Identification of poor-risk patients under 55 years old with Hodgkin's disease at diagnosis-a neural approach

The aim of this study is to assess the feasibility of using objective data at the time of diagnosis of Hodgkin's disease to predict those patients who were likely to die of progressive disease within four years. A neural approach was to be adopted for subsequent comparison with the traditional method based on a statistically derived index for the disease. All patients are aged between 15 and 55 years old and their data forms part of a multicentre investigation within the Scotland and Newcastle Lymphoma Group (SNLG) in the UK. The neural network first analysed 118 patients from the Newcastle centre alone. Outcome prediction was based on patient's age, clinical stage, lymphocyte count, haemoglobin level and bulk disease information. The overall classification rate in this case was 81.4%. Consequently, a blind analysis was carried out on 109 patients from the Edinburgh centre, where an overall classification rate of 82.6% was achieved. However, a crucial aim of the study was to compare the neural classification with the SNLG index in the case of poor-risk patients who had a low index at diagnosis.