A multicenter retrospective evaluation of Chronic Myeloid Leukemia (CML) therapy in Austria assessing the impact of early treatment response on patient outcomes in a real-life setting

Summary Background Several clinical trials in chronic phase (CP) chronic myeloid leukemia (CML) showed that early response to tyrosine kinase inhibitor (TKI) treatment results in an improved long-term survival and progression-free survival. This study assessed whether patients achieving early treatment response (ETR; partial cytogenetic response or BCR-ABL1 mRNA ≤10% at 3 months) in daily practice also have a long-term survival benefit. Methods The Retrospective Evaluation of Early response in CML for long-term Treatment outcome (R-EFECT), a multicenter, retrospective chart review, documented patients with newly diagnosed CML-CP starting first-line TKI therapy in routine clinical practice. The primary aim was to assess the 5‑year overall survival rate. Results Of the 211 patients from 12 centers across Austria (January 2004–May 2010), 176 (median age, 56 years) were included in the analysis. All patients received first-line therapy with imatinib. Overall, 136 patients (77.3%) achieved ETR (ETR+ group), whereas 40 (22.7%) did not reach ETR (ETR− group). The ETR+ group had higher 5‑year overall survival (92.5% vs. 77.5%, P = 0.018) and progression-free survival (95.6% vs. 87.5%, P = 0.06) rates compared with the ETR− group. As expected, more patients in the ETR− group were switched to another TKI. At the last contact, 120 patients were still on imatinib and 44 had switched to another TKI (25 to nilotinib, 15 to dasatinib, and 4 to bosutinib). Conclusion The data are in line with randomized trials demonstrating that ETR is associated with improved survival and thus confirmed these results in patients treated in daily clinical routine.


Introduction
According to the European Leukemia Network (ELN) guidelines for chronic myeloid leukemia (CML), BCR-ABL1 transcript levels (addressed as BCR-ABL1 throughout) of ≤10% according to the international scale (IS, BCR-ABL1 IS ) in the peripheral blood or a partial cytogenetic response (pCyR), <35% Philadelphia (Ph)+ metaphases, are defined as optimal treatment response at 3 months [1]. At 6 months and 12 months, the BCR-ABL1 IS should be <1% and ≤0.1%, respectively, or alternatively a complete cytogenetic response (CCyR; 0% Ph+ metaphases) at ≥6 months should be achieved to be classified as optimal response [1].
Several clinical trials have shown that patients with CML in chronic phase (CP) who achieve early molecular response (EMR, BCR-ABL1 IS ≤10% at 3 months) with tyrosine kinase inhibitor (TKI) therapy have better long-term responses, overall survival (OS), and progression-free survival (PFS) [2][3][4][5][6][7][8][9][10]; however, these data were obtained in patients selected for clinical trials. Therefore, it was of particular interest whether the survival benefit observed in these studies translates to patients treated in daily clinical practice.
This article presents data from a multicenter, retrospective analysis that assessed whether the achievement of an early treatment response (ETR) results in an improved survival in patients receiving TKI in clinical routine.

Study design and patients
The Retrospective Evaluation of Early response in CML for long-term Treatment Outcome (R-EFECT) study documented 211 patients with newly diagnosed CP-CML who received first-line TKI therapy in clinical routine between January 2004 and May 2010 in 12 centers across Austria. Only patients with documented BCR-ABL1 levels and/or % Ph+ metaphases at the 3-month visit were assessed. Patients who had received short-term (for ≤3 months) cytoreduction (e.g. hydroxyurea) or those who were subsequently treated with second or third generation TKIs (when available) could also be included. Patients who participated in interventional clinical trials were excluded.
The aim was to document outcomes in the clinical practice. We assessed the OS rate at 5 years in patients who achieved ETR, compared with those who did not reach ETR. Additional aims were to define the OS rates in the ETR+ and ETR-groups, the rates of progression to accelerated phase/blast crisis (AP/BC) at 5 years, and treatment responses at 6 months. In addition, ETR and OS rates were assessed based on Sokal risk scores at baseline [11].

Assessments and definitions
The ETR was defined as achievement of at least a pCyR (≤35% Ph+ metaphases) or BCR-ABL1 ≤10% or both at 3 months, which is in line with the optimal response defined in the 2013 ELN criteria [1]. Response categories at 6 months were defined as per 2013 ELN guidelines [1].
For evaluation of responses, priority was given to BCR-ABL1 IS values, which could be substituted by cytogenetic results if BCR-ABL1 IS values were unavailable or with raw (i.e. non-IS) BCR-ABL1 values if neither BCR-ABL1 IS nor cytogenetic data were available. This response evaluation was performed in this order, irrespective of whether additional parameters for evaluation were available and whether concordance between the response parameters were given or not. An exploratory analysis was performed to assess the concordance between BCR-ABL1 IS and BCR-ABL1 raw values [12].
A PFS was defined as the absence of progression to AP/BC (as per ELN guidelines) as labeled by the investigator [1] and OS and PFS rates were expressed as the percentage of all patients with data available for the respective endpoint.

