Clinical Application of Artificial Intelligence Methods in Diagnostic Imaging of Pulmonary Nodules

Condition(s) studied

Investigational drug(s) / intervention(s)

No intervention: This is a retrospective observational study. No intervention was administered. The AI-assisted diagnostic system was used as an observational assessment tool.

Study summary

This is a retrospective, multicenter observational study evaluating the clinical value of combining high-resolution (1024×1024 matrix) targeted CT scans with an artificial intelligence (AI) deep learning system to distinguish between benign and malignant pulmonary nodules. By analyzing existing medical records and CT images of adult patients (aged 18 and older) with pulmonary nodules (3-30 mm) who have confirmed diagnoses (via postoperative pathology）, this study aims to assess the diagnostic accuracy and efficiency of the AI system. The study plans to enroll approximately 10,000 cases from three hospitals in Shanghai. The primary goal is to evaluate the AI system's diagnostic performance (such as sensitivity, specificity, and Area Under the ROC Curve) and compare it with the performance of radiologists, ultimately exploring whether this combined approach can provide more accurate clinical diagnoses and reduce unnecessary procedures.

Eligibility

Primary outcome measure(s)

• The Area Under the Receiver Operating Characteristic Curve (AUC) for AI Model Diagnosis of Pulmonary Nodules — Baseline to 1 month after obtaining pathological results
  Using pathological results as the gold standard, calculate the area under the operating characteristic curve (AUC) of the AI model for diagnosing pulmonary nodules at the nodule level, and report the 95% confidence interval. If there are multiple nodules in the same patient, adjust according to patient clustering.

Trial sites (1)

More Shanghai East Hospital trials in China

Other trials for the same condition