P03 : NephroSense: Developing a CKD Screening Model to Explore Machine-Learning Breath Biomarker Testing
| Students | Zena Alsaadi
Sham Alsaadi |
| School | HCDSB - Bishop P. F. Reding Secondary School - Milton |
| Level | Intermediate 9/10 - Grade 10 |
| Group | Group 13 - Engineering and Computing V |
| Abstract | Current biomarker tests can take weeks before patients receive their lab results. Along with being time-consuming, they are painful and invasive. This project proposes the use of breath analysis methods for detecting VOC and VIC concentrations in patient breath samples, as these gases can serve as biomarkers for varying illnesses. A device for the screening of chronic kidney disease was developed as a proof-of-concept prototype; it integrates machine learning approaches to create a non-invasive device that requires approximately one minute to produce results without compromising accuracy metrics. It corroborates the feasibility of using breathomics in similar devices that focus on single-gas detection; it also proves its viability to adapt for clinical and institutional applications through the implementation of more generalized gas measurement techniques or machine-learning algorithms. |