Precise discrimination of mycobacterial pulmonary diseases via multimodal machine learning integrating chest CT and clinical markers
By
Yangyi Jin
Jindun Ding
Jinsheng Ouyang
Zhiye Yao
Liping Wang
Ruisong Xu
Xuewen Jin
July 9, 2026
Clinical Scorecard: Enhanced Differentiation of Mycobacterial Lung Diseases Using Multimodal Machine Learning with Chest CT and Clinical Data
At a Glance
Category Detail
Condition Mycobacterial lung diseases (MTB-LD and NTM-LD) Key Mechanisms Integration of clinical symptoms, hematological biomarkers, and HRCT features Target Population Patients with microbiologically confirmed mycobacterial lung disease Care Setting Clinical diagnostic settings
Key Highlights
Developed a multimodal machine-learning framework for differentiating MTB-LD from NTM-LD Random forest model achieved AUC of 0.92, sensitivity of 0.89, specificity of 0.93 Key predictive contributors included cystic bronchiectasis, tree-in-bud sign, fever, and laboratory biomarkers Multimodal integration outperformed single-modality approaches Study emphasizes the need for external validation before clinical implementation Guideline-Based Recommendations
Diagnosis
Utilize multimodal data for improved differentiation between MTB-LD and NTM-LD Management
Consider treatment paradigms specific to MTB-LD and NTM-LD based on accurate diagnosis Monitoring & Follow-up
Monitor clinical symptoms and imaging features for ongoing assessment of lung disease Risks
Conventional diagnostic methods may lead to misclassification and inappropriate treatment Patient & Prescribing Data
Patients diagnosed with either MTB-LD or NTM-LD
MTB-LD responds to standard anti-tuberculosis therapy; NTM-LD may require prolonged, species-specific regimens
Clinical Best Practices
Incorporate clinical symptoms and laboratory biomarkers in diagnostic processes Use HRCT features to enhance diagnostic accuracy for mycobacterial lung diseases Implement machine learning tools as exploratory decision-support aids Related Resources & Content