Statistical analysis
For all descriptive statistical comparisons of categorical data (frequencies and percentages) χ 2 -tests were used. Continuous data were compared using Mann-Whitney U-tests. 416 A multicenter retrospective evaluation of Chronic Myeloid Leukemia (CML) therapy in Austria assessing the. . .

Patient characteristics and initial therapy
Of the 211 patients, 176 (median age, 56 years; male, 59.7%, n = 105) were included in the analysis ( Table 1). The remaining 35 patients were excluded due to lack of clearly documented BCR-ABL1 levels or cytogenetic data at the 3-month visit.
Since imatinib was the only approved first-line TKI until the end of 2010, all patients assessed received first-line therapy with imatinib. Of the 176 patients, 164 (93.2%) started on 400mg/day, 11 patients (6.3%) started on lower doses, and 1 patient (0.6%) started on 600mg/day.
The median age was comparable in ETR+ and ETRpatients with 55 years and 59.5 years, respectively.
The median duration until the last follow-up/visit was 94.5 months (7.9 years), which was similar between the ETR+ and ETR-subgroups (97.0 months and 91.5 months, respectively, censoring patients who had died within the documentation period in order to avoid bias through the higher number of deaths in the ETR-group). Patients in the ETR+ group had original article higher 5-year OS and PFS rates compared with the ETR-group (OS: 92.5% vs. 77.5%, P = 0.018; PFS: 95.6% vs. 87.5%, P = 0.06). At the last follow-up, the differences between ETR+ and ETR-groups in OS and PFS were even more pronounced (OS: 88.1% vs. 67.5%, P = 0.003; PFS: 92.6% vs. 84.2%, P = 0.055) (Fig. 1).

Outcomes based on responses at 6 months
Of the 159 patients with available data for the 6-month visit, 101 (63.5%) reached an optimal response at 6 months, whereas 58 (36.5%) did not. Patients in optimal response at 6 months achieved higher OS and PFS rates at 5 years compared with patients who had not reached an optimal response at 6 months (OS: 95% vs. 84.5%, P = 0.02; PFS: 98.0% vs. 87.9%, P = 0.008). Similar results were seen at the last followup with higher OS and PFS rates in patients with optimal response vs. those without optimal response at 6 months (OS: 90.1% vs. 77.6%, P = 0.03; PFS: 96.0% vs. 82.8%, P = 0.004) (Fig. 2).  (Fig. 3).

Kinetics of responses
The ELN has defined molecular and/or cytogenetic criteria for responses (optimal, warning and failure, respectively) to TKI therapy at certain time points [1]. Interestingly, the majority of patients in the warning category at 3 months-defined as BCR-ABL1 >10% and/or 36-95% Ph+ metaphases-(n = 32) remained in warning category at 6 months (n = 13) or had a treatment failure (>95% Ph+ metaphases; n = 8) (Fig. 4). None of the patients in the warning category at 3 months switched to another TKI before month 6, and only 2 patients had a dose escalation of imatinib to 600 mg/day; however, 1 patient remained in the warning category and 1 patient had treatment failure at 6 months.

Therapy at last documented visit
At the last documented visit, more patients had been switched from imatinib to another TKI in the ETR-

Concordance of BCR-ABL1 raw and BCR-ABL1 IS
An additional analysis was performed to assess the feasibility of using BCR-ABL1 raw values for scoring the early responses at 3 and 6 months. Similar percentages of patients were found to have achieved ETR+ when using only BCR-ABL1 raw values vs. only BCR-ABL1 IS values (raw: 76.8% ETR+, 23.2% ETR-; IS: 71.6% ETR+, 28.4% ETR-). The OS at 5 years was also comparable between patients in whom response to TKI was analyzed by using BCR-ABL1 raw and patients in whom BCR-ABL1 IS values were obtained (raw: 87.1% for ETR+, 78.9% for ETR-; IS: 95.7% for ETR+, 79.3% for ETR-). This analysis supported the approach of using BCR-ABL1 raw values for scoring the early treatment responses in patients for whom IS values or cytogenetic data were unavailable (which affected 33 of 176 patients for ETR).

