6.0 STATISTICAL CONSIDERATIONS 
As a Phase II trial, this study has the goal of detecting responses in 
treated patients. To reduce the risk of treating patients on an unacceptable, 
inactive treatment protocol, there will be an early stopping rule. The study 
will proceed in two stages. In the first stage, 20 patients will be treated. 
If 3 or fewer responses are observed, the study will terminate. If 4 or more 
responses are observed, the study will proceed to a second stage, accruing an 
additional 20 patients. If the second stage is completed with a total of 11 or 
more responses observed among the ^ patients treated, it will be concluded that 
the treatment is deserving of further study. On the other hand, if the study 
terminates at the first stage or if the responses total 10 or fewer, the 
treatment will not be recommended for further study. Direct comparison of skin, 
blood and tumor will be made at each time point and compared to data obtained in 
patients previously receiving IL-2 alone in those treated with transduced TIL*s. 
6.1 AIM 1: ASSESSMENT OF EFFICACY 
We wish to distinguish between an objective response rate (CR + PR) of 20%, 
which is unsatisfactory, and a response rate of 40%, which is large enough to be 
of clinical interest. The study design is chosen to achieve this goal. The two- 
stage design described above has the following characteristics: 
1. The probability is less than 10% of making a "Type II error," i.e. , 
of rejecting the treatment when the true response rate is 40% or 
greater. 
2. The probability is less than 10% of making a "Type I error," or 
recommending the treatment for further study when the true response 
rate is 20% or less. 
3. The study design is optimal in the sense that it minimizes the expected 
sample size among designs with characteristics 1) and 2) , if the true 
response rate is 20% . 
The confidence interval reported for the response rate will be the exact interval 
based on the binomial distribution. In the unexpected eventuality of a very high 
response rate, this interval may differ substantially from the exact confidence 
interval based on the two-stage sampling design. In this case, both intervals 
will be reported. 
6 . 2 Aim 2: Evaluation of toxicity . 
Toxicities will be tabulated by type and severity. 
6.3 Aim 3: Trafficking and persistence of cells. 
Depending on the quantitative data obtained following recovery of gene 
marked cells . appropriate statistical tests will be performed to compare 
trafficking. Our assays (see data appendix) can detect by semiouantitative PCR 
1/10^ marked cells. Since biopsy samples will consist of at least a gram (10° - 
10^ cells) and blood samples 10° monanuclear cells/ml. we should have ^ 
excellent chance of observing TIL traffick and persistence. A result will be 
considered positive when the appropriate negative controls are truly negative and 
if it is repeated at least once. We would expect in each patient to generate a 
data matrix as shown on the following page: 
Recombinant DNA Research, Volume 14 
[849] 
