Comparative Computational Model for Rapid and Accurate Identification of Drug-Resistant Tuberculosis Using the G/Xpert XDR Molecular Diagnostic System

Abstract

Drug-resistant tuberculosis (DR-TB) remains a major global public health challenge due to delayed diagnosis, limited sensitivity of conventional methods, and evolving resistance patterns. Rapid molecular diagnostic systems such as the G/Xpert XDR assay have significantly improved detection of resistance to first-line and second-line anti-tuberculosis drugs. However, the variability in diagnostic outputs, population heterogeneity, and multi-drug resistance patterns necessitate advanced computational approaches to optimize diagnostic accuracy and predictive reliability.
This study proposes a comparative computational model integrating structured diagnostic outputs from the G/Xpert XDR system with analytical decision frameworks to enhance rapid identification of drug-resistant tuberculosis. The model is designed to evaluate diagnostic performance, optimize classification accuracy, and support early clinical decision-making. The approach is grounded in multicenter diagnostic evidence and programmatic implementation studies of Xpert MTB/XDR systems (Penn-Nicholson et al., 2022; Dorman et al., 2023).
The findings indicate that computational augmentation of molecular diagnostic outputs can improve sensitivity–specificity balance, reduce diagnostic uncertainty, and support scalable tuberculosis control strategies. According to global surveillance data, DR-TB continues to impose a significant burden on healthcare systems, requiring integrated diagnostic innovations (World Health Organization. Global tuberculosis report 2024. Geneva: WHO; 2024).
The proposed framework highlights the importance of combining molecular diagnostics with computational intelligence to achieve faster, more reliable, and scalable tuberculosis resistance detection systems.