An Intelligent Monitoring System Based on Near-Infrared Spectroscopy for Renal Function Evaluation

Severe sepsis can lead to renal dysfunction, which increases patient mortality. Assessing kidney function promptly and effectively is very important, in particular, in the intensive care unit (ICU). Several diagnostic methods have been developed to analyze renal state, including blood tests, urine tests, magnetic resonance angiography (MRA), ultrasonic testing (UT), and computed tomography angiography (CTA). However, the above approaches require longer testing time, experienced operators, and expensive costs, even the injection of special contrast media. Recently, near-infrared spectroscopy (NIRS) technology is utilized for continuously and non-invasively monitoring changes in tissue composition in many bio-applications. Since renal dysfunction may lead to peripheral vascular disease and limb edema, it may have the potential to indirectly assess renal status by detecting the changes in tissue hemoglobin and water content. Based on NIRS technology, this study proposed a smart system for renal function evaluation that indirectly assesses the renal status of patients with severe sepsis admitted to ICU. Here, several indexes, defined from the changes in tissue oxyhemoglobin, deoxyhemoglobin, and water content, were used to investigate the relationship between them and kidney status. Finally, a neural network technique was also applied in this system for effective classification of different kidney statuses. The experimental results indicated that kidney status was significantly reflected in some defined indexes and could be effectively evaluated by the designed system.