Although imatinib is highly effective in inducing remissions in CML patients, the long-term durability of response is not clear. Here we report updated results of a clinical trial investigating the feasibility and efficacy of collection and storage of PBSC from patients in complete cytogenetic remission (CCR) on imatinib, for use in autologous transplantation in the event of subsequent relapse. PBSC were collected from 36 patients [31 CP, 5 AP (at start of imatinib); median age 45 years (range 22–70); 21 males, 15 females; median time from diagnosis 25 months (mos) (6–90); median duration of imatinib treatment 13 mos (6–41); median time from CCR 7 mos (1–26)]. Patients were administered G-CSF (10μg/kg/day), and PBSC collection initiated on day +5 with a targeted minimum of 2x106 CD34+ cells/kg. Imatinib was continued during G-CSF administration and PBSC collection. The G-CSF dose was escalated in case of poor collection. The median number of CD34+ cells (106/kg) collected was 2.56 (0.31–6.19) with a median of 3 phereses (1–13). Five patients failed to collect the target number of CD34+ cells, achieving a median of 0.87x106 CD34+ cells after a median 5 collections. Seven patients required >6 collections to reach the target cell dose. There was no significant relationship between rapid (≤3) or slow (>3) collection, or failure to collect, and clinical characteristics such as age, sex, disease stage, prior interferon, time from diagnosis, time on imatinib, and duration of CCR. PBSC collections were evaluated for BCR/ABL contamination by cytogenetics and PCR (Q-PCR and nested RT-PCR). Ph+ cells were detected on cytogenetic examination in 1 or more collections from 5 patients and Ph- abnormal clones were detected in 4 patients. Patients with Ph+ PBSC were slower collectors than those with Ph- PBSC (median 8 vs. 3 collections, p=0.04). BCR/ABL mRNA was detected by PCR in 1 or more collection from 30 of 32 patients evaluated. Two patients, both with BCR/ABL mRNA detected in pre-mobilzation marrow, collected with a single PCR negative collection. Three patients had collections with low levels of BCR/ABL mRNA detected only with more sensitive nested RT-PCR. Of 134 separate collections analyzed, BCR/ABL mRNA was detected by Q-PCR in 113 (84%), by nested RT-PCR in 11 (8%), while 10 (8%) were PCR negative. Rapid collectors had significantly lower BCR/ABL mRNA levels in their collections compared to pre-mobilization marrow (p<0.05), whereas slow collectors did not show significant change in BCR/ABL levels. CD34+ cells isolated from PBSC showed significantly increased BCR/ABL mRNA levels compared with total nucleated cells (BCR/ABL:B2M ratio of 0.0006±0.0002 for NC vs. 0.03±0.01 for CD34+ cells, p=0.002). PBSC were injected into NOD/SCID mice to evaluate for presence of BCR/ABL+ progenitors capable of in vivo engraftment. Human cell engraftment was confirmed by flow cytometry in 3 of 4 patients, and BCR/ABL mRNA was detected in engrafted cells by Q-PCR. Our results indicate that cytogenetically negative PBSC collections can be obtained from CML patients receiving imatinib, but that mobilization is relatively poor. Rapid collectors have reduced BCR/ABL+ cell contamination in PBSC collections, and PCR negative collections are possible. However, the majority of PBSC products show evidence of persistent malignant stem cells. Additional strategies to enhance reliability and rapidity of collection and further deplete BCR/ABL+ stem cells in the PBSC product need to be explored.