AI and Cancer Research ── CML ── 2026-11-05
Chronic myeloid leukemia (CML): treatment, trials and AI
- Abstract
Accurate disease phase assessment remains clinically important in chronic myeloid leukemia (CML), as progression to accelerated or blast phase (AP/BP) is associated with therapeutic resistance and poor prognosis. We aimed to develop an interpretable machine learning (ML) framework integrating peripheral blood long non-coding RNA (lncRNA) expression and clinical variables for disease phase assessment in CML. Using peripheral blood from 305 treatment-naïve CML patients (85 AP/BP; 220 chronic phase) and 90 healthy controls, we identified a progression-associated ten-lncRNA signature via PCR array and real-time quantitative PCR (RT-qPCR) validation. Feature selection using Least Absolute Shrinkage and Selection Operator (LASSO) regression and multivariable logistic regression yielded eight key predictors. Five ML models were trained (n = 214), validated in an independent holdout test cohort (n = 91), and evaluated using the area under the ROC curve (AUC), calibration, and decision curve analysis. Interpretability was achieved using SHapley Additive exPlanations (SHAP). The final model integrated three lncRNAs (CCDC26, SNHG5, FENDRR) and five clinical variables. Among the evaluated algorithms, XGBoost demonstrated the most favorable overall performance, achieving AUCs of 0.9658 and 0.9656 in the training and independent holdout test cohorts, respectively. We developed an interpretable ML framework integrating peripheral blood lncRNAs and clinical variables to support disease phase assessment in CML. The proposed model demonstrated favorable performance within this single-center cohort and may provide complementary information for clinical evaluation. Further multicenter and longitudinal studies are required before broader clinical application.
Journal IF-equivalent: 7.6 (OpenAlex 2-year mean citedness, value as of 2026-11-04, retrieved 2026-11-05; not the official Clarivate IF)Reference: Bai Y, Zheng M, Li H, Song C, Huang B, Cai M, et al. An interpretable machine learning model integrating peripheral blood lncRNAs and clinical variables for phase classification in chronic myeloid leukemia. BMC Med Inform Decis Mak. 2026 Jun 29;26(1):332. doi:10.1186/s12911-026-03656-1. PMID: 42374374; PMCID: PMC13536685.Checked: Full text checked - Abstract
Comorbidities increasingly complicate targeted therapy for chronic-phase chronic myeloid leukemia (CP-CML). This study evaluated three comorbidity scoring systems (CCI, ACE-27, and CIRS-G) to predict molecular responses in flumatinib-treated CP-CML patients. Retrospective data from 559 patients (2018-2024) were analyzed. Machine learning algorithms, including XGBoost, were trained to predict 12-month major molecular response (MMR). SHapley Additive exPlanations (SHAP) analysis was employed to interpret non-linear associations and feature interactions. The XGBoost model demonstrated the highest predictive performance (AUC = 0.852). Integrating CIRS-G into the baseline clinical model provided the most substantial incremental value (ΔAUC = 0.078, p = 0.006), outperforming CCI and ACE-27. SHAP analysis revealed a non-linear threshold effect, suggesting that a CIRS-G score ≥ 8 may severely compromise therapeutic efficacy. Furthermore, interaction plots indicated an exploratory model-based association where advanced age and severe comorbidity clustered with lower predicted probabilities, while reduced dose interventions coincided with altered model predictions in this specific subpopulation. CIRS-G showed the strongest incremental predictive value among the evaluated comorbidity scores in this single-center cohort. Machine learning and SHAP analysis provide exploratory insights for individualized risk stratification, though the reported predictive performance may overestimate prospective validity due to temporal variations and retrospective design.
Journal IF-equivalent: 2.6 (OpenAlex 2-year mean citedness, value as of 2026-10-08, retrieved 2026-10-09; not the official Clarivate IF)Reference: Yang Y, Li Y, Wang J. Evaluating comorbidity scoring systems for flumatinib therapy in chronic myeloid leukemia: a machine learning and SHAP-based predictive analysis. Front Med (Lausanne). 2026 May 28;13:1849735. doi:10.3389/fmed.2026.1849735. PMID: 42292180; PMCID: PMC13253388.Checked: Full text checked - Abstract
Data-driven such as machine learning models in rare disease research faces a significant challenge: scarce data cause performance degradation and unstable explanations that undermine clinical trust. This study investigated whether synthetic data augmentation can maintain prediction performance and improve explainability stability for cardiovascular disease risk prediction in Chronic Myeloid Leukemia patients. Using 235 real-world patients' data, we employed DP-CGANS to generate synthetic records and evaluated their impact across decreasing sample sizes, comparing real-data-only versus real + synthetic data scenarios. Synthetic augmentation showed stable performance while real-data-only models degraded substantially. We observed synthetic data can significantly improve SHAP stability. These findings demonstrate that synthetic data augmentation has potential to address performance degradation and explanation instability in building reliable ML models for diseases with scarce patient data.
Journal IF-equivalent: 1.0 (OpenAlex 2-year mean citedness, value as of 2026-11-04, retrieved 2026-11-05; not the official Clarivate IF)Reference: Sun C, Dumontier M. Enhancing Explainable AI Stability with Realistic Synthetic Data for Cardiovascular Risk Prediction. Stud Health Technol Inform. 2026 May 21 [Epub ahead of print]. doi:10.3233/shti260317. PMID: 42174991.Checked: Abstract only