Using Hyperspectral Imaging for the Detection of Malaria

Abstract

The accurate diagnosis of malaria is crucial for ensuring an effective treatment and combating the developing drug resistance of the parasite. The standard diagnostic method, consisting of light microscopy in combination with Giemsa-staining, requires manual examination through a trained clinician. This is time-consuming and prone to human error. While emerging diagnostic innovations can overcome these challenges, they face limitations in the clinical implementation.

This work introduces a novel optical microscope utilising Hyperspectral Imaging in a push-broom mode, capturing images line by line as the sample moves. The system employs a prism-grating-prism configuration to generate spectral data, making it both portable and straightforward to assemble. With the intention of scanning and automatically diagnosing individual red blood cells based on their spectral signals, the requirement for chemical staining and experienced clinicians is removed. This alleviates the high diagnostic demand making the developed microscope ideal for clinical deployment across African countries.

The microscope has been used to analyse malaria-infected blood in the late stage through thin blood smears. After acquiring the spectral signal of the red blood cells, the samples have been Giemsa-stained to validate the cell infection. Selected fields of view have been captured with the developed microscope and compared to equivalent Giemsa-stained oil immersion microscopy images. This demonstrates the microscope's ability to capture the spectral signal of infected red blood cells without the need for staining, delivering results within minutes. A distinct spectral difference has been observed between healthy and malaria-infected cells. In particular, the absorption of the produced hemozoin is detectable, and the presence of a parasite has been observed through increased scattering. Furthermore, the hyperspectral microscope has been brought to Togo for an on-site feasibility study demonstrating its straightforward and applicable composition.

It is noteworthy, however, that this thesis work consists of the construction of the microscope and preliminary results for the malaria detection. To validate the results and its accuracy, an extended measurement campaign is required. When verified, this technology is promising to improve the detection of malaria in African countries. In addition, its applications may extend beyond infectious diseases as it has shown promising results in e.g. histopathology, highlighting its versatility and potential to transform diagnostic methods.

Popular Abstract

Malaria remains one of the most significant infectious diseases, particularly in sub-Saharan Africa, where the vast majority of cases and deaths occur. Despite medical advancements, diagnosing malaria still relies on a time-consuming method: Giemsa staining and manual examination of blood samples under a microscope. An experienced clinician must evaluate thousands of red blood cells before giving a positive diagnosis. Due to the long examination time, patients tend to self-medicate without proper diagnosis resulting in drug resistance and reducing treatment effectiveness.

But what if we could revolutionise malaria detection with cutting-edge technology? The implementation of hyperspectral imaging offers a promising solution. This emerging imaging technique enables material identification based on their unique spectral fingerprint. Just as every human has a unique fingerprint, every material creates its own characteristic signal when interacting with light. By capturing images based on spectral fingerprints rather than chemical staining, this system can automatically scan and identify infected cells, eliminating the need for trained clinicians and significantly speeding up diagnosis.

To achieve high-resolution spectral imaging, the system employs spatial scanning through a method known as push-broom imaging. The camera’s field of view is restricted by a slit, requiring the sample to be scanned line by line to generate a full image. The 2D detector captures one spatial dimension $x$ and the spectral dimension $\lambda$ simultaneously, while the second spatial dimension $y$ is acquired through sample movement. This automated scanning removes the necessity for manual scanning and counting of the red blood cells. A prism-grating-prism combination ensures precise spectral separation, allowing instant acquisition of full spectral information along the scanned spatial line. Since the device is portable and easy to assemble, it could be deployed in African clinics, bringing advanced diagnostic capabilities to regions where they are most needed.

Healthy red blood cells have a well-known spectral fingerprint dominated by oxygenated haemoglobin. When malaria parasites invade these cells, they consume haemoglobin and produce a by-product called hemozoin. This biochemical transformation alters the way light interacts with the infected cells, resulting in a spectral change. By analysing the signature absorption of hemozoin and the parasite’s effect on light scattering, the hyperspectral microscope can rapidly identify infections. These infections are verified through the standard diagnostic procedure of Giemsa-staining, followed by observation under an oil immersion microscope. This ensures reliable ground truth comparisons between stained and hyperspectral images. Therefore, the imaging system enables swift and accurate malaria detection, reducing the diagnosis time to just a few minutes.

Although the technology is still in development, further testing could validate its effectiveness and pave the way for widespread use in malaria diagnostics. Especially detecting early stages of the parasite is of high interest as this is the biggest diagnostic challenge. Beyond infectious diseases, the method has potential applications in histopathology, demonstrating its versatility in revolutionising medical imaging and diagnosis.

This pioneering work has the potential to transform malaria diagnosis, making it faster, more reliable, and accessible to regions where timely identification can save lives. With ongoing refinements, the microscope could soon become a crucial tool in the global fight against malaria.

• author: Müller, Michelle Franziska
• supervisor: Aboma Merdasa; Elias Kristensson
• organization: Combustion Physics; Department of Physics; Faculty of Medicine
• course: FYSM64 20251
• year: 2025
• type: H2 - Master's Degree (Two Years)
• subject: Physics and Astronomy
• keywords: Hyperspectral Imaging, Hyperspectral Microscopy, Malaria
• language: English
• id: 9203764
• date added to LUP: 2025-07-26 09:30:44
• date last changed: 2025-07-26 09:30:44