Multimodal-Imaging-Based Interpretable Deep Learning Framework for Distinguishing Brucella from Tuberculosis Spondylitis: A Dual-Center Study
Diagnostics · 22.11.2025 · 15 (23) · 2963
DOI: 10.3390/diagnostics15232963 ↗Kurz gefasst
Fachartikel in der Zeitschrift Diagnostics, online erschienen am 22.11.2025. Verfasst von Mayidili Nijiati und 10 weiteren Autor:innen. Beteiligte Einrichtung: Department of Radiology, The First Affiliated Hospital of Xinjiang Medical University, Urumqi 830002, China. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 19 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Key R&D Program of China, Science & Technology Program of Kashi Prefecture, Outstanding Youth Project of Xinjiang Uygur Autonomous Region Natural Science Foundation.
Abstract (Originalsprache)
Objectives
Brucella spondylitis (BS) and tuberculosis spondylitis (TS) are two causes of infection that share overlapping clinical and imaging features, complicating diagnoses. Early differentiation is critical, as treatment regimens differ significantly. This study aims to develop a deep learning framework using multimodal computed tomography (CT) and magnetic resonance imaging (MRI) data to accurately distinguish between these two conditions, improving diagnostic accuracy and patient outcomes.
Methods
In this study, imaging data were acquired from two centers using different MRI and CT protocols. Sagittal T1-weighted (T1WI) and T2-weighted imaging (T2WI), fat-suppression sequences (T2WI FSE), and sagittal CT data were collected. Image preprocessing included region of interest (ROI) segmentation, and normalization and augmentation techniques were used. A deep learning model, based on pre-trained GoogleNet architectures, was trained and evaluated against human radiologists using metrics including accuracy, sensitivity, and AUC to assess diagnostic performance.
Results
In this study, the GoogleNet deep learning model outperformed other architectures in classifying TS and BS, achieving AUCs of 95.97%, 91.24%, and 81.25% across training, test, and external validation datasets, respectively. In contrast, ResNet, DenseNet, and EfficientNet models showed lower AUC values. GoogleNet also demonstrated high accuracy (90.77% training, 83.04% test) and 90.91% sensitivity and 61.11% specificity in external validation. When compared to three radiologists, GoogleNet outperformed in diagnostic accuracy and speed, achieving an AUC of 88.01% and processing cases in 0.001 min. These findings highlight the potential of AI to enhance diagnostic performance and efficiency. Lastly, the explanation provided by the Grad-Cam model precisely localized major lesions.
Conclusions
This multimodal-imaging-based deep learning model could well differentiate TS and BS. Deep learning does not need manual feature extraction, selection, or model development, and has great potential in daily clinical practice.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Radiology, The First Affiliated Hospital of Xinjiang Medical University, Urumqi 830002, China.
- Department of Radiology, Xinjiang Medical University Affiliated Fourth Hospital, Urumqi 830002, China.
- Xinjiang Key Laboratory of Artificial Intelligence Assisted Imaging Diagnosis, Kashi 844000, China.
- Department of Deepwise AI Lab, Hangzhou Deepwise & League of PHD Technology Co., Ltd., Hangzhou 830054, China.
- Department of Spinal Surgery, The First People's Hospital of Kashi Prefecture, Kashi 844000, China.
- Department of Pharmacy, The First People's Hospital of Kashi Prefecture, Kashi 844000, China.
Förderung
- National Key R&D Program of China
- Science & Technology Program of Kashi Prefecture
- Outstanding Youth Project of Xinjiang Uygur Autonomous Region Natural Science Foundation
- Natural Science Foundation of Xinjiang Uygur Autonomous Region
- Innovative Team Project of the “Pearl River Scholar & Tianshan Talent” Collaborative Expert Studio
- Kashi Sci-Tech Plan Project
Zitieren
APA
Nijiati, M., Zhang, M., Huang, C., Chou, X., Shen, L., Ma, H., Ren, Z., Maimaiti, M., You, Y., Zou, X., & Wang, Y. (2025). Multimodal-Imaging-Based Interpretable Deep Learning Framework for Distinguishing Brucella from Tuberculosis Spondylitis: A Dual-Center Study. Diagnostics, 15(23), 2963. https://doi.org/10.3390/diagnostics15232963
Vancouver
Nijiati M, Zhang M, Huang C, Chou X, Shen L, Ma H, et al. Multimodal-Imaging-Based Interpretable Deep Learning Framework for Distinguishing Brucella from Tuberculosis Spondylitis: A Dual-Center Study. Diagnostics. 2025;15(23):2963. doi:10.3390/diagnostics15232963
BibTeX
@article{Nijiati2025,
author = {Nijiati, Mayidili and Zhang, Mei and Huang, Chencui and Chou, Xinyue and Shen, Lingyan and Ma, Haiting and Ren, Zhenwei and Maimaiti, Maimaitishawutiaji and You, Yi and Zou, Xiaoguang and Wang, Yunling},
title = {Multimodal-Imaging-Based Interpretable Deep Learning Framework for Distinguishing Brucella from Tuberculosis Spondylitis: A Dual-Center Study},
journal = {Diagnostics},
year = {2025},
volume = {15},
number = {23},
pages = {2963},
doi = {10.3390/diagnostics15232963}
}
EUVIMED