We read with interest the article by Cheng et al. addressing the antitumor effect of streptozocin and doxorubicin in the treatment of patients with advanced islet cell carcinoma.1 They concluded that they were unable to confirm the findings from the Eastern Cooperative Oncology Group trial published in 1992.2 In this trial of the 38 patients who received streptozocin and doxorubicin, 36 were evaluable and 69% showed any regression with 14% complete regression. For this study, criteria for response evaluation were a > 50% decrease in hormone production; a > 30% reduction in hepatomegaly; and measurable tumor reduction, with the product of the perpendicular diameters reduced > 50%. Routine use of computed tomography (CT) scan was not indicated. From this study it is in retrospect difficult to calculate how many patients had a response based only on hormone reduction or reduced liver size. Cheng et al. used CT scan, magnetic resonance imaging, or X-rays of bidimensional measurable lesions as the only criteria for response as defined by Miller et al.3 With the same streptozocin and doxorubicin regimen2 given to all of their 16 patients with islet cell tumors, Cheng et al. observed 1 objective response, 9 patients with stable disease, and 6 patients with progressive disease. The median overall survival was not yet reached, with > 60% alive after follow-up ranging from 10 to 67+ months. We would like to add six islet cell tumor patients to this series. From 1995–1999, we treated these patients according to exactly the same chemotherapy scheme. The patients were evaluated as described by Cheng et al. None of the patients were hormone producers. One patient experienced a partial response, three had stable disease, and two had progressive disease. Overall survival after initiation of chemotherapy ranged from 4 to 17+ months. Apart from cardiotoxicity (n = 1), which was most likely due to doxorubicin treatment (with cardiac failure responsive to antidiuretic treatment), no severe toxicity of the regimen was observed. Combining the data from Cheng et al. and ours results in 2 objective tumor responses out of 22 (9%; 95% confidence interval, 1–29%). Although we find with these criteria a lower response rate than Moertel et al., it is also true that in these two series a good effect on survival cannot be denied. In islet cell tumor studies, to report precisely on the objective regression of tumor size may be of major relevance. If there is a regimen that results in a realistic option to create a reduction in tumor size, this would potentially allow surgical tumor resection and thus increase the potential for curative tumor resection. If the main goal of this regimen is quality of life rather than remission, the objective response rate of only 9% with the streptozocin/doxorubicin regimen justifies more interest in interferon treatment of patients with islet cell tumors. After an initial study with leukocyte interferon,4 interferon-α doses of 5–6 MU 3–5 times per week resulted in biochemical responses in 50% of the patients and tumor responses in 12%.5 Surprisingly, this option is not even mentioned in the National Cancer Institute's PDQ information for health care professionals on islet cell carcinoma.6 Treatment with somatostatin analogs is mentioned in the PDQ, reporting improvement of symptoms but little to no effect on tumor size. Attempts to combine the several options have been initiated.5 Because of the rarity of the disease, it is difficult to conduct large studies to confirm earlier interesting results or to design studies to improve these results. It will be of major importance to consider future multicenter studies in order to improve the outcomes of patients in this category. H. de Vries*, N. H. Mulder*, E. G. E. de Vries*, * Division of Medical Oncology, Groningen University Hospital, Groningen, The Netherlands