Skip to content

Artificial Intelligence-based Differential Diagnosis Model for Pulmonary Infections: A Multicenter Retrospective Study

Artificial Intelligence-based Differential Diagnosis Model for Pulmonary Infections: A Multicenter Retrospective Study

Status
Active, not recruiting
Phases
Unknown
Study type
Observational
Source
ChiCTR
Registry ID
ChiCTR2600119375
Enrollment
Unknown
Registered
2026-02-26
Start date
2026-02-27
Completion date
Unknown
Last updated
2026-03-02

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Mycoplasma pneumoniae pneumonia, psittacosis pneumonia, pulmonary infections, etc.

Interventions

Gold Standard:Patients had definite etiological diagnosis results: Mycoplasma pneumoniae pneumonia (MPP) was confirmed or excluded by gold standards such as Mycoplasma nucleic acid testing (e.g., PCR)
viral pneumonia or fungal pneumonia was diagnosed via viral
Index test:Comprehensive artificial intelligence model based on deep learning

Sponsors

Wuxi People's Hospital
Lead Sponsor

Eligibility

Sex/Gender
All
Age
18 Years to No maximum

Inclusion criteria

Inclusion criteria: 1. Received a thin-slice chest CT scan without receiving targeted anti-mycoplasma treatment before the scan. 2. Have clear etiological diagnosis results (gold standard) : mycoplasma pneumonia: PCR, culture or serological specific antibody test positive. Other pneumonia (viral/bacterial/fungal) : positive result of corresponding nucleic acid/antigen test or culture. 3. Normal control: no evidence of pulmonary infection and no abnormal CT images. 4. Have complete baseline clinical data and high-quality CT imaging data. 5. All patients were over 18 years old;

Exclusion criteria

Exclusion criteria: 1. Poor CT image quality (severe motion or respiratory artifacts). 2. Combined with pulmonary diseases that seriously interfere with imaging features (such as active tuberculosis, lung cancer, diffuse pulmonary fibrosis). 3. The inflammatory lesion is too small (maximum diameter < 5mm) to reduce the interference of partial volume effect. 4. Missing clinical or imaging data.

Design outcomes

Primary

MeasureTime frame
Area under curve, AUC;

Secondary

MeasureTime frame
Sensitivity;Specificity;Dice similarity coefficient, DSC;Bland-Altman agreement;

Countries

China

Contacts

Public ContactXiaoyun Hu; Yan Wu

Wuxi People's Hospital

drxyh@foxmail.com+86 510 85350345

Outcome results

None listed

Source: ChiCTR (via WHO ICTRP) · Data processed: Mar 14, 2026