Neural Network-Based Calculator for Rat Glomerular Filtration Rate.

Area of Science:

• Renal Physiology
• Animal Models
• Computational Biology

Background:

• Glomerular filtration rate (GFR) is crucial for kidney health, with alterations indicating disease.
• Current GFR measurement via creatinine clearance (ClCr) involves lengthy 24-hour urine collection.
• This process can be stressful for animal models, impacting welfare and experimental outcomes.

Purpose of the Study:

• To develop a non-invasive, accurate method for estimating rat creatinine clearance (ClCr).
• To create a user-friendly tool that reduces experimental burden and enhances animal welfare.
• To provide a computational alternative to traditional timed urine collection for GFR assessment.

Main Methods:

• A neural network (NN) model was trained using historical data of rat plasma creatinine (pCr), body weight, and measured ClCr.
• The NN model was evaluated on training, validation, and test datasets.
• A publicly available interface, ACLARA, was developed for the NN calculator.

Main Results:

• The NN model demonstrated high accuracy in predicting rat ClCr, with correlation coefficients ranging from 0.856 to 0.902 across datasets.
• Prediction errors were low (0.178–0.203 mL/min), indicating reliable performance.
• The NN proved effective in comparing ClCr across experimental groups, often yielding more congruent results than direct measurement.

Conclusions:

• The developed NN calculator provides an accurate and efficient method for estimating rat GFR.
• ACLARA facilitates experimental procedures, reduces animal stress, and aligns with the 3Rs principles.
• This computational approach offers a valuable alternative for GFR assessment in preclinical research.