Related Experiment Video
Updated: May 9, 2026
03:22
Microbiological Rapid On-Site Evaluation for Pulmonary Infectious Diseases
Published on: March 1, 2024
Artificial intelligence-assisted rapid on-site evaluation in liver biopsy: a diagnostic accuracy study
Cheng-Cheng Du1, Yu-Xian Chen1, Chun-Hai Li1
1Department of Radiology, Qilu Hospital of Shandong University, Jinan, China.
Frontiers in Oncology
|May 8, 2026
Summary
Artificial Intelligence-assisted Rapid On-Site Evaluation (AI-ROSE) shows high diagnostic accuracy for liver biopsies, outperforming traditional cytology. Combining AI-ROSE with cytology significantly enhances intraoperative diagnosis reliability.
Area of Science:
- Medical Imaging
- Pathology
- Artificial Intelligence
Background:
- Liver biopsy is the standard for diagnosing liver lesions but is time-consuming.
- Artificial Intelligence-assisted Rapid On-Site Evaluation (AI-ROSE) offers real-time assessment of biopsy samples.
Purpose of the Study:
- To evaluate the diagnostic performance of AI-ROSE in liver biopsies.
- To compare AI-ROSE with exfoliative cytology and histopathology.
Main Methods:
- Prospective enrollment of 58 patients undergoing CT-guided liver biopsy.
- Triple assessment of biopsy samples: AI-ROSE, exfoliative cytology, and histopathology.
- Calculation of diagnostic metrics and agreement analysis using Kappa statistic.
Main Results:
- AI-ROSE demonstrated 92.31% sensitivity and 87.93% accuracy, surpassing exfoliative cytology.
- The parallel application of AI-ROSE and cytology achieved 96.15% sensitivity.
- Combined approach significantly improved agreement with histopathology (Kappa = 0.498).
Conclusions:
- AI-ROSE offers superior sensitivity and accuracy compared to exfoliative cytology for liver biopsy diagnosis.
- The synergistic use of AI-ROSE and cytology enhances intraoperative diagnostic reliability.
- AI-ROSE shows significant potential for optimizing clinical decisions and reducing missed diagnoses.
