Changing Treatment May Affect the Predictive Ability of European Treatment Outcome Study Scoring for the Prognosis of Patients with Chronic Myeloid Leukemia

Objective: Previous studies compared the predictive ability of the European Treatment Outcome Study (EUTOS), Sokal, and Hasford scoring systems and demonstrated inconsistent findings with unknown reasons. This study was conducted to determine a useful scoring system to predict the prognosis of patients with chronic myeloid leukemia (CML) and identify the probable factors that affect the scoring. Materials and Methods: This is a retrospective cohort study. The predictive ability of EUTOS and the factors that affect scoring were analyzed in 234 Chinese chronic-phase CML patients treated with frontline imatinib, including a few patients temporarily administered hydroxyurea for cytoreduction before imatinib. Patients were stratified into different risk groups according to each scoring system to assess the treatment outcomes and the predictive ability of EUTOS scores between patients who received imatinib during the entire follow-up period and patients who received altered treatment because of intolerance, progression, and treatment failure. Results: Sixty-one (26.0%) patients received altered treatments during the follow-up. In the EUTOS low- and high-risk groups, the 5-year overall survival was 94.6% and 84.7% (p=0.011), 5-year event-free survival was 92.6% and 77.6% (p=0.001), and 5-year progression-free survival (PFS) was 95.3% and 82.4% (p=0.001), respectively. The predictive ability of EUTOS was better than that of the Sokal and Hasford scores (p=0.256, p=0.062, p=0.073) without statistical significance. All three scoring systems were valid in predicting early optimal response. Kaplan-Meier analysis showed a high association between overall PFS and the EUTOS scores in the standard-dose imatinib group (p<0.001). Conclusion: This study suggests that the EUTOS scoring system could predict the outcome of chronic-phase CML patients treated with standard-dose imatinib. Altered treatment is a crucial factor that affects the prognostic impact of EUTOS scoring. Achieving complete cytogenetic response at 18 months is an essential factor in predicting the prognosis of patients with CML.


Introduction
As the firstline treatment for chronic myeloid leukemia (CML), imatinib is widely used after diagnosis and dramatically improves the overall survival (OS) of CML patients [1]. Predicting the prognosis is significant for the management of CML patients. Currently, the European Treatment Outcome Study (EUTOS), Hasford, and Sokal prognostic scoring systems are used for predicting the prognosis of CML patients [2,3,4]. The EUTOS scoring system is a novel prognostic scoring system that challenges the conventional Sokal and Hasford scoring systems in predicting the outcome of CML patients. However, recent studies examining the effectiveness of the EUTOS scoring system in predicting the prognosis of CML patients showed controversial results. For example, several studies from different regions of the world compared the clinical significance of the three prognostic scoring systems. Five studies found that EUTOS was better than the Hasford and Sokal systems in predicting the prognosis of CML patients [1,2,3,5,6]. In contrast, 3 studies showed that the EUTOS score does not predict prognosis in CML patients [7,8,9]. It is currently unknown what factors caused these controversial findings.
The purpose of this study was to compare the predictive ability of the Sokal, Hasford, and EUTOS prognostic scoring systems by stratifying CML-chronic-phase (CP) patients who received firstline imatinib mesylate at diagnosis into different risk groups. The possible factors that affect the prognostic ability of EUTOS were further explored according to the three scoring systems.

Patients
A total of 234 CML-CP patients (162 males, 72 females) who received imatinib mesylate (Novartis Oncology, Novartis Pharma Stein AG, Stein, Switzerland) treatment within 6 months of diagnosis at X Hospital between January 2004 and July 2014 were recruited for this study. CML-CP was diagnosed according to published diagnostic criteria [10], and all patients were treated with a standard dose of imatinib (400 mg/day) over 3 months. No other treatment was given, except for hydroxyurea temporarily administered for cytoreduction before imatinib in 9 patients.

Calculations of the Chronic Myeloid Leukemia Prognostic Indexes
The Sokal score was calculated using the following formula: Exp 0.0116 × (age in years-43.4) + 0.0345 × (spleen size-7.51) + 0.188 × [(platelet count/700) 2 -0.563] + 0.0887 × (blast cells-2.10). Patients with a score of less than 0.8 were assigned to the low Sokal risk group, patients with a score from 0.8 to 1.2 were assigned to the intermediate Sokal risk group, and patients with a score greater than 1.2 were assigned to the high Sokal risk group [11]. The Hasford score was calculated as follows: 0.666 (when age >50 years) + (0.042×spleen size) + 1.0956 (when platelet count >1500×10 9 /L) + (0.0584×blast cell count) + 0.20399 (when basophil count >3%) + (0.0413×eosinophil count) × 100. Patients with a score of less than 780 were assigned to the low Hasford risk group, patients with a score from 781 to 1480 were assigned to the intermediate Hasford risk group, and patients with a score higher than 1480 were assigned to the high Hasford risk group [12]. The EUTOS score was calculated as follows: (7×basophil count) + (4×spleen size), where the spleen was measured in centimeters below the costal margin and basophils as a percentage rate. Patients with a EUTOS score higher than 87 were assigned to the high EUTOS risk group, while patients with a EUTOS score of less than or equal to 87 were assigned to the low EUTOS risk group [10]. Event-free survival (EFS): the length of time from the date of initiating imatinib therapy to the date of failure according to the European Leukemia Net criteria, the date of stopping treatment due to imatinib intolerance, or the date of last follow-up in patients whose treatments did not fail [13].
Progression-free survival (PFS): the length of time from the date of imatinib therapy initiation to the date of progression to accelerated phase (AP)/blastic phase (BP) or to the date of death.
Complete cytogenetic response (CCyR): no Philadelphia chromosome was detected in the patient by G-banding analysis of bone marrow and no Philadelphia cell was detected in the patient when using fluorescence in situ hybridization analysis of peripheral blood.

