We have measured plasma calcitonin in 135 untreated eucalemic men with lung cancer and a control/smoker population. Calcitonin levels were determined by radioimmunoassay and validated by immunoextraction. Plasma immunoreactive calcitonin moieties were purified by immunoadsorbent chromatography, treated with mercaptoethanol and urea, and characterized by gel filtration. Artifacts in human calcitonin radioimmunoassays of cancer-patient plasmas were detected by parallel plasma incubations in a salmon calcitonin radioimmunoassay system which does not detect human calcitonin and by immunoprecipitation of tracer at the end of radioimmunoassay incubations. Heating fresh plasmas to 65 degrees C for 1.5 hours reduced radioimmunoassay artifacts without loss of calcitonin moieties. Such characterization of hypercalcitoninemia in each of the histopathological types of lung cancer has raised some important questions about the interpretation of plasma calcitonin radioimmunoassay measurements in lung cancer. Based on inhibition of tracer-antibody binding, plasma calcitonin seemed to be elevated in 18% (14/80) of basal plasma samples obtained from patients with epidermoid or with anaplastic lung cancer. Unequivocal hypercalcitoninemia (heat stable, causing no inhibition of antibody-tracer binding in the salmon calcitonin radioimmunoassays, and immunoextractable with human calcitonin antibodies) was not found in any of the apparently hypercalcitoninemic plasmas from persons with epidermoid or anaplastic lung cancer. By contrast, unequivocal hypercalcitoninemia was found in 27% (15/55) of plasmas from patients with small cell carcinoma or adenocarcinoma. Most of the immunoreactive calcitonin recovered from small cell and adenocarcinoma lung cancer plasmas with unequivocally elevated calcitonin is much larger than calcitonin monomer.