CML Patients Outcome after TKI Discontinuation: A Single Institution Experience in the US

Chronic myeloid leukemia (CML) patients who achieve a complete cytogenetic response (CCyR) with tyrosine kinase inhibitor (TKI) have an expected survival rate comparable to that of the general population. It is practice to continue TKI therapy indefinitely. However, this chronic therapy may impact patient's lives either by its associated adverse events, or its influence on real-life events (eg. pregnancy, financial burden). Therefore, treatment discontinuation has been increasingly sought. Discontinuation of therapy for patients with sustained MR4.5 has been associated with 30-50% sustained molecular response in previous studies. However, patients may sometimes choose to discontinue therapy for various reasons regardless of the response.

We retrospectively analyzed the outcome of 95 patients with CML who discontinued TKI therapy at a single institution. Median age at diagnosis was 50 years (range 26 - 75), 56 (59%) were females. Median follow up from diagnosis to treatment discontinuation was 120 months. TKI was the initial therapy in 62 patients (41 imatinib, 7 nilotinib, 12 dasatinib, 1 bosutinib, 1 ponatinib) and interferon in 33 patients (34%). At time of discontinuation, 67 patients (71%) were receiving their first TKI (48 imatinib, 6 nilotinib, 12 dasatinib, 1 ponatinib) while 25 (26%) and 3 (3%) patients were receiving their second and third TKI respectively. All patients achieved complete cytogenetic response (CCyR) and major molecular response (MMR) in a median of 3 (range 2-93) and 9 months (range 2-132) respectively. MR4.5 was achieved in 92 (97%) patients at a median of 17 months. Three patients had MMR as their best response. Patients received TKI for a median of 104 months (range 8-203) before discontinuation. Fifty two patients (54%) discontinued therapy due to adverse events, 27 (28%) electively due to sustained MR4.5, 7 (7%) due to pregnancy, 6 (6%) for financial reasons and 3 (3%) due to the occurrence of a second cancer. At time of discontinuation all patients were in CCyR; 90 patients (95%) were in MR4.5 and 5 patients (5%) in MMR. The median MR4.5 duration before discontinuation was 75 months (range 1-178); 84 had a sustained MR4.5 for at least 2 years and 59 for at least 5 years. After a median follow-up from treatment discontinuation of 23 months (0 to 113 months), 38 (40%) patients have lost their response at a median of 4 months (range 1-34) after discontinuation; 10 patients lost response after 12 months from discontinuation. Among patients with MR4.5 at discontinuation, molecular relapse occurred in 33 patients (37%) at a median of 4 months (range 1-34; 8 after 12 months from discontinuation). All 5 patients with MMR at discontinuation loss their response at a median of 8 months (range 2-14). Patients receiving imatinib, dasatinib or nilotinib had a median TKI treatment duration of 117, 64 and 65 months before discontinuation, and lost their response at a rate of 33%, 56% and 46%, respectively. Relapse rate for patients with MR4.5 sustained >2 years was 32%, whereas those with <2 years it was 82%; a similar analysis with cutoff of 5 years yielded a relapse rate of 15% and 77%, respectively. Among patients with b3a2 with MR4.5 at discontinuation, 35% lost response, compared to 33% for those with b2a2 and 53% for those with both b3a2 and b2a2. Twenty four relapsing patients were eventually retreated; 21 re-gained the response they had at the time of discontinuation in a median of 4.5 months after resumption of therapy. At last follow up, 82 (86%) patients were in MR4.5, 4 (4%) in MMR, 5 (5%) in CCyR, 3 (3%) in PCyR and 1 (1%) in mCyR.

This analysis shows that patients with less than MR4.5 at time of discontinuation have higher risk of relapse. The lower percentage of relapse seen in patients who had been on imatinib likely represents the longer period of treatment before discontinuation compared to the 2^nd generation TKIs. Late relapses (e.g., beyond a year) occur in a subset of patients. Although the majority of patients who lose their response after discontinuation respond to retreatment, treatment discontinuation is still not recommended outside clinical trials.