5-[(2-Aminoethyl)amino]-2-[2-(diethylamino)ethyl]-2H-[1]benzothiopyra no[4,3,2-cd]-indazol-8-ol trihydrochloride (CI-958) is the most active member of a new class of DNA intercalating compounds, the benzothiopyranoindazoles. Because of its broad spectrum and high degree of activity as well as a favorable toxicity profile in preclinical models, CI-958 was chosen for further development. The Phase I study described here was undertaken to determine the toxicity profile, maximum tolerated dose, and pharmacokinetics of CI-958 given as an i.v. infusion every 21 days. Adult patients with advanced refractory solid tumors who had adequate renal, hepatic, and hematological function, life expectancy, and performance status were eligible for this study. Written informed consent was obtained from all patients. Patients received a 1- or 2-h infusion of CI-958 at 21-day intervals. The starting dose was 5.2 mg/m2, and at least three patients were evaluated at each dose level before proceeding to a new dose level. A pharmacokinetically guided dose escalation design was used until reaching a predetermined target area under the plasma concentration versus time curve (AUC), after which a modified Fibonacci scheme was used. Forty-four patients (21 men and 23 women; median age, 59 years) received 162 courses of CI-958. Neutropenia and hepatorenal toxicity were the dose-limiting toxicities, which defined the maximum tolerated dose of CI-958 to be 875 mg/m2 when given as a 2-h infusion every 21 days. There were no tumor responses. Two patients had stable disease for >250 days. The recommended Phase II dose is 560 mg/m2 for patients with significant prior chemotherapy and 700 mg/m2 for patients with minimal prior chemotherapy. Pharmacokinetic analysis of plasma and urine concentration-time data from each patient was performed. At the recommended Phase II dose of 700 mg/m2, mean CI-958 clearance was 370 ml/min/m2, mean AUC was 33800 ng-h/ml, and mean terminal half-life (t1/2) was 15.5 days. The clearance was similar at all doses, and plasma CI-958 AUC increased proportionally with dose, consistent with linear pharmacokinetics. The percentage reduction in absolute neutrophil count from baseline was well predicted by AUC using a simple Emax model. The pharmacokinetically guided dose escalation saved five to six dose levels in reaching the maximum tolerated dose compared with a standard dose escalation scheme. This may represent the most successful application to date of this dose escalation technique.