Management of malignant teratoma: does referral to a specialist unit matter?

The causes of geographical differences in cancer survival among regions of the UK are unclear. Population-based audit of management of patients with non-seminomatous germ-cell tumours (NSGCT) in the west of Scotland enabled us to assess the relative contributions to outcome of recognised prognostic factors, treatment centre, and protocol treatment. Data on treatment and outcome were analysed for 440 (97%) of 454 men with NSGCT diagnosed between 1975 and 1989. All but 11 patients were treated at tertiary referral centres; 235 were treated at a single unit (unit 1) and 194 at four other units (2-5). 99 men have died, 89 (20%) from NSGCT. Independent prognostic factors for NSGCT survival were extent of tumour at diagnosis (p < 0.001), 5-year period of diagnosis (from 1975-79 to 1985-89, p < 0.001), and treatment unit (unit 1 vs units 2-5, p < 0.001). Unit 1, which had the best survival rates, treated most patients overall (53%), including the majority (70%) in the worst prognostic category (poor-prognosis metastatic disease). The proportion of men receiving nationally agreed protocol treatment was higher at unit 1 than elsewhere (97 vs 61%, p < 0.0001). However, analysis restricted to men who received protocol treatment, adjusted for other important prognostic variables, still showed a survival advantage for this unit (relative death rate units 2-5 vs unit 1, 2.82 [95% CI 1.53-5.19], p < 0.001). These findings suggest that centralisation of treatment for NSGCT improves outcome; the benefit seems to be additional to any advantage resulting from protocol treatment.