Evaluating comorbidity scoring systems for flumatinib therapy in chronic myeloid leukemia: a machine learning and SHAP-based predictive analysis - Summary - MDSpire
Advertisement
Evaluating comorbidity scoring systems for flumatinib therapy in chronic myeloid leukemia: a machine learning and SHAP-based predictive analysis
To evaluate the predictive value of three comorbidity scoring systems (CCI, ACE-27, and CIRS-G) for molecular responses at 6 and 12 months in flumatinib-treated CP-CML patients.
Key Findings:
XGBoost model achieved the highest predictive performance with an AUC of 0.852, indicating strong predictive capability.
CIRS-G integration into the baseline model provided the most significant incremental value (ΔAUC = 0.078, p = 0.006).
A CIRS-G score ≥ 8 may severely compromise therapeutic efficacy.
Advanced age and severe comorbidity were associated with lower predicted probabilities.
Interpretation:
CIRS-G demonstrated the strongest predictive value among the comorbidity scores evaluated, with machine learning and SHAP analysis offering insights for individualized risk stratification, which is crucial for optimizing treatment outcomes.
Limitations:
Predictive performance may overestimate prospective validity due to temporal variations.
Retrospective design may introduce biases, affecting the reliability of the findings.
Conclusion:
CIRS-G is the most predictive comorbidity score for flumatinib treatment outcomes in CP-CML patients, highlighting the need for personalized treatment approaches.
