Background:

Chronic myeloid leukemia (CML) is rare in children, with an estimated incidence of one per million annually. Pediatric CML exhibits a distinct breakpoint distribution in the BCR region and a higher proportion of genomic alterations compared to adults. Prognostic scores are essential for predicting outcomes and guiding therapy in CML. However, the utility of existing adult-derived scores, especially the EUTOS Long-Term Survival (ELTS) score, remains uncertain in children. This study evaluates the prognostic performance of the ELTS score in pediatric patients with chronic phase CML (CML-CP).

Objective:

To assess the prognostic value of the ELTS score in predicting long-term CML-specific survival among children diagnosed with CML-CP.

Materials and Methods:

We conducted a retrospective analysis of children (<18 years) diagnosed with CML-CP between 2001 and 2021 at our center. Baseline demographic, clinical, and laboratory data, including prognostic parameters for Sokal, EUTOS, and ELTS scores, were collected from medical records. The ELTS score was calculated for all patients. In order to identify the discriminative performance of the ELTS score, Sokal and EUTOS were included as comparators. Event-free survival (EFS) was defined as the time from diagnosis to progression (loss of hematologic response or transformation to accelerated/blast phase) or death. CML-specific overall survival (OS) was measured from diagnosis to CML-related death.

Results:

Of 178 children evaluated, 145 were included after excluding those with advanced disease (n=18) and incomplete baseline data (n=15). The median age was 14 years (range, 2–18), with 66.2% (96/145) male patients. The most common symptoms were abdominal discomfort (58%) and fatigue (43%). The median hemoglobin was 9.3 g/dL (range,4-13), median WBC count 1,84,000/µL (range, 5,900-9,89,600), median platelet count 4.2 lakhs/µL (range, 20,000-14,00,000), median blast percentage 3% (range, 0-14), and median spleen size 12 cm (range, 0-24).

Based on ELTS, 31.7% were low-risk, 41.4% intermediate-risk, and 26.9% high-risk. Sokal stratification categorized 35.9% as low-, 38.6% as intermediate-, and 25.5% as high-risk. The median follow-up was 174 months. The 12-month major molecular response rate was 54%. Median EFS was 131 months (IQR: 51–177.5). Due to limited access to second-generation TKIs, dose escalation was the main strategy for suboptimal responders. Median OS was 150 months (IQR: 84-192).

Median OS across ELTS risk groups was:

  • Low-risk: 156 months (IQR: 113–195)

  • Intermediate-risk: 149.5 months (IQR: 93–192.5)

  • High-risk: 145 months (IQR: 30–179)

Five-year OS rates were 95.7% (low), 83.3% (intermediate), and 69.2% (high-risk), with the ELTS score demonstrating significant survival discrimination (log-rank p < 0.001). In comparison, the Sokal score showed a non-significant trend (p = 0.078), and the EUTOS score lacked discriminatory power (p = 0.23). C-index for ELTS was 0.671, superior to Sokal (0.594) and EUTOS (0.543). A combined model yielded only a marginal improvement (C-index = 0.683).

On multivariate analysis, only the ELTS score retained independent prognostic significance:

  • Intermediate-risk: HR = 4.35 (95% CI: 1.43–13.23, p = 0.010)

  • High-risk: HR = 8.96 (95% CI: 2.53–31.81, p < 0.001)

Notably, 3 patients categorized as low-risk by Sokal were high-risk by ELTS and had poor outcomes.

Conclusion:

Our findings provide real-world evidence supporting the prognostic utility of the ELTS score in pediatric CML. ELTS outperformed Sokal and EUTOS in risk stratification and independently predicted survival outcomes. These results underscore the need for further prospective, multi-center validation in children with CML-CP.

This content is only available as a PDF.
2025
Sign in via your Institution