Mortality and cancer incidence among 999 neurological patients exposed to chronic alpha-particle irradiation from 232ThO2 (Thorotrast) administered for cerebral angiography in 1935-1947 was compared to that of 1480 similar patients examined with cerebral angiography without Thorotrast in 1946-1963 (controls). The ratio of standardized mortality/morbidity ratios (SMRs) for patients exposed to Thorotrast and controls was the relative risk, which was analyzed by multiplicative regression (RRREG). For mortality from all causes, RRREG was significantly increased mainly due to cancer, benign and unspecified tumors and benign liver conditions, while RRREG for all other causes combined was not significantly increased. The RRREG was significantly associated with the injected amount of Thorotrast for cancer and for benign liver conditions, while no other dependence on the amount of Thorotrast was seen. Cancer incidence was significantly increased, caused mostly by liver cancer, leukemia, metastases and cancer at unspecified sites. The risk for cancer other than liver, hematological, brain, metastases and cancer at unspecified sites combined was also significantly increased, but the temporal trend of RR for this category of sites did not indicate a radiation effect. A significant association between relative risk of cancer and injected amount of Thorotrast was largely accounted for by liver cancer. Thus increased mortality among patients exposed to Thorotrast is due mainly to cancer and benign liver diseases, and increased cancer incidence is caused by high risks of liver cancer, leukemia and ill-defined types (metastases, etc.). The time trend and dependence of the amount of Thorotrast injected do not support that cancer at most other sites is related to radiation from Thorotrast.