This study was designed to evaluate dosimetric, pharmacokinetic, and other treatment-related parameters as predictors of outcome in patients with advanced B-lymphocytic malignancies. Fifty-seven patients were treated with radiolabeled Lym-1 antibody in early phase trials between 1985 and 1994. Logistic regression and proportional hazards models were used to evaluate treatment parameters for their ability to predict outcome, taking into account patient risk group based on Karnofsky performance status and serum lactic dehydrogenase. The occurrence of a partial or complete response (31 of 57 patients) and development of human antimouse antibody (HAMA) predicted improved survival using a time-dependent proportional hazards model. The final multivariate model for survival with parameters significant at P ≤0.05 included overall response and pretreatment risk group. Although some of the dosimetric and pharmacokinetic parameters were predictive in univariate analyses, only longer half-time of radionuclide in the blood showed any indication of improved prediction beyond that provided by the lactic dehydrogenase/Karnofsky performance status-based risk groups. Splenic volume, splenectomy, and malignant tissue Lym-1 reactivity were not contributory. In this patient group, the effect of radiolabeled Lym-1 treatment as indicated by measurable tumor response was associated with improved survival. Development of HAMA was also associated with improved survival, indicating that concern about HAMA should not preclude exploration of radioimmunotherapy. Although dosimetry has a role in determining safety based on dose to normal organs, when adjusted for baseline clinical features, dosimetric and pharmacokinetic parameters showed limited ability to improve outcome prediction.