Researchers applied machine learning to clinical and nutritional datasets to predict extrauterine growth restriction (EUGR) in preterm infants during transitional nutrition. The retrospective models flag infants at high risk for growth failure early, offering a potential decision support tool to tailor nutrition plans in neonatal intensive care units. In pharmacology, a separate deep‑learning framework predicted drug‑induced nephrotoxicity from chemical and biological features, aiming to identify renal risk earlier in preclinical pipelines. Both studies demonstrate how AI can shift risk assessment earlier in care and development pathways, though prospective validation will be required before deployment in clinical trials or regulatory submissions.