A study of risk distribution scores for patients with chronic myeloid leukemia (CML) indicated that risk distribution according to EUTOS long-term survival (ELTS) and Sokal score and concordance between the scores was different between different age groups of patients with the disease.
The ELTS is the most recent prognostic score for risk stratification of patients with CML and has been shown to be superior for predicting overall survival (OS) and leukemia-related survival. However, data are limited and differences between the two systems are small.
Fausto Castagnetti, MD, PhD, of the University of Bologna, Italy, recently presented findings of a study comparing the two systems at the 60th ASH Annual Meeting and Exposition.
Many prognostic scores have been proposed for risk stratification; the ELTS score is based on a large cohort of patients treated with imatinib as a first-line therapy, and is specifically developed to discriminate the probability of dying from CML.
“Given the different weight of the variable ‘age’ in ELTS and Sokal score formulations, we hypothesized a different predictive value of the two scores in young patients (under 30 years old), adults (between the ages of 30 and 64), and elderly patients (those 65 years and older),” Castagnetti said.
Of the 904 included patients, 559 had been treated with imatinib as first-line therapy and 345 had been treated with nilotinib as first-line therapy. Patients were enrolled in six multicenter studies (NCT00481052, NCT00769327, NCT01535391, NCT00514488, NCT00510926, observational trial CML/023) conducted by the GIMEMA CML WP.
The majority of patients (70%) were adults, with 22% (n=202) elderly; median follow-up was 6.4 years.
In the young and adult patient groups, the risk distribution was comparable between the Sokal and ELTS, but differed greatly in elderly patients in predicting low and intermediate risk patients. In elderly patients (≥ 65 years old), the Sokal score suggested 9%, 70%, and 21% were at low, intermediate, or high risk, respectively. This differed considerably from the ELTS score, where 24%, 54%, and 22% were at low, intermediate, or high risk, respectively. Only 8% of low ELTS patients had a low Sokal score, and only 48% of high ELTS score had a high Sokal score.
“Half of the intermediate and high Sokal risk patients were potentially misclassified,” Castagnetti said. “One-third of patients stratified by ELTS score have a different Sokal score (in any risk category). These discrepancies were more evident in elderly patients.”
Similar outcomes were found when stratifying patients by molecular response by risk. Overall, both scores were able to predict significantly different probabilities of MR3, MR4, OS, and leukemia-related survival, but in elderly patients 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=.001) and MR4 (82%, 61%, and 50% in low-, intermediate-, and high-ELTS score patients, respectively; P=.005).
Interestingly, in elderly patients both scores predicted the OS, while only the ELTS score predicted a significantly different leukemia-related death probability (cumulative incidence 2%, 6%, and 14% in low-, intermediate-, and high-risk patients, respectively; P=.049).
“Consequently, especially in elderly patients, the use of ELTS score is strongly recommended to assess the baseline disease-risk and to select patient candidates to a frontline treatment with second-generation TKIs, minimizing the risk of unnecessary over-treatment,” Castagnetti said.