AI Science Tracker
EN

Learned correction improves retrospective infliximab concentration estimates in young patients

Cincinnati Children's Hospital Medical Center; University of Cincinnati; Connecticut Children’s Medical Center; Nationwide Children’s Hospital; Children’s Hospital of Wisconsin · 30 Aug 2025 · biology

Discovery

Learned correction improves retrospective infliximab concentration estimates in young patients

Source check: Source checked. Not independent scientific confirmation.

Finding: XGBoost corrected errors from a population-pharmacokinetic Bayesian model using measured infliximab concentrations in young patients with Crohn’s disease. Internal cross-validation reduced concentration prediction error, but some cases worsened and no external or prospective treatment-benefit evaluation was shown. The result concerns measured drug concentrations, without established improvements in dosing safety or disease outcomes.

Technology: XGBoost error correction integrated with population-pharmacokinetic Bayesian estimation

Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.

AI contribution
XGBoost learned residual prediction patterns and adjusted the prior pharmacokinetic estimates; comparison with measured subsequent concentrations supplies the specific clinical validation result. The pre-existing Bayesian model/dashboard alone is not treated as a new discovery.
Test and result
The source analyzed 292 successive-concentration pairs from 93 patients aged 2–22 years. XGBoost learned prediction errors using clinical features and prior drug measurements. Five-fold evaluation reported RMSE 3.78 ± 0.85 micrograms/mL versus 4.8 for Bayesian estimation alone. Absolute error improved modestly on average; the source also reports worsened predictions in a subset.
Limitation / still unknown
Multiple observations per patient were present, but explicit patient-grouped folds were not established in the reviewed Methods. Correlation-based feature selection and algorithm comparisons may affect internal performance; independent final evaluation was absent. Low concentrations below quantification were excluded. This is retrospective evaluation, without demonstrated clinical remission, safety improvement or validated continuous-learning deployment.
Evidence sources and annotation context

Evidence context (descriptions follow the cited sources; test type is a catalog classification):

Evidence sources: https://link.springer.com/article/10.1007/s40262-025-01564-7

Source checked · editorial details

The source page was checked against this entry. Independent confirmation is not assessed by this label.

Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.

Review date: 21 Nov 2026. Review scope: Read actual primary clinical cohort/consecutive-concentration design, Bayesian and XGBoost residual-learning/tuning/evaluation Methods, measured-concentration error Results and limitations. Patient-grouped fold independence and supplementary implementation were not established; dosing procedures or clinical recommendations not represented. Exact original article front DOI/title and fresh PubMed primary relationship metadata checked..

No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.

Editorial assessment: Impact 1/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.

Open in the full catalog for filters, charts and related discoveries. Browse biology discoveries.