Statistical Analysis
Data were analyzed using SPSS 17.0 (SPSS Inc., Chicago, IL, USA). Normally distributed continuous variables were presented as mean ± standard deviation, and non-normally distributed continuous variables were presented as medians with interquartile ranges. Kaplan-Meier methods and log rank tests were applied to analyze the time-to-event data. The 5-year EFS, PFS, and OS and the cumulative incidence of PCyR, CCyR, and MMR were compared using the chi-square test. A value of p<0.05 was considered statistically significant.

Baseline Characteristics
The mean age at diagnosis was 40.3±13.2 years. Median followup duration of imatinib treatment was 20.5 months (range: 9-120 months) months. The median duration from diagnosis of CML to imatinib initiation was 19 (range: 6-115) days. The CCyR rates at 12 months and 18 months were 56.4% and 65.4%, respectively.
According to the EUTOS scoring system, 149 patients (63.7%) were classified as low-risk and 85 patients (36.3%) were classified as high-risk. Using the Sokal scoring system, 66, 70, and 98 patients were classified into low-, intermediate-, and high-risk groups, while 94, 93, and 47 patients were classified into low-, intermediate-, and high-risk groups when the Hasford scoring system was used. The baseline characteristics of patients with CML-CP are shown in Table 1.

Outcomes
The 5-year OS, EFS, and PFS of the 234 patients were 91.0%, 87.2%, and 90.6%, respectively. The 5-year OS, EFS, and PFS were 94.6% and 84.7% (p=0.011), 92.6% and 77.6% (p=0.001), and 95.3% and 82.4% (p=0.001) for low and high EUTOS risk groups, respectively. Significant differences in 5-year OS, EFS, and PFS were observed between the low EUTOS and high EUTOS risk groups. However, there were no significant differences between groups classified by the Sokal (p=0.137, p=0.106, p=0.110, respectively) and Hasford (p=0.256, p=0.062, p=0.073, respectively) prognostic scoring systems. Moreover, Kaplan-Meier analysis of OS, EFS, and PFS showed significant differences between the low and high EUTOS risk groups (p<0.001, p<0.001, p<0.001) (Figure 1), but no significant difference was  There were significant differences in 5-year OS between low and high EUTOS risk score groups in patients that received standard-dose imatinib treatment (97.5% and 89.1%, p=0.03), whereas it was not different in the patients who received altered treatments (83.9% and 76.7%, p=0.479). Furthermore, Kaplan-Meier analysis showed a high association between overall PFS and EUTOS scores in patients who received standard-dose imatinib treatment (p<0.001), but no significant correlation in patients who received altered treatment (p=0.246) (Figure 2).

Discussion
Previous studies presented controversial findings on whether EUTOS is more useful in predicting the survival of CML patients than the Sokal and Hasford scoring systems. This study demonstrated that the EUTOS, Hasford, and Sokal scoring systems were all effective in predicting early optimal response (PCyR at 3 months, CCyR at 12 months and 18 months, and MMR at 18 months) in CML-CP patients who were treated with imatinib as frontline therapy, but the EUTOS scoring system was more effective as a prognostic indicator of OS, EFS, and PFS than the Sokal and Hasford scoring systems. A novel finding in this study is that altered treatment is a key factor that affects the prognostic ability of the EUTOS scoring system.
Among the 3 studies not supporting the predictive effect of EUTOS in the prognosis of CML patients [7,8,9], 33.6% of 282 CML-CP patients in the study of Marin et al. were treated with imatinib as frontline therapy, but the treatment was changed to a second-generation tyrosine kinase inhibitor (2 nd TKI) [9]. In the study of Jabbour et al. [8], 84.7% of 465 CML-CP patients were Therefore, CML patients who did not achieve CCyR at 18 months were more likely to progress AP/BP or death and CCyR at 18 months should be considered as an essential end-point to assess the survival of CML patients.

Conclusion
In conclusion, the EUTOS prognostic scoring system is an effective prognostic tool in assessing the outcome of CML-CP patients who were treated with standard-dose imatinib. Altered treatment should be considered as a key factor when using the EUTOS scoring system. CCyR at 18 months during therapy is an essential end-point to assess the survival of CML-CP patients. Our findings may have important clinical implications, but they should be confirmed in a larger cohort and validated in a prospective observational study.

Acknowledgments
We are thankful to Fu Bin for revising the manuscript. This work did not receive any specific grant from any funding agency in the public, commercial, or not-for-profit sector.

Ethics
Ethics Committee Approval: This is a retrospective cohort study; Informed Consent: This is a retrospective cohort study.

Figure 2.
Progression-free survival using European Treatment Outcome Study score for chronic myeloid leukemia-chronicphase patients who received imatinib or altered treatment. (A) Progression-free survival using European Treatment Outcome Study score for chronic myeloid leukemia-chronic-phase patients treated with standard-dose imatinib. There was a significant difference between the risk groups (p<0.001). (B) Progressionfree survival using European Treatment Outcome Study score for chronic myeloid leukemia-chronic-phase patients who received altered treatment. There was no significant difference between the risk groups (p=0.246).