Background: Prior to TKIs, CML was uniformly fatal unless the patient underwent allogeneic-HSCT. With TKIs, some patients achieve deep molecular responses with bcr-abl <0.01% (MR4); a portion of such patients can achieve long-term TFR, possibly cure, after TKI cessation. This approach is now considered the goal for every newly diagnosed CML patient. We present a single centre experience, with long-term follow up, of a cohort of CML patients undergoing TFR1 and TFR2 attempts.

Method: Our database of 173 patients diagnosed with CML from 1994 to 2022 was searched for patients who had the potential to be eligible for a TFR attempt (pending adequate response to TKI): CML diagnosed in 1st chronic phase (CP1); no allogeneic HSCT in CP1; >3 yrs treatment with TKI; adequate follow-up. 120 patients fulfilled these criteria.

Results: 70/120 (58%) achieved accepted criteria to attempt TFR: >3 years of TKI treatment; a deep molecular response (≥MR4) and underwent TFR attempt. Median age at diagnosis was 49.5 (20-81) yrs; 42% were female. Sokal score was low in 38%; intermediate in 29%; high in 33%. 66% of patients received multiple TKIs; 23% had received interferon. TKI at time of TFR1 attempt was imatinib (40%), nilotinib (30%), dasatinib (27%), and ponatinib (3%). Median time from diagnosis to TFR was 8.5 (4-23) yrs and median time on TKI was 7.5 (3.4-17.9) yrs.

36/70 (51%) remain in TFR with median follow-up of 46 (1-180) mths including; 16 patients now >5 yrs and 2 now >10 yrs in ongoing TFR; the cohort includes 2 patients with historically acquired bcr-abl mutations and 1 with a historically acquired additional cytogenetic abnormalities (no longer present at TFR1 attempt). All 3 are long-term TFR successes.

Median time to relapse was 5 (2-36) mths, 25/34 (73.5%) relapsed within 6 mths of TKI withdrawal and 28/34 (82.4%) within 12 mths. Time of relapse for 1 patient is unknown. The remaining 5 patients relapsed at 13, 14, 18, 20, and 36 mths. All patients with CML recurrence (bcr-abl >0.1%) restarted TKI with a rapid return to major molecular response (MMR) or better (2 are too early for assessment) with median time to MMR of 3 (1-12) mths. No patient developed CML with high-risk features or blast phase.

The strongest predictors for TFR success were duration of MR4 prior to attempting TFR1 (Table 1) - 60% (25/42) of patients with >5 yrs of ≥MR4 prior to cessation remain in TFR, compared to 55% (11/20) with 3-5 yrs and 0% (0/7) of patients with <3 yrs (p<0.03 for <3 vs 3-5 yrs and <3 vs >5 yrs, p=0.79 for 3-5 vs >5 yrs) ; prior treatment with interferon (81% - 13/16 - of patients who received interferon remain in TFR compared to 37% - 20/54 - who had not received interferon, p=0.0035), and a Sokal score <0.8 (64% of patients with low Sokal score remain in TFR compared to 45% with a score 0.8-1.2 and 45% with a score >1.2, p=0.35 for <0.8 vs 0.8-1.2, p=0.16 for <0.8 vs >1.2, p=1 for 0.8-1.2 vs >1.2).

6 patients had transient rises in their bcr-abl (>0.01%), ranging from 0.012 - 0.033, but these 6 returned to MR4 or better and remain in TFR for 8-77 mths.

13 (38%) of the 34 failed TFR1 patients attempted TFR2. Median time on TKI between 1st and 2nd attempts was 5.7 (4.8-8.9) yrs and had a median time ≥MR4 of 5.4 (4.6-8.8) yrs. Median follow up was 12 (1-27) mths. Currently, 9 (69%) patients remain in TFR2. Of the 4 patients who relapsed, all did so in the first 6 months after TKI withdrawal.

17 patients (25%) developed a withdrawal syndrome. Symptoms included fatigue, myalgias, and arthralgias. All but 4 patients had resolution over a median of 11 (1-25) mths. 3 patients benefited from a course of prednisolone. 4 patients have had persistent symptoms for median 12.5 (10-19) mths.

Conclusion: Of 120 patients with adequate duration of TKI, 58% were eligible for TFR1 attempt which, to date, is successful in 51%. Prior interferon treatment, longer period in ≥MR4 and a Sokal score <0.8 predicted a successful TFR. For those that fail TFR1, TFR2 appears safe, and with short follow-up, currently 9/13 (69%) patients remain off treatment.

In our cohort, overall 64% (45/70; 36 TFR1 + 9 TFR2) of patients eligible for TFR attempts remain off TKI giving an actual or functional cure rate of 38% (45/120). Historically acquired bcr-abl mutations or additional cytogenetic abnormalities are not necessarily contraindications to a TFR attempt - with 3 long-term TFR successes in our cohort. TFR is safe with no cases of disease progression, and all patients with TFR failure and adequate duration of follow-up regaining MMR.

No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution