The objective of the present study was to investigate the association between human T-lymphotropic virus type-1 (HTLV-1) infection and cancer risk in a longitudinal study. The study population consisted of 2729 atomic bomb survivors in Nagasaki Prefecture, Japan, who had no previous history of cancer at baseline. The baseline survey, including analysis of antibody to HTLV-1, took place during 1985-1987 and follow-up was performed until the end of 2001. There were 553 incident cases of malignant neoplasms during the observation period. After adjustment for sex, age and other variables, HTLV-1 infection was not associated with the risk of developing cancers of all sites, excluding adult T-cell leukemia (rate ratio 1.0, 95% confidence interval [CI] 0.76-1.4), stomach, colon and rectum, lung, female breast or other minor sites, but was associated with increased risk of liver cancer (rate ratio 2.1, 95%CI 1.0-4.6). The point estimate of the rate ratio for thyroid cancer was 3.0, but this was not significantly higher than 1 because of the small number of events (n = 11) and low prevalence of HTLV-1 seropositivity. These findings support the idea that HTLV-1 infection is not associated with an increased general cancer risk. Confounding by hepatitis C virus (HCV) and the interaction between HTLV-1 and HCV may explain the increased risk of liver cancer among HTLV-1 carriers. Further follow-up may be required to determine if HTLV-1 carriers are at increased risk of thyroid cancer.