The article by Riou et al,1 published in the October 1995 issue of the Archives, was interesting and provides potentially useful data for clinicians. Unfortunately, it seems that the authors incorrectly analyzed the data and overestimated the rate of other cancers in patients with melanoma.
The authors stated, "The incidence of lymphoma in our melanoma series was 12 of 664 (followed up over a period of 41 months)...."1(p1059) However, if I understand correctly, 6 of the lymphoma cases occurred before the melanomas, not during the 41 months of follow-up. It, indeed, seems that all the observed incidence rates per 100 000 in Table 21 were calculated as though the full number of observed cancers occurred during the 41 months of follow-up. In fact, fewer than half of them did.
Using the authors' calculations (Table 21), the risk of any kind of cancer in this group is