The Use of EUTOS Long-Term Survival Score Instead of Sokal Score Is Strongly Advised in Elderly Chronic Myeloid Leukemia Patients

Background. Many prognostic scores have been proposed for risk stratification of chronic myeloid leukemia (CML) patients: the Sokal is the oldest and the most widely used score; the EUTOS long-term survival (ELTS), based on a large cohort of patients treated frontline with imatinib (IM), is the most recent score. The ELTS score, compared to Sokal score, showed superior ability to predict overall and leukemia-related survival, but data are limited and differences were small. Further evidence is required to support and to implement the clinical use of ELTS score.

Aims. Given the different weight of the variable "age" in ELTS and Sokal score formulations, we hypothesized a different predictive value in specific age groups. Consequently, the aim of our study was to compare the prognostic value of ELTS and Sokal scores in a cohort of CML patients treated frontline with tyrosine kinase inhibitors (TKIs), according to the age, < 30 years, 30-64 years, or ≥ 65 years old.

Methods. Nine hundred and four adult patients were included, 559 treated with IM and 345 treated with nilotinib (NIL). Patients were enrolled in six multicenter studies (NCT00481052, NCT00769327, NCT01535391, NCT00514488, NCT00510926, observational trial CML/023) conducted by the GIMEMA CML WP. The intention-to-treat population of each study was analyzed. Definitions: major molecular response (MR3 or MMR), BCR-ABL^IS <0.1%; deep molecular response (MR4), BCR-ABL^IS <0.01% with > 10.000 copies; progression, transformation according to ELN criteria; leukemia related death (LRD): death after progression.

Results. Median age, 52 years (range 18-86). Age distribution: < 30 years, 68 pts (8%); 30-64 years, 634 pts (70%); ≥ 65 years 202 pts (22%). Median follow-up 77 months (range: 24-109 years). The risk according to the two scoring systems was as follows: 57% low, 30% intermediate and 13% high ELTS score, 40% low, 39% intermediate and 21% high Sokal score, respectively; in elderly patients (≥ 65 years), 24% low, 54% intermediate and 22% high ELTS score, 9% low, 70% intermediate and 21% high Sokal score, respectively. The risk distributions were comparable in patients treated with IM or NIL. The concordance between the two scores, in particular in the low (L) and the high (H) risk categories, was good in patients < 30 years (87% L-L and 80% H-H, respectively) or 30-64 years old (68% L-L and 85% H-H respectively); in contrast, in patients ≥ 65 years old, only 8% of low ELTS patients had a low Sokal score, and only 48% of high ELTS score had a high Sokal score. Overall, both scores were able to predict significantly different probabilities of MR3, MR4, overall survival (OS) and LRD, but in elderly patients (> 65 years) only the ELTS score was able to predict the achievement of MR3 (99%, 87% and 75% in low, intermediate and high-risk patients, respectively; p=0.001) and MR4 (82%, 61% and 50% in low, intermediate and high ELTS score patients, respectively; p=0.005). Interestingly, in elderly patients both scores predicted the OS, while only the ELTS score predicted a significantly different LRD probability (cumulative incidence 2%, 6% and 14% in low, intermediate and high-risk patients, respectively; p=0.049). The results were similar considering patients < 30 years, 30-59 years, or ≥ 60 years old.

Summary/Conclusion. The risk distribution according to ELTS and Sokal score and the concordance between the two scores was different in young adults (< 30 years), adults (30-64 years) and elderly (≥ 65 years) patients, and the number of patients potentially misclassified by the Sokal score was particularly relevant in the elderly group. In elderly CML patients treated with IM or NIL as frontline therapy the ELTS score was able to predict the achievement of MR3 and MR4 and long-term leukemia-related survival, whereas the Sokal score was not able to find any significant difference. Consequently, especially in elderly patients, the use of ELTS score is strongly recommended to assess the baseline disease-risk and to select patients candidate to a frontline treatment with second generation TKIs, minimizing the risk of unnecessary over-treatment.