Discussion
This study aimed to elucidate whether the survival benefit seen with early responses in controlled clinical trials [2][3][4][5][6][7][8][9][10] can also be translated into daily clinical practice in Austria. Therefore, this study retrospectively analyzed charts from 12 major institutions across Austria of patients diagnosed with CML-CP between 2004 and 2010. As quantification of BCR-ABL1 levels by IS was not established at all participating sites during the project period, this study evaluated ETR (BCR-ABL1 ≤10% and/or ≤35% Ph+) instead of a purely molecular response at 3 months, which is in concordance with the 2013 ELN criteria [1]. Of note, in this study the substitution of non-IS BCR-ABL1 values (BCR-ABL1 raw ) for scoring early responses of patients for whom neither BCR-ABL1 IS nor cytogenetic data were available was accepted. This original article

Fig. 5
Therapy at last documented visit. ETR early treatment response (partial cytogenetic response and/or BCR-ABL1 ≤ 10% at 3 months)  [4,9]; DASISION, 64% on imatinib reached EMR [2,10]; BFORE, 57.3% on imatinib reached EMR [14]). It is of note that a substantial number of patients on imatinib do not reach optimal early responses in daily routine. Also, the negative impact of intermediate or high risk Sokal scores at diagnosis seen in clinical trials was replicated in this study. Patients in the low-risk group were more likely to reach ETR and have higher 5-year OS and PFS rates compared with the intermediate/high-risk group.
Patients in the ETR+ group had higher 5-year OS and PFS rates than patients in the ETR-group (92.5% vs. 77.5%, P = 0.018). In the imatinib arm of the EN-ESTnd trial, the OS and PFS rates at 5 years were >95% for patients reaching EMR, whereas the OS was 79.1% and PFS was 79.3% for patients who failed to reach EMR [4]. Similarly, in the imatinib arm of the DA-SISION trial, the 5-year OS and PFS rates of patients reaching EMR were 95.4% and 93.1%, respectively, and were 80.5% and 71.9%, respectively, for patients not reaching EMR [2]. Within the limits of cross-trial comparisons, the measured efficacy parameters obtained in this study appear to be comparable. The 5-year OS and PFS calculated for the R-EFECT study are slightly lower than that seen in the large clinical trials. This is possibly driven by the more heterogeneous unselected patient population, especially including those with more comorbidities seen in daily clinical practice. Notably, comorbidities have a strong impact on survival in CML patients treated with TKIs, although this holds true primarily for the second and third generation TKIs [15][16][17] but not for imatinib [18][19][20]. Overall, the results confirm the efficacy and long-term treatment outcome of imatinib and show that the predictive value of reaching ETR for better survival does translate to daily clinical routine.
It is not unexpected that patients failing ETR were more likely to switch from imatinib to another TKI. Interestingly, patients who were in the warning category at 3 months as per ELN guidelines [1] were more likely to remain in the warning category or worsen leading to treatment failure by 6 months, rather than to improve to an optimal response. Of the 32 patients in warning at 3 months, 13 were still in warning, 8 had a treatment failure and only 9 showed an optimal response by 6 months. None of the 32 warning patients have switched to another TKI before the 6-month visit. Interpretations in this respect have to be made with caution since for large parts of the diagnosis period of this study, imatinib was the only approved TKI in Austria. It remains to be determined if a consequent early switch to other TKIs in patients failing ETR is associated with an improved outcome both in clinical trials and in current routine in CML care. Currently, only limited data are available suggesting that the switch to a second generation TKI may improve outcome for patients not achieving an ETR [21].

K original article
22 of the 283 patients (~8%) died by 5 years, including 16 (5.7%) CML-related deaths [4]. Similarly, in the imatinib arm of the DASISION study, 26 of 258 patients (10%) died by 5 years and 17 (6.6%) of these deaths were CML-related [2]. Due to the retrospective nature of the current study, attribution of a cause of death from routine records may have been difficult and must therefore be considered with caution.
In summary, these data from a real-life setting support the findings from randomized trials demonstrating that ETR achievement is associated with superior OS and PFS. Patients who achieved ETR are less likely to switch from imatinib to other TKIs, and patients who fail to achieve ETR should be monitored closely and treated according to available guidelines.

Limitations
Since this study relied exclusively on retrospective data, the observations made here have to be complemented with prospective data before drawing definitive conclusions. Prospective registries might be able to close this gap. In addition, according to the Austrian Pharmaceutical Act, this was an observational study with no hypothesis testing and no control group. There were no local law requirements for monitoring [9]. Consequently, no monitoring was applied and source data verification was not applicable. Missing data were not queried, resulting in a deliberate number of missing data points. A total of 12 major CML treatment centers in Austria participated in the study; however, a bias in the patients selected cannot be excluded.