Prospective evaluation of peripheral blood CD26+leukemia stem cells in chronic Myeloid Leukemia patients during TKI discontinuation (Flower-TFR Study)

Background and rationale: In order to better identify CML patients (pts) suitable for an efficacious treatment free remission (TFR) are warranted additional biological criteria to molecular response. Leukemia stem cells (LSCs) are supposed to be the reservoir of disease. We first showed in a cross-sectional study that residual circulating CD34+/CD38-/CD26+ CML-specific LSCs are still detectable in the peripheral blood (PB) of the majority CML pts in sustained TFR (66%) despite stable and deep molecular response. Aims: In prospective FLOWER-TFR multicenter study we monitored by flow-cytometry the number of circulating CD26+LSCs in CML pts from the time of TKI discontinuation until molecular relapse, if any. Methods: CML pts meeting the current molecular criteria for TKI withdrawal entered this study. At time of stopping TKI treatment (baseline) and at +1, +2, +3, +6, +12 months (mos) after discontinuation and at any time of molecular relapse, CML pts were evaluated for number of PB CD34+/CD38-/CD26+LSCs by centralized flow-cytometry analysis and for BCR-ABL transcript levels by QRT-PCR assay. Results: 72 consecutive CML pts were enrolled. After a median observation time of 11 mos since TKI withdrawal (1-37 mos), 20/72 (28%) pts lost their molecular response and restarted TKI treatment while 52/72 (72%) are still in TFR; of note 12/72 (17%) pts have so far discontinued the treatment for ≤ 6 months. The median time to relapse after discontinuation was 4 mos (range 2-7 mos) (Table 1). At the time of discontinuation, residual CD26+LSCs, were detectable in 37/72 (51%) pts: of those 25/37 (67%) sustained TFR and 12/37 (33%) lost response. The median number of detected CD26+LSCs was 0.0237μ/L (range 0.0077-0.1197) with minimal fluctuation at different time points. On the other hand, 35/72, 49% pts showed no detectable CD26+LSCs at time of discontinuation: 27/35 (77%) pts maintained TFR and 8/35 (23%) pts lost response. No statistical correlation between BCR-ABL/ABLIS ratio and number of residual CD26+LSCs was found. However, we observed that pts in which both LSCs and BCR-ABL copies were detectable had the highest percentage of TFR loss while pts with both undetectable LSCs and BCR-ABL copies had the lowest probability to TFR loss (Table 1). Conclusions: Our results confirm that CD26+LSCs are detectable at time of TKI discontinuation and during TFR. Moreover, the persistence of “fluctuating” values of CD26+LSCs do not hamper the possibility to maintain a stable TFR. Pts discontinuing TKIs with no detectable CD26+ and no detectable BCR-ABL copies appear to have less probability to undergo TFR loss (21%) compared to pts with both detectable CD26+LSCs and BCR-ABL (TFR loss 40%). However, no correlation between BCR-ABL/ABLIS ratio and number of residual CD26+LSCs was found. Additional studies evaluating CD26+LSCs ability to modulate the immune system through a variable expression of immune response inhibitory molecules are ongoing.