Although the existence of subclinical Pneumocystis curinii infection in pediatric patients with solid tumors or hematologic malignancies has been documented, similar data are lacking in adults. In addition, data are needed to define the epidemiology of this agent in adult malignancies to assess the validity of the methodology employed in antigen detection, and to elucidate the value of these methods in the diagnosis, prophylaxis, and prognosis of P curinii infection in adults with cancer. The study was designed to determine the incidence of P curinii antigenemia in ambulatory patients with solid tumors or hematologic malignancies. The authors also sought to determine if antigenemia a s detected by a counterimmunoelectrophoresis test correlated with any clinical parameter. Patients included in the study were ambulatory, asymptomatic, afebrile, adult cancer patients seen in the clinic for follow-up or treatment. Coded sera were electrophoresed against high-titered rabbit antiserum to P curinii organisms. Two hundred fortyseven patients were studied, including 172 hematologic malignancies (average age, 57 years), 109 men and 63 women; 75 solid tumors (average age, 55 years), 39 women and 36 men. One hundred three healthy adults served as controls. Only five patients had positive antigen (2%). All of these patients had hematologic malignancies and were women. None of the control sera were antigen-positive. We conclude that the incidence of P curinii antigenemia in asymptomatic adults with neoplastic disease is extremely low. A positive P curinii antigen in the absence of clinical symptoms most likely represents subclinical infection. Positive antigen does not always indicate active disease, but probably reflects mobilization of antigen during generalized inflammatory response or possible pulmonary insult. In making the decision to treat consideration should be given to clinical presentation and history.