Research Journal of Cell Sciences
AI-Enhanced Multi-Omics Framework for Predicting Antibiotic Resistance in Pediatric Cancer Patients
Abstract
Avani Agarwal and Nineesha Chumbhale
Antibiotic resistance poses a critical challenge for pediatric cancer patients whose immune systems are compro- mised by chemotherapy. This paper presents an AI-driven multi- omics framework designed to tackle this issue. Using Convolu- tional Neural Networks (CNNs), the framework predicts bacterial resistance patterns, while Multi-Task Neural Networks (MTNNs) evaluate patient-specific drug responses. In trials, the CNN model correctly identified resistant bacteria, such as Klebsiella pneu- moniae (with a 92% probability of resistance), and susceptible strains like Escherichia coli (35% probability of resistance). These insights guided the selection of alternative treatments, including colistin and ceftriaxone. This AI-powered approach marks a step forward in personalized antibiotic therapies, aiming to minimize resistance development and improve health outcomes for vulnerable patients.