[Research of the glomerular filtration rate estimation model]

Area of Science:

• Nephrology
• Biomedical Engineering
• Data Science

Background:

• Glomerular filtration rate (GFR) is a key indicator of kidney function.
• Current GFR measurement using radioisotope markers is complex, costly, and involves radiation exposure.
• Developing simpler, accurate methods for GFR estimation is a global research priority.

Purpose of the Study:

• To optimize existing glomerular filtration rate (GFR) formulas for Chinese renal function data.
• To develop a novel, more accurate GFR estimation formula using BP neural network technology.

Main Methods:

• Optimization of three existing GFR formulas using multi-function optimization techniques.
• Development of a new GFR formula based on BP neural network technology applied to Chinese renal function data.
• Comparison of the new formula's accuracy using RMSE and P30 metrics against existing formulas.

Main Results:

• Optimized existing GFR formulas, improving their computational accuracy for the Chinese population.
• The newly developed BP neural network-based GFR formula demonstrated higher precision than previously existing formulas.
• The new formula achieved superior RMSE and P30 values, indicating enhanced accuracy and reliability.

Conclusions:

• BP neural network technology provides a simpler and more precise method for estimating GFR.
• The developed GFR formula offers a significant improvement in accuracy for assessing renal function in the Chinese population.
• This approach holds promise for more accessible and reliable kidney function monitoring.