
Correlation Observed Between Leukemia Stem Cells and Response in CML
Key Takeaways
- A lower number of CD26+ LSCs at diagnosis correlates with optimal molecular response in CML patients at 3, 12, and 24 months.
- CD26+ LSCs measurement could serve as a predictive tool for TKI response, offering additional diagnostic information.
New research reveals that CD26+ leukemia stem cell levels at diagnosis predict treatment response in chronic myeloid leukemia patients, enhancing TKI therapy strategies.
Final results from the prospective FLOWERS study demonstrated a correlation between the amount of CD26þ leukemia stem cells (LSCs) plus at diagnosis and at the molecular response in patients with chronic myeloid leukemia (CML). The findings suggest the number of CD26+LSCs at diagnosis could represent an additional tool for predicting tyrosine kinase inhibitor (TKI) response.1
There was a median value of 7.14 cells/µL at the time of diagnosis in patients. There was no statistical difference observed during TKI treatment. However, a significant correlation was observed between a low amount of CD26+LSCs at diagnosis and an optimal molecular response at 3 (P =.03), 12 (P =.004), and 24 (P =.009) months. There were 3 tertiles identified in the study: <3.21 cells/µL; between 3.21 and 19.21 cells/µL; and >19.21 cells/µL. A higher optimal response was observed in patients in the first tertile (P =.027, P = .015 vs P =.079).
“First, we confirmed that in all 242 patients [peripheral blood] PB CD26+ LSCs were measurable at diagnosis albeit with a great value variability between patients ranging from 0.01 to 698 cells/µL (median, 7.14 cells/µL),” stated Anna Sicuranza et al, authors of the study. “However, we showed that PB CD26+ LSCs rapidly and deeply reduce during TKI treatment remaining detectable with fluctuating similar values from then on.”
The study had a total enrollment of 242 patients with CML. The median age of patients at diagnosis was 60 years (range, 18-90). Sokal scores were high in 16% (n = 38) of patients, intermediate in 36% (n = 87) of patients, and low in 44% (n = 106) of patients. EUTOS scores were high in 8.2% (n = 20) of patients, low in 83.6% (n = 202) of patients, and unknown in 8.2% (n = 20) of patients. ELTS scores were high in 14% (n = 32) of patients, intermediate in 35% (n = 85) of patients, low in 47% (n = 114) of patients, and unavailable in 4% (n = 11) of patients.
Cytogenetic abnormalities were reported in 5.3% (n = 13) of patients. Complete cytogenetic response was achieved at a median time of 4 months (range, 3-13 months). There was a median observation time of 66 months. Following the observation time, 35% (n = 84) of patients switched to TKI treatment; 43% (n = 36/84) due to failure or resistance, and 57% (n = 48/84) due to intolerance. Overall, 54% (n = 132) of patients began treatment with imatinib (Gleevec), 30% (n = 72) with nilotinib (Tasigna), and 16% (n = 38) with dasatinib (Sprycel). Of the total number of patients, 39% (n = 95) received cytoreductive therapy (hydroxyurea) before TKIs.
In the imatinib, nilotinib, and dasatinib cohorts, there were median values of 5.53 cells/µL, 11.98 cells/µL, and 13.27 cells/µL, respectively. A significant correlation between the bulk of CD26+ LSCs at diagnosis and Sokal score was observed (P =.018). Levels of CD26+ LSCs were significantly higher in the high Sokal risk group compared with the intermediate- or low-risk groups (22.65 cells/µL vs 5.60 cells/µL vs 6.16 cells/µL). There was no correlation found between the bulk of CD26+ LSCs and low or high EUTOS scores (P =.109) and high, intermediate, and low ELTS scores (P =.224).
It was determined that an optimal molecular response at 3, 12, and 24 months was always correlated to a lower amount of CD26+ LSCs. Patients with CML with an optimal molecular response at 3 months showed a median CD26+ LSCs at diagnosis of 6.21 cells/µL. Suboptimal responses had a median of 19.87 cells/µL at diagnosis (P =.03). At 12 months, patients showed a median amount of CD26+ LSCs of 5.50 cells/µL at diagnosis and suboptimal responders had a median of 16.87 cells/µL (P =.004). At 24 months, patients with an optimal molecular response had a median CD26+ LSCs value at diagnosis of 6.05 cells/µL compared with suboptimal responses, showing a median value of 20.52 cells/µL (P =.009). In patients who failed molecular response after first-line TKI treatment and switched treatment (n = 36), patients had a significantly higher median CD26+ LSCs at diagnosis, accounting for 14.59 cells/µL compared with a median CD26+ LSCs value of 5.82 cells/µLin the no-switch group (P =.034).
“Although the exact significance of the presence and persistence of these unique cells in the natural history of CML is not yet fully understood, our data demonstrate a correlation between the amount of [peripheral blood (PB)] CD26+ LSCs at diagnosis and the molecular response,” concluded Sicuranza et al, authors of the study. “Given these results, flow cytometry measurement of the absolute number of PB CD26+ LSCs represents not only an easy and rapid diagnostic tool, it could also furnish additional information for predicting TKI response.”





































