Loading
Life sciences · Journal article
International Journal of Drug Delivery Technology · October 12, 2026
No summary has been generated for this record yet. What follows is drawn from its source metadata only.
Journal article.
No findings were extractable from the material analysed.
Safety was not reported in the material analysed. Check the source before drawing any conclusion about harm.
The source did not state who this applies to in practice.
Graded across the dimensions that decide whether you should act, each from what the source actually supports. There is no single score, and where a dimension was not assessed it says so.
This record has not been graded across any dimension yet. Treat the label above as provisional and read the source.
What is missing. This record has no bottom line, key findings, reported figures, evidence dimensions. That is a gap in the analysis, not a judgement about the study.
Background.Pediatric malignancies remain a leading cause of disease-related mortality in children, yet prediction of cancer emergence, optimal therapy selection, and long-term survivorship is constrained by fragmented multimodal data and limited transfer of adult oncology models to childhood disease biology.We developed an autonomous foundation model (PedOnco-FM) that jointly learns from clinical, genomic, imaging, histopathology, and laboratory data to support the full pediatric oncology care continuum. Methods.A retrospective cohort of 350 children (0-18 years) was assembled across four tertiary centers, spanning acute lymphoblastic leukemia (ALL), acute myeloid leukemia (AML), neuroblastoma, Wilms tumor, medulloblastoma, osteosarcoma, and Ewing sarcoma.PedOnco-FM employs self-supervised pretraining followed by task-specific fine-tuning for risk stratification, treatment-response prediction, progression forecasting, and survivorship simulation.Performance was benchmarked against Random Forest, XGBoost, and a deep neural network using nested cross-validation with an external held-out test set.Results.PedOnco-FM achieved an accuracy of 0.931, F1-score of 0.921, and an area under the receiver operating characteristic (ROC) curve (AUC) of 0.961 for outcome prediction, exceeding XGBoost (AUC 0.908) and the deep neural network (AUC 0.926).Predicted risk strata separated five-year survival cleanly (low risk 0.93, intermediate 0.74, high 0.51).MYCN amplification, TP53 mutation, and minimal residual disease (MRD) were the dominant predictors.Survivorship simulation reproduced observed late-event hazards with a concordance index of 0.84.Conclusions.An autonomous foundation model integrating multimodal pediatric data delivers accurate, interpretable, and clinically actionable predictions across the disease lifecycle, supporting earlier risk-adapted intervention and individualized survivorship planning.Prospective multi-center validation is the necessary next step toward clinical deployment.