Mouse lung automated segmentation tool for quantifying lung tumors after micro-computed tomography.

Area of Science:

• Preclinical Cancer Research
• Medical Imaging Analysis
• Computational Biology

Background:

• Lung cancer survival rates have stagnated since the 1970s, necessitating improved research tools.
• Genetically engineered mouse models offer high translational relevance for studying human lung cancer mutations.
• Traditional methods for quantifying lung tumors in mice are insufficient for longitudinal studies and therapy response assessment.

Purpose of the Study:

• To develop a fully-automated alternative to manual segmentation for analyzing micro-computed tomography (microCT) images of mouse lung tumors.
• To create the Mouse Lung Automated Segmentation Tool (MLAST) for precise and efficient quantification of tumor burden.
• To validate MLAST's performance against established methods and assess its utility in preclinical drug efficacy trials.

Main Methods:

• Development of the Mouse Lung Automated Segmentation Tool (MLAST) for automated thoracic region identification and lung field segmentation.
• MLAST categorizes lung tissue into soft tissue, intermediate, and lung categories to quantify tumor burden.
• Validation of MLAST against manual scoring, manual segmentation, and histology; application in a Kras/Lkb1 non-small cell lung cancer efficacy study.

Main Results:

• MLAST accurately quantifies tumor burden by measuring decreased lung volume and increased soft/intermediate tissue.
• Validation confirmed MLAST's precision and sensitivity compared to manual methods and histology.
• Application in an efficacy trial demonstrated MLAST's ability to precisely quantify tumor growth inhibition following drug treatment.

Conclusions:

• MLAST provides a fully-automated, efficient, and validated method for analyzing microCT lung cancer data in mice.
• The tool accelerates data analysis, enabling larger study sizes and mid-study therapeutic readouts.
• Automated image analysis tools like MLAST are crucial for high-throughput, quantitative results in preclinical imaging research.