To determine whether we could distinguish asbestos-related lung cancers from unrelated ones, we typed and quantified by electron-optical methods the asbestos fibers in the lungs of 75 men with lung cancer. All but eight men had some history of asbestos exposure. On the basis of combined amosite and crocidolite (AC) concentrations, we divided the subjects into three groups (AC fibers per gram of dry lung): low (less than 10(5)); intermediate (10(5) to 10(6)); and high (greater than 10(6)). Age, smoking history, latent period, and type and location of tumors were similar in all three groups. Of 62 evaluated subjects, zero of 14 in the low group, seven of 29 in the intermediate group, and five of 19 in the high group had asbestosis. Epidemiologic studies suggest that persons exposed to concentrations of asbestos that can cause asbestosis are at increased risk for lung cancer. Thus, the subjects in our intermediate and high concentration groups may have been at increased risk for cancer, even when they did not have asbestosis. Because large burdens of asbestos do not always cause pulmonary fibrosis, asbestosis may be a poor marker of fiber-related lung cancer.