AI-powered rapid detection of multidrug-resistant Klebsiella pneumoniae with informative peaks of MALDI-TOF MS

Sep 30, 2026

Klebsiella pneumoniae is a highly virulent pathogen driving rising antibiotic resistance, while MALDI-TOF MS has enabled rapid microbial identification yet remains insufficiently explored for resistance prediction in large clinical cohorts. This study builds machine-learning models with feature-level interpretability from MALDI-TOF MS spectra to rapidly predict ciprofloxacin (CIP), cefuroxime (CXM), and ceftriaxone (CRO) resistance. Across 28,000+ isolates from two hospitals, top models achieve independent test accuracy 0.7858, sensitivity 0.7289, and specificity 0.8127, supported by enriched resistance-associated m/z peaks and time-stable performance with inter-hospital variation.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages12

Reading Time30 min

AccessFree to download

Abstract

Motivation

Results

Availability and implementation

Introduction

Data and text mining AI-powered rapid detection of multidrug-resistant Klebsiella pneumoniae with informative peaks of MALDI-TOF MS Jang-Jih Lu 1, Chia-Ru Chung2, 􀀃 , Hsin-Yao Wang3 , Yun Tang4, Ming-Chien Chiang2, Li-Ching Wu5 , Justin Bo-Kai Hsu6 , Tzong-Yi Lee4,7 , Jorng-Tzong Horng2, 􀀃 1 Division of Clinical Pathology, Taipei Tzu Chi Hospital, Buddhist Tzu Chi Medical Foundation, New Taipei City, 231016, Taiwan 2 Department of Computer Science and Information Engineering, National Central University, Taoyuan, 320317, Taiwan 3 Bruker, Billerica, MA, 01821, United States 4 Institute of Bioinformatics and Systems Biology, National Yang Ming Chiao Tung University, Hsinchu, 300193, Taiwan 5 Department of Biomedical Sciences and Engineering, National Central University, Taoyuan, 320317, Taiwan 6 Department of Computer Science and Engineering, Yuan Ze University, Taoyuan, 320315, Taiwan 7Center for Intelligent Drug Systems and Smart Biodevices (IDS2B), National Yang Ming Chiao Tung University, Hsinchu City, 300193, Taiwan 􀀃 Corresponding author. Department of Computer Science and Information Engineering, National Central University, No. 300, Zhongda Rd., Zhongli District, Taoyuan City, 320317, Taiwan. E-mail: [horng@db.csie.ncu.edu.tw](horng@db.csie.ncu.edu.tw); Department of Computer Science and Information Engineering, National Central University, No. 300, Zhongda Rd., Zhongli District, Taoyuan City, 320317, Taiwan. E-mail: [chungchiaru@ncu.edu.tw](chungchiaru@ncu.edu.tw) Associate Editor: Thomas Lengauer Abstract Motivation: Klebsiella pneumoniae is a highly virulent superbug with rising antibiotic resistance worldwide. While matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) has transformed microbial identification, its application to antimicrobial resistance prediction remains underexplored, particularly for large clinical cohorts. In this study, we developed machine-learning models with feature-level interpretability using MALDI-TOF MS data to rapidly predict resistance to ciprofloxacin (CIP), cefuroxime (CXM), and ceftriaxone (CRO) in K. pneumoniae. Results: Using more than 28 000 isolates from two hospitals, the best-performing models reached an independent test accuracy of 0 .7858, with sensitivity of 0.7289 and specificity of 0.8127. Several resistance-associated m/z signals—including 3657, 4341, 4519, 4709, 5070, 5409, 5921, 5939, and 6516—were consistently enriched in resistant isolates, offering interpretable spectral markers linked to resistance. Performance remained stable in timebased validation but declined across hospitals, suggesting sensitivity to geographic variability in resistance profiles. Overall, this study demonstrates that combining MALDI-TOF MS with machine learning enables rapid and interpretable prediction of resistance to commonly used fluoroquinolone and cephalosporins in K. pneumoniae. These findings highlight the clinical potential of such models for supporting empiric therapy and emphasize the importance of incorporating local data or adaptive strategies to improve generalizability across healthcare settings. Availability and implementation: Data available on request from the authors. 1 Introduction Klebsiella pneumoniae, a Gram-negative bacterium, can cause a variety of serious infections (Ko et al. 2002) . These infections are often treated with antibiotics, but overusing and misusing these drugs can lead to the development of antibiotic-resistant K. pneumoniae (Gasink et al. 2009, Kim et al. 2016) . This study focused on three antibiotics commonly used to treat K. pneumoniae infections: ciprofloxacin (CIP), cefuroxime (CRO), and ceftriaxone (CXM) . These drugs work by binding to different proteins in the bacterial cell, inhibiting their growth and replication. Detailed information is provided in Table 1, available as supplementary data at Bioinformatics Advances online (Rawat and Nair 2010, Kim et al. 2019) . However, the emerg

FAQs

What clinical problem does the study address?

The study targets rising antibiotic resistance in Klebsiella pneumoniae and the need for faster, more reliable antimicrobial susceptibility testing than methods that can take up to 24 hours.

How does the proposed approach make predictions?

It uses machine-learning models built from MALDI-TOF MS data with feature-level interpretability to predict resistance to ciprofloxacin, cefuroxime, and ceftriaxone.

Which spectral features support resistance prediction?

The models identify multiple consistently enriched resistance-associated m/z signals (including 3657, 4341, 4519, 4709, 5070, 5409, 5921, 5939, and 6516) that function as interpretable spectral markers.

Sep 30, 2026

Uploaded by Alex Sinclair