One hundred and fourteen patients with clinical or pathological stages IIIB and IV Hodgkin's disease have been treated with MVPP chemotherapy followed by radiotherapy to sites of previously bulky disease. The minimum follow-up is 2 years with a median of 5 years. The overall remission rate was 92 per cent with 74 per cent achieving CR. A discriminant analysis showed that the presence of bulky disease was the only independent factor that predicted a lower chance of CR (66 per cent vs 82 per cent, P=O*O45). The 5-year survival was 70 per cent overall and 85 per cent for CR patients. A Cox multivariate analysis demonstrated that stage 111 disease, age less than 36 years and female sex were all variables which independently predicted a more favourable prognosis in terms of overall survival. However a similar analysis for survival of CR patients showed that age less than 36 years and the absence of bulky disease were the only two factors to independently predict a more favourable outcome. Of 14 patients who did not receive the chemotherapy according to protocol 5 have relapsed compared to 7 of 70 who did. A Cox analysis confirmed that this variable was the only one of significant prognostic import in predicting relapse-free survival. Using the Cox analyses we have been able to devise a scoring system which accurately predicts the outcome for these patients. This model may be useful in determining which patients have a worse prognosis following treatment with MVPP thus allowing more intensive therapy to be given to such patients while minimising treatment for those with favourable features.