successfully marked, or in whom failure to engraft with 
the marked marrow, will be removed from the study. 
10.0 STATISTICAL CONSIDERATIONS 
10.1 Evaluation of Origin of Relapse and Comparison of the 
Contribution of Peripheral Blood cells and Marrow to 
Recovery: We plan to initially study ten patients. The 
accrual rate is 10 per year. Since relapsed patients are 
utilized, we estimate that 80% of patients may eventually 
relapse in two years. Thus, eight patients will be 
available for analysis of retroviral marking. Since 5 x 
10 4 to 5 x 10 6 retrovirally marked blast cells may be 
infused with each marrow and peripheral blood collection, 
it is probable that several of the eight patients will 
produce marked cells at relapse if the cells infused are 
the origin of the relapse. The LNL-6 GINa marking vector 
has been approved for use in man (See Appendix B) . 
We have calculated that if 1.4xl0 10 marrow cells are 
exposed to virus after a Ficoll Hypaque separation (which 
results in 5 fold reduction in the number of cells to 
2 . 8xl0 9 ) , and if the marrow is purged which results in a 
reduction of the number of cells to 5.6x10®, and if only 
30% of the marrow is used for marking 1.68x10 s ), and if 
the marking or transducing frequency is 3% with LNL-6 or 
GIN in CML marrow, or GINa in peripheral blood and the 
maximum ratio of CML blasts is one CML blast per 100 
nucleated cells at a maximum, the maximum number of 
marked leukemic blast cells present at storage will be 
5xl0 4 . The ratio of marked leukemia cells to the total 
number of cells frozen (5x10®) is 1 in 10,000. This 
ratio is clearly consistent with detection of NEO by PCR 
at the time of storage by PCR, and clearly will be 
adequate at the time of relapse since the growth of the 
leukemia cells, if relapse arises from the marrow, will 
be greater and therefore easily detected by PCR. 
At the time of relapse, the ratio of marked leukemia 
blasts/total unmarked cells will be higher. The presence 
of a total of 5xl0 4 marked leukemia blast cells in the 
autologous cells infused after TBI, VP16, and cytoxan is 
clearly sufficient to generate a polyclonal relapse if 
relapse occurs from the infused cells. These 
considerations suggest that a pilot trial be conducted 
with at least 10 patients to test if relapse occurs from 
infused autologous cells or from the residual disease 
present systemically. 
10.2 The difference in the percentage of cells positive for 
each of the two safety modified viruses used for the 
peripheral blood or marrow cells in the marrow after 
transplant and at 6 monthly intervals following 
transplant will be used to measure the relative 
[870] 
Recombinant DNA Research, Volume 15 
