A new method for evaluating lung volume: AI-3D reconstruction.

Area of Science:

• Medical Imaging
• Thoracic Surgery
• Artificial Intelligence in Medicine

Background:

• Accurate lung volume measurement is critical for successful lung transplantation.
• Traditional methods for lung volume assessment have limitations.
• AI-powered 3D reconstruction offers a novel approach to volumetric analysis.

Purpose of the Study:

• To evaluate the clinical utility of an AI-3D reconstruction system for lung volume measurement.
• To assess the practical value of AI-derived lung volumes in donor-recipient size matching for lung transplantation.

Main Methods:

• Retrospective analysis of data from 75 subjects undergoing plethysmography and CT scans.
• Comparison of actual total lung volume (aTLV) from plethysmography with predicted total lung volume (pTLV) and AI-3D reconstruction CT lung volume (AI-3DCTVol).
• Correlation and absolute error analysis performed on overall, male, and female groups.

Main Results:

• AI-3D reconstruction showed significant correlation with actual total lung volume (ICC=0.792) across all participants.
• AI-3D reconstruction demonstrated higher accuracy in females (ICC=0.615) compared to males (ICC=0.413).
• Both AI-3D reconstruction and pTLV formulas showed statistical differences compared to plethysmography values (p<0.05).

Conclusions:

• AI-3D reconstruction is a convenient and promising method for assessing lung volume.
• This technology holds significant potential for improving donor-recipient size matching in lung transplantation.
• Further research may refine AI applications in pre-transplant evaluations.