Available summaries of radiology reports were examined among 899 primary brain cancer cases (age 40 years and older) diagnosed in Connecticut residents in selected years from 1965 to 1988 and reported to the population-based Connecticut Tumor Registry. Adjustment for the lower sensitivities of radiologic tests used before the advent of computerized tomography (CT) suggested that the introduction of CT (by itself) could account for little of the secular increase in brain cancer rates. Examination of trends in age-standardized rates for histologically confirmed brain cancers in the elderly, after excluding those diagnosed only by radiologic tests, did not support the idea that the secular trend in the elderly was largely artifactual.