neck node metastases were not predicted by any host or tumour factor.
Twenty-one patients out of 67 receiving primary surgery had recurrence at the primary site compared with 26 patients out of 50 receiving primary irradiation. Neck node metastases occurred in 16 out of 67 patients receiving surgery and in eight out of 50 receiving radiotherapy. The difference was statistically significant for recurrence at the primary site (X21 = 4.261; p = 0.039) but not significant for neck node metastases (X 21= 0.661;p = 0.416). The data were further analysed using Cox's proportional hazards model for survival and no host or tumour factors were found to be predictive of eventual outcome apart from poorly differentiated histology. This adversely affected survival (X21 = 6.4444; p = 0.011). If patients not treated were included in the model, treatment became a significant factor in improving the survival (X21 = 4.4197; p = 0.034).
Radiotherapy appears to be at least as good as surgery for treating patients with an early carcinoma of the postcricoid region. We would recommend radiotherapy is used in patients with no detectable neck node metastases and in tumours <5 cm long. The complication rate from radiotherapy was reduced when compared with that of surgery.