In order to evaluate the clinical benefit of Nocardia rubra cell wall skeleton (N-CWS), a randomized controlled study was performed with inoperable lung cancer patients entered in 5 institutions from October 1978 to June 1981. Patients without pleural effusions were treated initially with conventional therapies such as chemotherapy and/or radiotherapy, according to common protocol, and then patients in performance statuses 0 to 3 were randomized into control and N-CWS groups with stratification into 16 categories according to 4 histological types and 4 clinical stages In the N-CWS group, 400 micrograms N-CWS were initially injected once or twice into the bronchial tumor using a fiberoptic bronchoscope, and subsequently 200 micrograms of N-CWS were injected at monthly intervals into the skin from the shoulders to upper arms. Of 309 patients, 118 patients in the N-CWS group and 108 patients in the control group were eligible for statistical analysis. There was statistically no significant difference in survival rate between the control and the N-CWS groups. According to histological type, significant prolongation of the survival period was observed in patients with small-cell carcinoma. The 97 patients with pleural effusions were initially randomized into control and N-CWS groups. In the control group, local chemotherapy with Adriamycin was performed and, in the N-CWS group, local administrations and monthly intracutaneous injections of N-CWS were given. Tube thoracostomy was performed in both groups. The local response rate was statistically greater in the N-CWS group than in the control group, and survival period was also prolonged significantly in the N-CWS group. The main adverse reactions to N-CWS were skin lesions in the injected sites and fever, but these were temporary and not serious.