Background: Including Chronic Myeloid Leukemia (CML) patients with deep molecular responses (MR4.5) and those with suboptimal responses provides valuable insights into treatment-associated metabolic changes. This study aimed to characterize the metabolomic alterations associated with CML and identify potential biomarkers for treatment response, particularly in patients achieving a deeper molecular response versus those with poorer responses. Methods: Plasma samples were collected from 51 chronic-phase CML patients and 24 healthy controls. CML patients were classified into two groups based on molecular responses: T1 (BCR-ABL1 IS ≤ 0.0032%) and T2 (BCR-ABL1 IS > 0.0032%, 1.5 included glutamate, hypoxanthine, and D-galactonic acid. In the T2 group, significant increases in malate and 5-aminoimidazole-4-carboxamide ribonucleotide were observed, reflecting disruptions in purine metabolism, the tricarboxylic acid cycle, and amino acid metabolism. The pathway enrichment analysis highlighted significant alterations in CML energy metabolism, nucleotide synthesis, and amino acid biosynthesis. Conclusions: CML patients exhibit pronounced metabolic changes, particularly in energy and nucleotide metabolism, which are linked to treatment response. These findings provide novel insights into CML biology and suggest potential biomarkers for monitoring treatment efficacy and predicting outcomes and therapeutic targets for improving treatment outcomes and overcoming tyrosine kinase inhibitor resistance.
Background: Multiple myeloma (MM) diagnosis requires specialised testing that is not universally available, contributing to diagnostic delays. The aim of this study was to develop and internally validate statistical and machine learning (ML) models that discriminate MM from clinically similar conditions using exclusively routine laboratory parameters. Methods: This retrospective diagnostic prediction model study included 422 consecutive patients referred to hematology for suspected MM at a tertiary care center (2017–2025). Fifteen routine demographic and laboratory variables were used to train seven ML algorithms and two ensemble methods. Model performance was evaluated on a held-out test set (30%) using the area under the receiver operating characteristic curve (AUC) with bootstrap 95% confidence intervals, calibration metrics, and decision curve analysis. Internal validation comprised repeated 10-fold cross-validation and bootstrap optimism correction (1000 resamples). Results: Of 422 patients, 203 (48.1%) were diagnosed with MM. Random Forest, Gradient Boosting, and Ensemble Stacking showed similarly high discrimination on the test set (AUCs 0.950–0.955). Total protein (multivariable-adjusted OR: 11.95), calcium (OR: 9.62), hemoglobin (OR: 4.34), and albumin (OR: 0.16) were the strongest independent predictors. Ensemble Stacking showed the best calibration (Brier score 0.084; calibration slope 1.14). At a ≥95% sensitivity threshold, Random Forest maintained 77.3% specificity. Risk stratification effectively separated patients, with observed MM rates ranging from 2.4% (very low risk) to 97.2% (very high risk). Conclusions: Machine learning models using routine laboratory parameters achieved excellent diagnostic performance for discriminating MM from its clinical mimickers in a real-world referral cohort. Following further development with larger datasets and external validation, these models could support triage and prioritization of patients already under evaluation for suspected MM, particularly where access to specialized testing is delayed; use in lower-prevalence, first-contact settings would require recalibration and dedicated validation.
Background/Objectives: Essential thrombocythemia (ET) is a JAK-STAT-driven myeloproliferative neoplasm with an incompletely defined metabolic phenotype. This study aimed to characterize the plasma metabolome of ET and to determine whether it differs according to JAK2V617F status, platelet count (≥450 vs. 1.5 or <0.67) and a Benjamini–Hochberg false discovery rate (q) below 0.05; identifications were annotated at Level 1 (retention time-matched authentic standards) or Level 3 (molecular formula), and pathways were examined by over-representation analysis in Reactome. Results: PCA and PLS-DA separated ET from controls (PLS-DA R2Y = 0.96; nested cross-validation Q2 = 0.91; permutation p = 0.001). Nineteen metabolites met the criteria: seven Level 1 identifications (succinate, an aminobutyrate isomer, uridine and N-acetylphenylalanine higher; proline, glucuronolactone and quinaldate lower in ET) and twelve Level 3 candidates dominated by platelet-derived eicosanoids, including 11-dehydro-thromboxane B2 (detected in 26% of controls and 99% of patients) and thromboxane B2 (4% and 84%), prostaglandin D2/E2 and B1/A1 isomers, a hydroxy-docosahexaenoic acid isomer, and corticosteroid intermediates. Reactome over-representation analysis, which rests mainly on Level 3 candidates, ranked synthesis of prostaglandins and thromboxanes and arachidonate metabolism highest (FDR 7 × 10−5), followed by fatty acid metabolism, pyrimidine catabolism and, on the basis of Level 3 candidates only, mineralocorticoid biosynthesis. The signature was similar in direction across cytoreductive-therapy, platelet-count and JAK2V617F strata. Conclusions: These exploratory data describe a plasma metabolite signature of aspirin-treated ET dominated by platelet-derived eicosanoids, together with changes in succinate and pyrimidine-related metabolites and candidate changes in corticosteroid intermediates; because aspirin exposure and disease status were fully confounded, the signature requires confirmation with targeted methods in independent cohorts.