There is concern about possible association between PUVA treatment and an increased risk of noncutaneous cancer. An alteration in the risk of cancer among persons with psoriasis has also been postulated. To test this hypothesis, for nearly two decades we have prospectively followed 1380 patients who first began PUVA treatment for psoriasis in 1975-1976. We compare the risk of noncutaneous cancer in our cohort with that expected based on general population incidence rates. The overall risk of noncutaneous cancer was nearly identical to that expected in general population. For three separate sites, we noted significant increases: thyroid cancer (RR = 3.57, 95% CI = 1.16-8.34), breast cancer (RR = 1.81, 95% CI = 1.19-2.64), and central nervous system neoplasms (RR = 2.80, 95% CI = 1.13-5.57). Since 1987, however, the risk of central nervous system neoplasms has not been elevated (RR = 0.00, 95% CI = 0.00-3.35) and the relative risk of breast cancer was lower than in the prior decade and not statistically significant. There was no association between higher levels of exposure to PUVA and the risk of any of these cancers. We did not detect any significant increase in the risk of lymphoma or leukemia. Our study does not support the hypothesis that long-term PUVA treatment increases the risk of noncutaneous cancer.