NC 
Probability of marked relapse 
1 
0.010 
10 
0.096 
100 
0.634 
1000 
0.999 
3000 
1.000 
Since both relapses observed in our previous study were marked and only 
30% of the marrow was exposed to a vector, the above table suggest that it 
is unlikely that fewer than 100 progenitor cells from the reinfused marrow 
gave rise to the relapse. 
In this study of children with AML, we expect the two year continuous 
complete remission (CCR) rate to be approximately 60%. Thus we should 
observe 14 relapses in the 35 patients treated with ABMT. This assumes that 
ABMT will show no improvement in disease control over the historical data. 
With the same assumptions used to derive the previous table, the following 
table provides estimates of detecting at least one marked relapse and the 
expected number of relapses which will be observed. 
NC 
Prob (Exp) 
1 
0.13 (0.1) 
10 
0.75 (1.3) 
100 
1.00 (8.9) 
1000 
1.00 (14.0) 
3000 
1.00 (14.0) 
If we observe 14 unmarked relapses and there is evidence that normal 
progenitor cells are marked with both vectors, we will be 95% confident that 
the true failure rate of purging is less than 20%. Note that this result alone 
does not demonstrate the necessity of purging. That follows from the previous 
experience of the gene marking study without purging in AML. 
Recombinant DNA Research, Volume 17 
[245] 
