A smear test aiding with an automated slide reader acquires equivalent sensitivity to MGIT960 culture in pulmonary tuberculosis diagnosis

Smear microscopy remains widely used for tuberculosis (TB) diagnosis due to its rapid turnaround and low cost, despite its relatively low sensitivity. With the advent of molecular tests, there is growing debate on whether the smear test remains relevant in clinical settings. An artificial intelligence (AI)-aided automated smear slide reader system (Zopomed, Zhengzhou, China) was evaluated at Beijing Chest Hospital. Sputum samples from suspected TB patients were used to prepare both Ziehl-Neelsen (ZN) and auramine O-stained slides. These were scanned by the automated reader under predefined field-of-view settings. Manual light-emitting diode fluorescence microscopy (for auramine O slides), MGIT960 culture, and Xpert MTB/RIF assay served as reference standards. Among 352 clinically confirmed TB patients, the detection rates were 18.75% by manual microscopy, 20.74% by automated ZN (1,000 fields), 32.67% by automated auramine O (300 fields), 30.68% by MGIT960 culture, and 42.05% by Xpert. Expanding the field of view significantly enhanced detection rates for both staining methods, with auramine O staining consistently outperforming ZN. The automated auramine O method (300 fields) identified 78.70% of culture-positive and 66.89% of Xpert-positive TB cases. All methods maintained specificities exceeding 99% when tested on 158 patients with non-mycobacterial pulmonary diseases. AI-aided auramine O smear testing represents a sensitive, rapid, and cost-effective approach to TB diagnosis. With the one-time investment in the automated system, this technique achieves sensitivity comparable to MGIT960 culture, offering a practical diagnostic solution for high-throughput clinical laboratories and resource-limited settings. IMPORTANCE: This study demonstrated that an AI-aided automated smear microscopy system, when used with auramine O-stained slides scanning 300 fields, achieves a sensitivity comparable to MGIT960 culture, while offering the advantages of rapid, same-day results and significant cost-effectiveness; this one-time investment in equipment, which is compatible with routine staining protocols, provides a highly practical and scalable solution to enhance tuberculosis diagnosis in both high-throughput laboratories and resource-limited settings.

Authors

Institutions

Publication Details

Journal
Microbiology Spectrum
Published
2026-06-26
DOI
https://doi.org/10.1128/spectrum.03970-25
Primary Topic
Image Processing Techniques and Applications
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

A smear test aiding with an automated slide reader acquires equivalent sensitivity to MGIT960 culture in pulmonary tuberculosis diagnosis

Xia Yu, 申阿东, Hairong Huang, Haoran Wang et al.
Microbiology Spectrum
Image Processing Techniques and Applications
article

A smear test aiding with an automated slide reader acquires equivalent sensitivity to MGIT960 culture in pulmonary tuberculosis diagnosis

Xia Yu, 申阿东, Hairong Huang, Haoran Wang, Chao Zhang, Yuqing Chen
article en

Abstract

Smear microscopy remains widely used for tuberculosis (TB) diagnosis due to its rapid turnaround and low cost, despite its relatively low sensitivity. With the advent of molecular tests, there is growing debate on whether the smear test remains relevant in clinical settings. An artificial intelligence (AI)-aided automated smear slide reader system (Zopomed, Zhengzhou, China) was evaluated at Beijing Chest Hospital. Sputum samples from suspected TB patients were used to prepare both Ziehl-Neelsen (ZN) and auramine O-stained slides. These were scanned by the automated reader under predefined field-of-view settings. Manual light-emitting diode fluorescence microscopy (for auramine O slides), MGIT960 culture, and Xpert MTB/RIF assay served as reference standards. Among 352 clinically confirmed TB patients, the detection rates were 18.75% by manual microscopy, 20.74% by automated ZN (1,000 fields), 32.67% by automated auramine O (300 fields), 30.68% by MGIT960 culture, and 42.05% by Xpert. Expanding the field of view significantly enhanced detection rates for both staining methods, with auramine O staining consistently outperforming ZN. The automated auramine O method (300 fields) identified 78.70% of culture-positive and 66.89% of Xpert-positive TB cases. All methods maintained specificities exceeding 99% when tested on 158 patients with non-mycobacterial pulmonary diseases. AI-aided auramine O smear testing represents a sensitive, rapid, and cost-effective approach to TB diagnosis. With the one-time investment in the automated system, this technique achieves sensitivity comparable to MGIT960 culture, offering a practical diagnostic solution for high-throughput clinical laboratories and resource-limited settings. IMPORTANCE: This study demonstrated that an AI-aided automated smear microscopy system, when used with auramine O-stained slides scanning 300 fields, achieves a sensitivity comparable to MGIT960 culture, while offering the advantages of rapid, same-day results and significant cost-effectiveness; this one-time investment in equipment, which is compatible with routine staining protocols, provides a highly practical and scalable solution to enhance tuberculosis diagnosis in both high-throughput laboratories and resource-limited settings.

Microbiology Spectrum
Kunming Children's Hospital (CN), Beijing Chest Hospital (CN)
Natural Science Foundation of Beijing Municipality, Beijing Nova Program
Openalex Percentile: Top 6%
Image Processing Techniques and Applications
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.