We read with interest the report of De SanJose et al., 1 in which the authors performed a retrospective seroepidemiologic study to determine the presence of polyomavirus SV40 antibodies using an enzyme immunoassay (EIA) in patients with lymphoma in Spain. The authors reported detection rates of 5.9% for SV40 antibodies in patients with lymphoma and 9.5% for controls. Based on these preliminary results, the authors concluded "there was no serological evidence of widespread SV40 infection and no association of SV40 seropositivity with human lymphomas in Spain". It is important to point out that the Institute of Medicine of the National Academies recently concluded "that the biological evidence is strong that SV40 is a transforming virus" and "that the biological evidence is of moderate strength that SV40 exposure could lead to cancer in humans under natural conditions". 2 These conclusions by the Institute of Medicine are in agreement with the recent analyses 3 of controlled molecular, pathologic and clinical data of 1,793 cancer patients that indicated a significant excess risk of SV40 associated with non-Hodgkin's lymphoma [odds ratio (OR) ฯญ 5.4, 95% confidence interval (CI) 3.1-9.3] as well as primary brain cancers (OR ฯญ 3.8, 95% CI 2.6 -5.7), primary bone cancers (OR ฯญ 24.5, 95% CI 6.8 -87.9) and malignant mesothelioma (OR ฯญ 15.1, 95% CI 9.2-25.0). Experimental data indicate that SV40 may be etiologically meaningful in the development of a specific subset of human cancers; results include expression of SV40 mRNA and/or T-ag in cancer cells, integration of SV40 sequences in some cancers and complex formation of SV40 T-ag protein with p53 and pRb in some tumor specimens. [2][3][4] These are important findings because SV40 oncogenesis is mediated in large part by T-ag and its effects on tumor-suppressor proteins. [2][3][4] While De SanJose et al.