Active immunotherapy with living BCG was conducted on 98 patients with various types of cancer. The candidates for this therapy were patients with residual or inoperable cancer of the colorectum, liver, breast, biliary tract, lung, and other organs with a follow-up of 4-58 months. Eleven of the 98 (11%) were able to survive for as long as 37-58 months (mean survival time 42.5 months) because of this treatment and are still living. Another 11 patients are also alive more than 24 months after starting treatment. Thirty-seven patients, however, succumbed within 12 months despite BCG immunotherapy. On the other hand, 37 patients in the control group, who shared the same clinical status and did not receive BCG therapy during this period, underwent unhappy courses for 2-12 months (mean survival time 8.7 months). The pretreatment immunoresponsiveness of these 98 patients was suppressed, as measured by the following immunologic parameters: T-cell subpopulation in the peripheral blood, stimulation index of PHA, and skin tests to DNCB, KLH, PPD, and PHA. All of these parameters improved shortly after initiation of BCG injections in 22 patients who survived more than 24 months. In contrast, in patients who died within 12 months, immunoresponsiveness remained suppressed throughout the course. This result has suggested that there was an apparent correlation between the effectiveness of BCG and immunoresponsiveness. In addition, a good correlation was observed between the duration of inflammatory reactions at BCG injection sites and clinical prognoses. Moreover, it was shown that a relatively high amount of BCG (20-80 mg as an initial dosage) and repeated injections of living BCG were necessary to obtain a sufficient enhancing effect on the immunocompetency of these late-stage cancer patients. The most conventional criterion used to determine an optimal time for booster injections of BCG was measurement of the PPD-evoked skin reaction at the BCG injection site, that is, Koch's phenomenon. When a marked flare-up reaction of more than 2.5 X 2.5 cm in size was observed, the effect of BCG was considered to be continuing, and no additional booster injection was needed. The mean interval between the first and second BCG injections was 6.2+/-1.1 months in patients who survived more than 2 years. In contrast, the duration of this reaction was only transient in ineffective cases. The most frequent side effects of this therapy were fever and malaise; these complications occurred in 62% of the cases. No severe side effects, such as dissemination, anaphylactic shock, or granulomatous hepatitis, have been experienced throughout this study, even in patients to whom a total dosage of more than 200 mg of living BCG were injected.