8.0 STUDY SIZE AND STATISTICAL CONSIDERATIONS 
8.1 Transduction of Malignant Progenitor Cells 
The transduction efficiency for clonogenic neuroblastoma cells from fresh 
tumor is between 0.3 and 3%. At present, there is no way of assessing 
whether "clonogenic cells" and neuroblastoma "stem cells" are identical or have 
the same transduction efficiency. If we assume that transduction efficiency for 
clonogenic cells and stem cells is identical and take a figure of 1% as average 
efficiency, then only 0.3% of the total malignant "stem cells" present in a 
harvested marrow will be transduced in the 30% aliquot exposed to the 
vector. This makes it unlikely that marrow relapse from <10 progenitors 
would be detected unless >50 patients were studied - a number not currently 
feasible. Polyclonal relapse (possibility 2 in section 7.2 above) would be more 
readily detected with a probability that would increase as the number of NB 
"stem cells" present increased. The marrow relapse rate in patients receiving 
ABMT for relapsed/refractory neuroblastoma is >80% at 12 months. A 
polyclonal relapse would therefore be detectable in a cohort of 12 individuals 
with a probability of >95% even if only 100 cells contribute to the relapse. 
In any one individual, there is >60% chance of detecting a marked relapse if 
1000 cells contribute, and >95% chance if 3000 cells contribute. 
8.2 Transduction of Normal Progenitors 
Because the probability of transducing a human stem cell in a patient 
recovering from chemotherapy is unknown (see Appendk A) and because the 
number of pluripotent stem cells contributed by an autologous marrow graft 
is unknown, it is not possible to assess the number of patients required to 
undertake this subsidiary aim. But if transduction into early progenitors is 
possible, then data on the value of this approach for investigating stimuli which 
modify stem cell behavior could readily be obtained within the first 12 patients, 
since even a single positive patient would be informative. Should the first 12 
patients show no evidence of early progenitor marking, this component of the 
study would not be continued separately from the primary aim of analyzing the 
source of relapse. 
8.3 Statistical Considerations for Evaluation of ABMT as Therapy 
We expect to accrue 8 to 10 patients per year on this study. The drug 
combination will be considered effective if and only if we were able to produce 
at least a 30% response rate with acceptable toxicity. Using a significance 
level of a = 0.05 yields the following table. 
Recombinant DNA Research, Volume 14 
[721] 
