AlveolEye: rapid and precise lung morphometry guided by computer vision

Area of Science:

• Pulmonary Pathology
• Computational Biology
• Histomorphometry

Background:

• Accurate lung tissue evaluation is crucial for understanding development, injury, and drug effects.
• Manual histological measurements (MLI, ASVD) are time-consuming and prone to inter-observer variability.

Purpose of the Study:

• To develop an open-source, semi-automated tool, AlveolEye, for rapid and reproducible lung morphometry.
• To improve the precision and throughput of MLI and ASVD calculations from H&E stained lung sections.

Main Methods:

• Developed AlveolEye, a computer vision-assisted software for semi-automated analysis.
• Validated AlveolEye against manual measurements using standard H&E stained lung tissue.
• Assessed generalizability across species and impact on inter-observer variability.

Main Results:

• AlveolEye-assisted MLI calculations closely matched manual measurements.
• The tool preserved trends observed in injury models and human tissues.
• Significantly reduced variation among analyzers, especially those with less experience.

Conclusions:

• AlveolEye offers a reproducible and efficient method for lung morphometry.
• The semi-automated design allows for investigator control and adaptability.
• Facilitates larger sample sizes and improves statistical power in preclinical lung research.