Multi-Wavelength Optoelectronic System with Machine Learning for Online Hemodialysis Monitoring

An ultraviolet (UV) optoelectronic system is proposed to online monitor uremic toxins from the patients during hemodialysis (HD). The system uses three sets of UV lighting emitting diodes (LEDs) to measure the absorbance of creatinine, potassium and β2-microglobulin in patients’ dialysate. The LEDs are selected at the representative wavelengths for the uremic toxins using machine learning methods. Collective lenses are added to increase the light relative intensity, reduce stray light, and enhance the absorption measurement. Equipped with Internet of Things (IoT) technology, the system can remotely online assess the HD treatment progress and prevent dialysis complications. It was tested with 18 sessions of clinical trials to show its potential in online HD monitoring capability.