A semi-automated technique for adenoma quantification in the Apc^Min mouse using FeatureCounter

Area of Science:

• Oncology
• Bioinformatics
• Medical Imaging

Background:

• Colorectal cancer is a significant global health burden.
• The Apc^Min mouse model is crucial for studying intestinal neoplasia, mimicking human colorectal cancer mutations.
• Manual adenoma counting in Apc^Min mice is labor-intensive and lacks standardization.

Purpose of the Study:

• To develop and validate an automated, image-based method for quantifying intestinal adenomas in Apc^Min mice.
• To improve the efficiency, standardization, and data richness of tumor burden assessment in this model.
• To provide a more reliable alternative to manual adenoma counting.

Main Methods:

• Photographic documentation of Apc^Min mouse small intestines.
• Image processing using an ImageJ macro (FeatureCounter) for automatic feature detection.
• Application of a machine learning pipeline for adenoma identification and quantification.

Main Results:

• The automated method achieved high specificity (80% TNR) and sensitivity (87% TPR) in adenoma detection compared to manual counting.
• Tumor burden measurements derived from automated analysis effectively distinguished between high- and low-burden mice, comparable to manual methods.
• The automated strategy demonstrated increased speed and reduced experimenter bias.

Conclusions:

• The developed automated image analysis and machine learning pipeline offers a faster, more consistent, and informative approach to quantifying intestinal adenomas in Apc^Min mice.
• This method enhances the reliability of results in colorectal cancer research using this important animal model.
• The strategy provides a practical solution for standardized and efficient tumor burden assessment.