Neck dissection for non-squamous malignancy.

Of 1030 patients who underwent neck dissection (radical, modified or selective) in a 27-year period 103 had malignant neck nodes from a primary site in the head and neck with a histological diagnosis other than squamous carcinoma. There were 71 men and 32 women in this group with a mean age of 55 years. 28 patients had neck dissection as part of their initial treatment and 75 for later nodal recurrence. Five-year survival was 52% (40-63%). Survival was site dependent, best for thyroid tumours and worst for tumours of the major salivary glands (chi 1(2) = 6.52, P < 0.05). Histology significantly affected survival, best for papillary tumours and worst for melanoma and undifferentiated tumours (chi 1(2) = 3.85, P < 0.05). Survival was worse with advanced N stage but varied little with node level. The number of nodes invaded had a highly significant effect on survival (chi 4(2) = 23.94, P < 0.001), but extracapsular rupture had no effect. Advanced T stage at the time of surgery had a significant adverse effect on survival using univariate analysis, but this effect disappeared using multivariate analysis. In the 75 patients who had neck dissections for nodal recurrence the presence of a simultaneous recurrence at the primary site had no significant effect on survival. These patients had a better 5-year survival than patients having neck dissection for squamous disease, but the usual predictors of survival in squamous carcinoma do not always apply to non-squamous malignancy.