TKIs discontinuation has become a safe and feasible option for selected CML patients (pts). Sokal risk score, longer treatment and deep molecular response (DMR) duration were associated with lower probability of major molecular responde (MMR) loss after stopping TKIs (Etienne, 2010-Saussele, 2018). Moreover, the impact of bcr-abl transcript type on outcome and, particularly, on TFR probability and duration has been recently demonstrated (Claudiani, 2017-D’Adda, 2019). Previous INF therapy > 12 months was associated with a better outcome after imatinib discontinuation (Ross, 2013).
We retrospectively evaluated our CML pts who stopped TKI after sustained DMR (sDMR) in order to assess whether the type of previous treatment could influence the probability of a durable TFR.
Bcr-abl transcripts were determined by RQ-PCR analysis performed in accordance with EAC protocol (Gabert, Leukemia 2003) and to the standards of the Italian National Network Labnet. Criteria for TKI discontinuation was sustained DMR (MR4 or better) for at least 2 years. After TKI withdrawal, RQ-PCR for bcr-abl was performed every month during the first year and every 2 months thereafter. TKI treatment was reintroduced if DMR loss occurred.
Sixty-seven pts discontinued TKIs, 18 of them after less than 5 years of treatment because of pregnancy desire (3), intolerance (6), patient's desire/non compliance with therapy (5), enrollment in study protocol with preplanned discontinuation programs (4). At discontinuation, median age was 63 years (30–85), median time from TKI start 85 months (30–190), median duration of sDMR 48 months (24–153). Sokal distribution was 49%, 31% and 18% for low, intermediate and high risk (one pt was not evaluable). Thirty-eight pts stopped imatinib, 24 nilotinib (19 as first line, 5 as second line treatment), 5 dasatinib. Before imatinib, 15 pts received INF, for a median time of 60 months (3–256). Median follow up after TKI stop was 34 months (4–100, > 24 in 49 pts, <12 in only 2 pts). Twenty-seven (40 %) pts lost DMR. Median time off-therapy for these pts was 3 months (1–19), only 2 lost DMR after 6 months (+16 and +19 month). One pt aged 87 years has not yet resumed therapy because of age and severe kidney failure; he remains in MR3 at 47 months after TKI discontinuation. Therapy was restarted in 26 pts (14 in MR3, 11 in MR2, 1 in MR1), 24 achieved a second DMR after a median interval of 2 months (1–18); 2/26 pts are in M3 after 20 and 22 months. Neither cytogenetic relapses, nor progression were documented. One pt died in DMR for pancreatic cancer. Univariate analysis showed no difference in relapse risk according to age, gender, type of TKI (imatinib vs 2GTKI), duration of sDMR and Sokal score risk, while the e14a2 vs e13a2 transcript type (p = 0.02), duration of TKI therapy > 60 months (p = 0.03) and prior INF therapy (p = 0.02) were significantly associated with better outcome after TKI discontinuation, particularly for e13a2 transcript pts whose 12 months probability of TFR increased from 20% to 80% for INF pretreated pts (figure 1). At multivariate analysis the type of bcr-abl transcript (p = 0.014) and prior INF (p = 0.035) remained independently significant prognostic factors.
The e14a2 transcript confirmed as a favorable prognostic factor for TFR maintenance in CML pts receiving TKIs. However INF treatment before TKI improved TFR rate after TKI withdrawn, particularly in e13a2 transcript pts, and actually reverses the e13a2 transcript type negative prognostic impact on TFR maintenance in pts treated with TKI only