Background:
For patients with chronic-phase chronic myeloid leukemia (CML), treatment-free remission (TFR) is increasingly becoming a goal of therapy. While the safety of TFR has been established [Mahon, Lancet Oncol 2010; Ross, Blood 2013], the ability to predict success following attempted TFR remains limited. Recent publication of Euro-Ski [Saussele, Lancet Oncol 2018], the largest tyrosine kinase inhibitor (TKI) cessation study to date, demonstrated that duration of MR4 (BCR-ABL1 <0.01% IS) prior to attempted TFR was the main factor predicting TFR success.
Aim:
To identify the predictors of TFR in a single academic centre.
Methods:
We performed a retrospective analysis of adult CML patients receiving their primary CML management at the Royal Adelaide Hospital between January 2008 and March 2018, reviewing both clinical and molecular data. Criteria for qualifying for a TFR attempt included a minimum of 3 years (yrs) of TKI therapy and 2 yrs of deep molecular response (DMR: BCR-ABL1 <0.0032% IS; MR4.5). History of blast crisis, allogeneic stem cell transplantation and non-quantifiable atypical BCR-ABL1 transcripts precluded TFR qualification. Patients were monitored with monthly BCR-ABL1 qRT-PCRs for at least the first 12 months of the TFR attempt. In the event of molecular recurrence, defined as loss of major molecular response (MMR; BCR-ABL1 ≥ 0.1% IS) on a single test, TKI was recommenced. Our population also included 20 patients enrolled in the Australian CML8 (TWISTER) study where restart criteria was more stringent, requiring TKI restart in the event of BCR-ABL1 becoming detectable on 2 consecutive tests.
Results:
A total of 298 patients were treated at our institution within the defined time frame and 280 patients qualified for inclusion into our retrospective analysis. TFR eligibility was attained in 114 patients and 96 (84%) attempted TFR. Table 1 details patient characteristics of patients attempting TFR. Of the 82 patients with >12 months of follow-up, 52% (n=43) remain off TKI at 12 months in MMR. Variables were assessed by univariate Cox proportional hazards regresssion analysis for their association with TFR. The most significant finding was that patients attempting TFR with e14a2 BCR-ABL1 transcripts were more likely to remain in TFR at 12 months (65%; n=24/37) in comparison to the e13a2 transcript (34%; n=10/29), p = 0.008. This advantage also translated to patients with both e14a2 and e13a2 transcripts when grouped with the e14a2 cohort and compared with e13a2 alone, p = 0.006. The negative effect of the e13a2 transcript was further confirmed on multivariate analysis (Figure 1a) as patients with either e14a2 or both transcript types were 2.24 times more likely to remain in TFR at 12 months compared with the e13a2 transcript, p=0.032. Patients with sustained MR4.5 >3.4 yrs prior to cessation were more likely to remain in TFR at 12 months (42 vs. 64%, p = 0.014). We postulated that the higher rate of TFR in patients with e14a2 might be due in part to the longer time in MR4.5 prior to cessation. The median duration of MR4.5 prior to stopping in the e14a2 cohort was 4.1 yrs (2.05 - 10.76) compared to 3.01 yrs (2 - 10.41) in the e13a2 group (Table 2). Cumulative incidence curves of all 280 patients in our analysis demonstrated that by 6 yrs of TKI therapy, 70% of patients with e14a2 transcripts achieved MR4.5 whereas only 52% of patients with e13a2 transcripts attained MR4.5; confirming that patients with e14a2 transcripts are more likely to achieve DMR earlier (Figure 1b). Furthermore by 8 yrs, 48% of patients with e14a2 transcripts became eligible for a TFR attempt compared with only 32% of e13a2 transcripts (Figure 1c). While patients with e13a2 transcripts eventually achieve the same frequency of MR4.5 as the e14a2 group, the earlier achievement of MR4.5 in e14a2 patients may have contributed to the difference in TFR success.
Conclusion:
The factors that we identified as most predictive for TFR success were duration of MR4.5 and the presence of the e14a2 transcript, which has not been described previously. We also observed earlier achievement of MR4.5 in the e14a2 cohort, consistent with other studies [Jain, Blood 2016]. These observations, taken together, raise important questions about the impact of transcript type on disease biology, drug sensitivity, and immunological response which warrant further investigation.
Shanmuganathan: Novartis: Honoraria, Other: Travel sponsorship; Janssen: Honoraria; Royal Adelaide Hospital Research Fund: Other: Scholarship; Bristol-Myers Squibb: Honoraria, Other: Travel sponsorship. Branford:Qiagen: Honoraria, Membership on an entity's Board of Directors or advisory committees; Novartis: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding, Speakers Bureau; Cepheid: Honoraria; BMS: Honoraria, Membership on an entity's Board of Directors or advisory committees, Speakers Bureau. Yong:BMS: Honoraria, Research Funding; Novartis: Honoraria, Research Funding; Celgene: Research Funding. Hiwase:Celgene: Research Funding; Novartis: Research Funding. Yeung:Novartis: Honoraria, Research Funding; BMS: Honoraria, Research Funding; Pfizer: Honoraria; Amgen: Honoraria; Specialised Therapeutics Australia: Honoraria. Ross:BMS: Honoraria; Novartis: Consultancy, Honoraria, Research Funding, Speakers Bureau; Celgene: Research Funding. Hughes:BMS: Honoraria, Research Funding; Novartis: Honoraria, Research Funding; Takeda: Honoraria.