Comment and response to: the significance of early warning in chronic myeloid leukemia

Dear Editor:

We read with interest the correspondence by Mario Tiribelli and colleagues [1] about our recent paper regarding the clinical significance of ‘warning’ signs (especially on the long-term outcomes) in patients with chronic myeloid leukemia (CML) receiving upfront tyrosine kinase inhibitor (TKI) treatment according to the European LeukemiaNet (ELN) 2013 recommendations [2]. We thank the authors for their gentle comments on this work.

We fully agree with the authors that the ELN 2013 definitions of ‘warning’ signs at third and sixth months of frontline TKI therapy consider both cytogenetic and molecular response status. On the other hand, only the degree of molecular response is considered to define a warning response at 12 months [2]. So if we combine these two response categories together, in a subset of patients there can be discordant results between the cytogenetic and molecular responses at these time points, and as the authors stated, among their cohort these patients were 9.3% and 11.6% of the entire cohort at 3 and 6 months, respectively [3]. And they concluded that patients having at least one warning sign at 3 or 6 months had inferior outcomes, but not statistically different to that observed in patients with concordant cytogenetic and molecular warning results [3]. However it is more than a possibility that patients may achieve optimal cytogenetic responses at 3 and 6 months (Ph+ ≤ 35% and Ph+ 0%, respectively) while having a molecular warning response but the opposite is unlikely. Supporting this, in a recent article, the rates of patients with both optimal cytogenetic and molecular warning responses at 3 and 6 months were 6.6% and 7.0%, respectively [4]. So, although the rates of patients with discordant results are relatively small, combining these two monitoring modalities may be worthy while defining patients with warning responses.

The probable impact of BCR-ABL1 transcript type on the response status is another issue that is gaining interest recently. As the authors had previously shown, patients with b2a2 BCR-ABL1 transcript had higher rates of warning response (WR) at 3 and 6 months than those with b3a2 type [5]. Similarly in another cohort, although not translating into an overall survival (OS) benefit, patients with b3a2 transcript were shown to achieve earlier and deeper responses compared to those with b2a2 [6]. The impact of BCR-ABL1 transcript type on the response status as well as long-term outcomes warrants further evaluation.

As the authors stated, there is also uncertainty in the clinical relevance of a late warning in terms of long-term outcome. Recently, we performed a study in a cohort consisting of 73 CML patients in chronic phase (CML-CP) treated with first-line imatinib 400 mg daily with a median of 45 months. Seventy of these patients had a BCR-ABL1 testing available at 12 months, and of these patients 34 (49%) had a BCR-ABL1 (IS) transcript level ≤ 0.1% (optimal response (OR)), and the number of patients with a BCR-ABL1 (IS) transcript level of 0.1–1% (WR) and >1% (failure) were 23 (33%) and 13 (18%), respectively (unpublished data). The complete cytogenetic response (CCyR) rate at 12 months was 82%. Patients with a warning sign at 12 months continued imatinib with the same dose, and the cumulative CCyR rates were similar between patients with OR and WR (p = 0.227), and patients with OR and WR both had significantly superior CCyR rates than patients with failure (p < 0.0001 and p < 0.0001, respectively). Optimal responders had significantly higher rates of cumulative major molecular response (MMR) than patients with WR (p = 0.031), and both optimal and warning responders had significantly higher MMR rates than those of patients with a failure response (p < 0.0001 and p < 0.0001, respectively). Event-free survival (EFS) was comparable between patients with OR and WR, and both groups had superior EFS than patients with failure. Patients with WR and failure had similar OS, and the optimal responders had superior OS than both patients with WR and failure; however, the difference was not significant. As the authors concluded, our findings together with the current data available indicate that achieving CCyR at 12 months is most probably one of the strongest determinants of favorable long-term outcome in terms of survival [1]. And no study yet demonstrated a survival benefit of switching to another TKI in patients with a WR at 12 months, whereas this maneuver may result in achieving deeper molecular responses among this subgroup of patients [1,2].

Patients with cytogenetic and/or molecular warning signs have inferior outcomes in terms of achieving long-term molecular responses than that of patients with OR; however, this does not necessarily translate into unfavorable survival outcomes especially if CCyR is achieved and maintained. As the authors concluded, it is obvious that prospective studies are needed to explore the possible benefit of treatment switch in CML patients with warning signs.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.