Deep learning-based classification of the mouse estrous cycle stages.

Area of Science:

• Veterinary Medicine
• Machine Learning Applications
• Preclinical Research

Background:

• Accurate determination of estrous cycle stages in female rodents is crucial for preclinical research.
• Current methods relying on vaginal smear cytology are time-consuming, require extensive training, and can yield inconsistent results.

Purpose of the Study:

• To develop and validate a machine learning model for automated estrous cycle stage determination.
• To improve the efficiency and consistency of estrous staging in rodent models.

Main Methods:

• A machine learning model named SECREIT was trained using 2,096 microscopic images of vaginal cytology.
• The model was tested on a dataset of 736 images, evaluating its performance for each estrous stage.
• Classification accuracy was compared against human examiners using a separate set of 100 images.

Main Results:

• SECREIT achieved an area under the receiver-operating-characteristic curve of 0.962 or higher for each estrous stage.
• In a test with 100 images, SECREIT demonstrated classification accuracy of 91%, comparable to human examiners (91% and 79%).
• The model provided classifications in just 11 seconds, significantly faster than manual evaluation.

Conclusions:

• SECREIT offers a rapid, accurate, and consistent method for determining estrous cycle stages in rodents.
• This machine learning tool has the potential to accelerate preclinical research involving female rodents.
• Automated estrous cycle staging can enhance the reliability and efficiency of studies requiring precise hormonal phase determination.