How Portable Raman Spectroscopy Systems Can Lead to Affordable Kidney Disease Diagnostics

Recently, a team of researchers from Portugal examined a new method that could improve accessibility to kidney disease diagnostics. This study, led by Alessandro Fantoni from the Polytechnic University of Lisbon (IPL), the Center of Technology and Systems (UNINOVA-CTS), and the Associated Lab of Intelligent Systems (LASI), showcased how the advancements made in portable Raman spectroscopy could be applied to detect disease biomarkers in urine samples. The study’s findings were published in the journal Sensors, and they demonstrate how portable instrumentation is being used to improve clinical diagnostics and advance this critical field (1).

How are kidney diseases diagnosed, and what are the current challenges?

Kidney disease is normally diagnosed by taking blood or urine tests from the patient (2). Generally, patients are recommended to test for kidney disease when they experience symptoms such as shortness of breath, blood in urine, weight loss or poor appetite, and excessive urination (2).

Although these tests are quite common, kidney disease detection can occasionally run into issues such as complex laboratory tests and unstable biomarkers. These traditional diagnostic approaches involve expensive, time-consuming laboratory procedures, creating barriers to timely detection and treatment (1).

What did the researchers do in their study?

In their study, the researchers looked to address the above challenges by using a portable Raman system that can acquire high-quality spectra from complex liquids such as urine while remaining cost-efficient and suitable for point-of-care applications.

The foundation of their work draws on the “Starter Edition” methodology of the OpenRAMAN project, which allowed the team to optimize the Raman system’s performance through careful calibration (1). This included adjusting the laser’s temperature, evaluating emission spectra at different ranges, and optimizing acquisition parameters using ethanol spectra. System validation was achieved through the analysis of five urine samples, which showed the device’s ability to capture consistent and sensitive spectra capable of detecting subtle variations in urine composition (1).

What were some of the unique features of this study?

One of the unique features of this study was how the researchers utilized artificial intelligence (AI). The team used a neural network to classify Raman spectra, which accelerated this step in their experimental procedure. Using methanol and ethanol solutions as a baseline, the supervised model achieved 99.19% accuracy and 99.21% precision within a short three-minute training time (1). This high performance highlights the potential for automating the classification of urine spectra and ultimately supporting diagnostic decisions (1).

What do these results mean for the industry?

The researchers showed that with further testing, portable instrumentation can become more widespread for disease diagnosis, which in turn will come with financial benefits. Traditional Raman instruments can cost upwards of tens of thousands of euros (1). Portable Raman instruments, on the other hand, could be priced at less than five thousand euros (1). This cost reduction, combined with the system’s simplified design, could make it a feasible option in both clinical and non-clinical settings, including emergency care and remote regions where access to advanced laboratories is limited.

Despite its successes, the prototype does face challenges. High noise levels and the low intensity of characteristic peaks in the acquired spectra remain technical hurdles (1). The study suggests that using a higher-power laser could mitigate these issues by reducing fluorescence interference and enhancing spectral clarity (1). Additionally, while the ethanol optimization showed system stability, the direct impact of these adjustments on urine spectra requires further investigation.

Looking ahead, the researchers plan to expand the system’s testing by analyzing a larger dataset of urine samples from both healthy and diseased individuals. This next step is crucial for identifying specific biomarkers linked to kidney diseases and evaluating the system’s diagnostic accuracy in real-world contexts (1). They also envision incorporating more sophisticated optical elements to further improve sensitivity and signal strength (1).

Although refinement and validation are still necessary, the research team demonstrates that combining portable instrumentation with AI could result in the creation of more affordable, accessible, and non-invasive healthcare technologies.

References

1. Domingos, C.; Fantoni, A.; Fernandes, M.; et al. Low-Cost Raman Spectroscopy Setup Combined with a Machine Learning Model. Sensors 2025, 25 (3), 659. DOI: 10.3390/s25030659
2. National Health Service, Diagnosis – Chronic Kidney Disease. NHS.uk. Available at: https://www.nhs.uk/conditions/kidney-disease/diagnosis/ (accessed 2025-09-23).