Prior to the introduction of the tyrosine kinase inhibitors (TKI) into clinical practice, the only curative therapy for chronic myeloid leukaemia (CML) was allogeneic hematopoietic stem cell transplantation (HSCT). HSCT is nowadays recommended for patients in accelerated phase, blastic phase or with the T315I mutation and for patients who experience second-line TKIs (TKI2) failure or intolerance. The impact of TKI2 for CML on the results of subsequent HSCT has not yet been clearly established, although preliminary data does not suggest an increase of non relapse mortality (NRM) in patients previously treated with these agents. To assess whether exposure to TKI2s before HSCT adversely affects outcome, we retrospectively analyzed 31 patients with CML reported to the French registry for HSCT between January 2001 and December 2008 who received a first HSCT for imatinib resistant or intolerant CML subsequently treated with either nilotinib or dasatinib or both.
The median age at diagnosis was 39.8 years (range, 19–61). At the start of TKI, 15 (48%) patients were in chronic phase, 12 (39%) patients were in accelerated phase and 4 (13%) patients were in blast crisis. For chronic phase patients, 6 (40%) patients were classified as low Sokal risk, 7 (46%) as intermediate risk, and 2 (14%) as high risk. After imatinib failure (among the 27 patients in chronic or accelerated phase at diagnosis), 21 and 6 patients received dasatinib and nilotinib, respectively. Among these patients, 9 patients received a third line therapy, including 5 sequential therapies with both drugs. The best response to the second and third-line treatments with TKI2 was a complete molecular response in 2 patients, a major molecular response in 2 patients, a complete cytogenetic response (CCR) in 7 patients, a complete hematologic response in 14 patients and no response in 2 patients. 14 patients eventually failed TKI2 treatment because of resistance, whereas 8 were considered as intolerant. Of 14 patients who developed resistance to TKI2, 2 had a mutation identified (T315I in both cases).
Median interval from diagnosis to HSCT was 19 months (range, 3–151). At time of transplant, 21 patients were in chronic phase, 10 in accelerated phase and none in blast crisis. 19 patients received a graft from an unrelated donor whereas 12 a match related donor. Stem cell source was peripheral blood, bone marrow or cord blood in 20, 8 and 3 patients, respectively. The conditioning regimen was myeloablative in 21 patients combining either TBI and cyclophosphamide (9 patients) or high dose IV busulfan and cyclophosphamide (12 patients), and a RIC for 10 patients. All patients engrafted successfully: median time to neutrophil and platelet recovery was 18 days and 21 days respectively. Grade 2–4 acute graft-versus-host disease (aGVHD) was observed in 11 (37.9%) patients; grade 3–4 aGVHD occurred in 6 (20.6%) patients. Chronic GVHD was observed in 15 (60%) of 25 patients alive after day 100. The median follow-up after HSCT is 27 months (range, 1.2–50.2). At time of analysis, 11 (35%) patients died, 7 (22.5%) from NRM and 4 (12.5%) from progression of disease. NRM was due to infection (3 patients), GVHD (2 patients), post transplant lymphoma disease (1 patient) and unknown cause (1 patient). The 1-year overall survival (OS) was 79.2% (95% CI, 64,3-94,1%) and the estimated 2-year OS was 55,5% (95% CI, 35,0-75,9%). One-year relapse and NRM rates were respectively 10,3% (95% CI, 2.5–24.6%) and 19.1% (95% CI, 6.7–36.2%). In univariate analysis, no variable had a significant impact on outcome among Sokal score, disease phase at diagnosis, Grathwol score, age at HSCT, time from diagnosis to HSCT, or quality of response before HSCT. In a multivariate analysis, only quality of response (at least better than CCR) was significantly associated with a better outcome in terms of OS (p=0.0459, HR=0.17, 95% CI=0.03-0.97).
In conclusion, TKI2 prior to HSCT did not result in an increased risk of NRM. However, our observation that patients in cytogenetic or molecular response at time of HSCT have a significant better outcome underscores the importance of a stringent prospective molecular monitoring under TKI2 therapy and the need for prognostic factors of response under TKI2. Such elements could help to better define after initiating TKI2 treatment patients that could really benefit from HSCT and the best timing of the procedure.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.