Skip to content

A Study on the Diagnostic Accuracy of Differentiating Benign and Malignant Peripheral Pulmonary Diseases and Malignant Subtypes Based on Ultrasound Dual-Modality Technology Combined with Machine Learning: A Multicenter Study

A Study on the Diagnostic Accuracy of Differentiating Benign and Malignant Peripheral Pulmonary Diseases and Malignant Subtypes Based on Ultrasound Dual-Modality Technology Combined with Machine Learning: A Multicenter Study

Status
Recruiting
Phases
Unknown
Study type
Observational
Source
ChiCTR
Registry ID
ChiCTR2500110502
Enrollment
Unknown
Registered
2025-10-14
Start date
2025-10-15
Completion date
Unknown
Last updated
2025-10-20

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

Conditions

1. Malignant lesions: Main types: Peripheral lung cancer, among which the most common subtypes are lung adenocarcinoma and lung squamous cell carcinoma. One of the goals of this study is to distinguish between these two subtypes. 2. Benign lesions: Infectious/inflammatory lesions: Such as pulmonary tuberculosis ball, inflammatory pseudotumor, organizing pneumonia, etc. Benign tumors: Such as pulm

Interventions

Gold Standard:The establishment of the gold standard relies on lung lesion tissues obtained from patients through surgical resection or needle biopsy. Diagnoses are independently evaluated in accordan
Index test:1.Data collection method: The use of dual-modal ultrasound imaging, i.e., simultaneous acquisition of conventional grayscale ultrasound (B-mode) images and dynamic sequences of contrast-enh

Sponsors

Peking University Shenzhen Hospital
Lead Sponsor

Eligibility

Sex/Gender
All
Age
18 Years to No maximum

Inclusion criteria

Inclusion criteria: 1. Aged 18 years or older; 2. Having ultrasonic images that can clearly show lung lesions; 3. Patients with pathological results after surgery or lung tumor puncture; 4. Research subjects who have been informed of and consented to the research protocol of this project, and have provided a written informed consent form.

Exclusion criteria

Exclusion criteria: 1. Patients with pathologically confirmed malignant tumors metastasized to the lungs from other sites; 2. Patients with poor-quality acquired ultrasonic images; 3. Patients with missing clinical data; 4. ROI delineation failure criteria (any of the following): Dice similarity coefficient (DSC) between two physicians' delineations 2.5 mm; proportion of regions with uneven contrast agent perfusion > 20%; lesion boundary visibility < 70% (such cases need to be recorded separately and the proportion reported).

Design outcomes

Primary

MeasureTime frame
Area under the curve;Recall;Weighted F1;Specificity;Sensitivity;Negative predictive value;Precision;Positive Predictive Value;

Secondary

MeasureTime frame
Brier score;Macro-Averaged Recall;

Countries

China

Contacts

Public ContactShi Yu

Peking University Shenzhen Hospital

13823578405@qq.com+86 138 2357 8405

Outcome results

None listed

Source: ChiCTR (via WHO ICTRP) · Data processed: Feb 4, 2026