COMPREHENSIVE LIPIDOMIC PROFILING COMBINED WITH MACHINE LEARNING REVEALS BASELINE SIGNATURES PREDICTIVE OF FIRST LEG ULCER DEVELOPMENT IN SICKLE CELL DISEASE

@article{santos2026comprehensive,
  title = {{COMPREHENSIVE LIPIDOMIC PROFILING COMBINED WITH MACHINE LEARNING REVEALS BASELINE SIGNATURES PREDICTIVE OF FIRST LEG ULCER DEVELOPMENT IN SICKLE CELL DISEASE}},
  author = {Felipe Custódio Carvalho DOS SANTOS and Álex Aparecido Rosini SILVA and Marcos Eduardo Souza ABREU and Fábio Eudes LEAL and André Rolim BELISáRIO and Brian Scott CUSTER and Carla Luana DINARDO and Luiz De Melo Amori FILHO and Scott BOWLER and Lishomwa NDHLOVU and Paula Fraiman Blatyta CASELLI and Dahra TELES and Shannon KELLY and Ester Cerdeira SABINO and Sheila De Oliveira Garcia MATEOS},
  journal = {Hematology, Transfusion and Cell Therapy},
  year = {2026},
  month = oct,
  doi = {10.1016/j.htct.2026.106636},
  url = {https://doi.org/10.1016/j.htct.2026.106636},
  note = {CC BY},
}

TY  - JOUR
TI  - COMPREHENSIVE LIPIDOMIC PROFILING COMBINED WITH MACHINE LEARNING REVEALS BASELINE SIGNATURES PREDICTIVE OF FIRST LEG ULCER DEVELOPMENT IN SICKLE CELL DISEASE
AU  - Felipe Custódio Carvalho DOS SANTOS
AU  - Álex Aparecido Rosini SILVA
AU  - Marcos Eduardo Souza ABREU
AU  - Fábio Eudes LEAL
AU  - André Rolim BELISáRIO
AU  - Brian Scott CUSTER
AU  - Carla Luana DINARDO
AU  - Luiz De Melo Amori FILHO
AU  - Scott BOWLER
AU  - Lishomwa NDHLOVU
AU  - Paula Fraiman Blatyta CASELLI
AU  - Dahra TELES
AU  - Shannon KELLY
AU  - Ester Cerdeira SABINO
AU  - Sheila De Oliveira Garcia MATEOS
PY  - 2026
DA  - 2026/10/23
JO  - Hematology, Transfusion and Cell Therapy
DO  - 10.1016/j.htct.2026.106636
UR  - https://doi.org/10.1016/j.htct.2026.106636
AB  - Introduction: Human T-cell lymphotropic virus type 1 (HTLV-1) infection can lead to different clinical outcomes, such as asymptomatic infection, HTLV-1-associated myelopathy/tropical spastic paraparesis (HAM/TSP), and adult T-cell leukemia/lymphoma (ATLL). The intracellular metabolic differences associated with these phenotypes remain poorly characterized. Aim: This study aimed to characterize and compare intracellular CD4+ T-cell metabolic profiles across HTLV-1 clinical groups. Methodology: We analyzed 63 leg ulcer–naïve participants from the REDS-III Brazil cohort with longitudinal follow-up (median 2.21-years). Baseline plasma lipidomics was acquired by untargeted LC-MS. Supervised classifiers for incident first leg ulcer were developed using leakage-resistant preprocessing and strict separation of model development and evaluation; performance was estimated by stratified five-fold cross-validation within the training set, with a stratified 20% hold-out split used as an internal validation. Results: During follow-up, 18 participants developed a first leg ulcer. Baseline clinical characteristics did not show clear discrimination between individuals who developed leg ulcers and those who remained ulcer-free. Training-set feature selection identified a panel of 17 putatively annotated lipids (MSI level 2), enriched in sphingomyelins and glycerophospholipid-derived lysophospholipids, reflecting biologically coherent lipid remodeling patterns. Models based on this panel achieved consistent discrimination (cross-validated AUC 0.80–0.88) and reproduced a high-specificity, moderate-sensitivity operating profile in the hold-out split (AUC up to 0.89). Discussion: The high-specificity predictive performance observed across machine learning models was accompanied by biologically coherent lipid alterations involving sphingomyelins, lysophospholipids, and fatty acids. These metabolites converge on pathways related to endothelial dysfunction, chronic inflammation, oxidative stress, and vascular homeostasis, indicating that systemic metabolic differences are already present at baseline before the first clinical manifestation of leg ulcers. Conclusion: Baseline plasma lipidomic profiling identified a reproducible molecular signature associated with future first leg ulcer development in sickle cell disease, supporting its evaluation as a prognostic approach for early risk stratification. Validation in larger prospective cohorts is warranted. Financial Support: This research was supported by Financiadora de Estudos e Projetos (FINEP), grant number 01.22.0589.01, by Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), grant number 112969/2025-9, by National Institutes of Health, grant number HHSN2682011, and by funding from the Universidade Municipal de São Caetano do Sul (USCS).
SN  - 2531-1379
SN  - 2531-1387
ER  -