clinically indicated. The indicadon(s) for treatment, dosage, and dates of treatment will be 
recorded on the study flow sheets. 
14.0 STATISTICAL CONSIDERATIONS 
14.1 Sample Size 
The sample size of 3 patients per dose level was selected to allow expeditious assessment 
of toxicity prior to further treatment. 
14.2 Feasibility 
14.2.1 Patient Accrual 
Approximately 150 new colorectal and hepatocellular cancer patients are seen each 
year by the Gastrointestinal Oncology Service at University of California, San 
Francisco. Of the patients with colorectal cancer, approximately 50 will have liver 
metastases as the sole or predominant site of disease and will be candidates for 
hepatic artery infusion therapy. 
14.2.2 Study Duration 
Based on current activity patterns and anticipated concurrent protocols involving 
malignant liver tumors, the accrual of 18-24 patients is expected to take 
approximately 6-9 months. 
14.3 Analysis 
14.3.1 Pharmacokinetics Analysis 
The study will explore the relationships among pharmacokinetic parameters (see 
Appendix D), toxicity, and biological efficacy. This analysis will utilize methods 
stated in Section 14.3.3. 
14.3.2 Analysis of Gene T ransfer Efficiency 
The study will explore the relationship between dose of ACN53 and efficiency of 
transduction (gene transfer). This analysis will utilize methods stated in Section 
14.3.3. 
14.3.3 Analysis of Clinical Efficacy 
Evaluation of clinical efficacy is not a primary objective of this study. However, 
treatment effect will be estimated as the proportion of patients with objective 
responses (complete or partial response) following ACN53 therapy. Chi -square 
tests and logistic regression will be used to analyze which variables are significant 
predictors of response. 
14.3.4 S urvi val Analysis 
The Kaplan-Meier method will be used to estimate progression-free survival. 
[690] 
Recombinant DNA Research, Volume 20 
