Evaluating comorbidity scoring systems for flumatinib therapy in chronic myeloid leukemia: a machine learning and SHAP-based predictive analysis - Scorecard - MDSpire
Advertisement
Evaluating comorbidity scoring systems for flumatinib therapy in chronic myeloid leukemia: a machine learning and SHAP-based predictive analysis
Clinical Scorecard: Assessing Comorbidity Scoring Systems for Flumatinib Treatment in Chronic Myeloid Leukemia: A Predictive Analysis Utilizing Machine Learning and SHAP Techniques
At a Glance
Category
Detail
Condition
Chronic-phase chronic myeloid leukemia (CP-CML)
Key Mechanisms
Comorbidity scoring systems (CCI, ACE-27, CIRS-G) to predict molecular responses
Target Population
Patients with CP-CML receiving flumatinib
Care Setting
Retrospective cohort study at a single center
Key Highlights
CIRS-G showed the strongest predictive value among comorbidity scores.
XGBoost model achieved an AUC of 0.852 for predicting major molecular response.
CIRS-G score ≥ 8 may severely compromise therapeutic efficacy.
Advanced age and severe comorbidity are associated with lower predicted probabilities.
Machine learning and SHAP analysis provide insights for individualized risk stratification.
Guideline-Based Recommendations
Diagnosis
Utilize comorbidity scoring systems to assess patient risk.
Management
Consider comorbidity burden when prescribing flumatinib.
Monitoring & Follow-up
Monitor molecular responses at 6 and 12 months of therapy.
Risks
Increased adverse drug reactions and reduced compliance due to comorbidities.
Patient & Prescribing Data
559 patients with CP-CML treated with flumatinib from 2018 to 2024.
Initial dose of flumatinib recorded; reduced doses may alter model predictions.
Clinical Best Practices
Integrate comorbidity indices into treatment planning.
Utilize machine learning for predictive analytics in treatment outcomes.
