EDITOR—Studies in pancreatic cancer, such as that of Greenway,1 are difficult to recruit into, and a review of the literature will confirm that they are unusual. During the past three years two other national studies in advanced pancreatic cancer have been done (unpublished data). These were much larger, and histological or cytological proof was an entry requirement. This was not a requirement for inclusion in Greenway's study. Only 17 of the 49 patients entered into Greenway's study had confirmation of their disease, 12 in the flutamide group and five in the placebo group. Thirty two patients had evidence of only local disease at entry into the study; 11 of these had open surgery either before or after they entered the study. Presumably, histological proof of diagnosis was obtained in this group, which leaves a further six patients who had cytological proof of diagnosis obtained from ascitic fluid.

Two patients in the group treated with flutamide survived for over three years, compared with a maximum of just over one year in the group treated with placebo. This length of survival in someone with advanced ductal adenocarcinoma of the pancreas is unusual and brings into question the original diagnosis. Greenway comments that three of the patients in the flutamide group with histological or cytological proof of diagnosis were the longest survivors in the trial. We would be interested to know the exact diagnostic details of the two patients surviving for three years and whether they remain alive and well. It is also disappointing that there is no significant benefit in using the generalised Wilcoxon test when the whole group is analysed. Although use of the log rank test gives a significant result, this test gives greater statistical weight to differences in treatment at later time points. The two patients with longest …