Table 2.

Results of tumor cell purging in mobilized blood cells before and after serial purging steps

Tumor cell frequency
Log decrease
Exp.	Mobilized blood aphereses	CD34+ selected fresh cells	CD34+ cells after culture	NGFR+ selected cells
MM 3	< 1:49334	< 1:49334	N.D.	< 1:49334
	0	0	-	0
MM 6#	1:2449	1:11791	< 1:89606	< 1:89606
	0	0.68	> 1.56	> 1.56
MM 10	> 1:6211	> 1:28818	N.D.	1:28852
	0	< 0.67	-	0.67
MM 12#	1:4239	< 1:89606	< 1:89606	< 1:89606
	0	> 1.32	> 1.32	> 1.32
MM 13#	1:10646	< 1:89606	< 1:89606	< 1:89606
	0	> 0.92	> 0.92	> 0.92
MM 14#	1:490	1:11162	< 1:109697	< 1:109697
	0	1.36	> 2.35	> 2.35
MM 15#	N.D.	1:12789	1:2669	1:26667
	N.D.	0	0	1
MM 16	N.D.	> 1:12422	1:28852	1:89256
	N.D.	0	0.37	0.86
MM 18#	N.D.	1:238	< 1:44823	< 1:44823
	N.D.	0	> 2.27	> 2.27
MM 19	N.D.	1:89286	< 1:89606	< 1:89606
	N.D.	0	> 0	> 0
POOLED DATA	1:5291	1:12270	1:84602	1:208333
	0	0.36	1.2	1.59

Mobilized blood cells were analyzed after different purging steps for tumor load, where the specific MM marker was available. DNA from 1 × 10⁴, or 2 × 10⁴ cells was amplified to detect MM contaminants. For 6 patients (^#) serial dilutions of DNA were performed, with 0.5 logs step dilutions. The quantification of malignant cells in each cell fraction was calculated according to the single-hit Poisson model, and expressed as tumor cell frequency (1:x). The logarithmic decrease (lower part of cells) was calculated with respect to the foregoing unpurged sample indicated with "0". When the frequency of tumor contamination was below the detection threshold of the system, PCR were scored as negative (clear-shaded cells). The pooled data represent frequency determinations, according to the statistical method by Taswell.

N.D. = not done