Recent advances in computing and imaging have allowed doctors to better identify and treat malignancies in patients, specifically, malignancies which develop without symptoms and are poorly identified in their early stages.

Worldwide, gastric cancers accounted for 6.8% of all cancers in 2012, with mortality rates of 8.8%, ranking it as one of the top five most deadly forms of cancer. Due to the possibility of rapid growth, detecting such malignancies early on is crucial because it allows for less aggressive treatments.

To date, the diagnosis of such gastric pathologies falls into two procedures: gastroendoscopy (visual exploration of the stomach under white light) and biopsies for the collection of tissue for analysis. Although biopsies provide the medical field with highly-accurate results, they have proven to be very demanding and time consuming.  If done correctly, biopsies are not only effective, but crucial in identifying such malignancies, but the chance for error is high. Results of biopsies are strongly dependent on the accuracy of samples collected from the damaged tissue.  Because of this, there was a demand to find additional sources to help identify target areas for biopsy collection.

Sergio Ernesto Martinez Herrera, at the Universite of Paris-Saclay and Universite Versailles Saint-Quentin en Yvelines, considered multispectral imaging to help identify tissue to sample, in hopes of improving accuracy and decreasing the time of the overall procedure. Herrera argues that multispectral imaging could be a great method for identifying areas to sample due to chemical and structural changes in the gastric tissue during the development of malignancies and the different ways in which histopathological tissues may reflect light of different wavelengths compared to normal tissue. Multispectral imaging allows the medical professional to analyze absorption, scattering coefficients and blood content, which can provide important information for identifying precancerous lesions. Overall, multispectral imaging can be used to analyze, both spectrally and spatially, the surface of the gastric mucosa.

FluxData’s FD-1665 was one of the technologies used in Herrera’s study of multispectral imaging to improve the detection of pre-cancerous lesions in digestive endoscopies. In this research, specifically, Herrara chose to look at the reflection of gastritis pathology under various conditions. Gastritis is part of the stages of cancer development which results from a chronic infection by H. pylori (which is directly linked to gastric lesions). Herrara studied mice that developed a chronic infection after inoculation of H. pylori. A total of 25 mice were used for this study, 15 of which were infected with H. pylori (pathological group) and 10 that were not (control group). The mice were then sacrificed at different times post inoculation, to analyze changes in tissue during the development of the pathology. The tissues were analyzed using multispectral imaging to identify spectral regions that show significant changes as the disease progresses. The goal is to link these spectral signatures to the diagnosis of gastritis and the development of cancer.

(Image by: Sergio Ernesto Martinez Herrera)

Segmentation of the stomach from mice. a) Original image b) monoband image centered at 520 nm c) result from a sobel filter d) opening on the image e) threshold and inversion f) entropy filter and identification of the centroid of the stomach g) estimated mask of the stomach and h) segmented stomach. 

FluxData’s FD-1665 was used in Herrara’s research to overcome limitations with other imaging systems. Using the FD-1665, he could increase the spectral range of the system by acquiring the NIR wavelengths and the reduction of the acquisition time by collecting a multispectral image in a single shot. It is known that light encodes chemical and structural information from the tissue in a non-invasive way providing quantitative information for objective diagnosis.  NIR is a strong candidate for the differentiation of pathology due to its ability to express biological changes in water content and light scatter due to oxygen saturation.

The results from the mice model study show significant variations in the mucosa measured at different stages of the pathology between the two groups analyzed. The conclusions from such analysis show that three wavelength bands are candidates for the diagnosis of gastritis: between 430 to 450 nm, 460 to 590 nm, and 620 to 680 nm.

This research shows that the introduction of multispectral imaging for malignancy detection may provide a practical and accurate tool that can be used routinely for medical procedures of various types. This technique can be beneficial in the early stages by offering information to improve detection of gastric lesions, or it can be beneficial later in tracking the performance of medical treatments and the tissue’s response to various medications. Additionally, extending multispectral imaging techniques to other diseases which are difficult to detect early can help in the diagnoses and evaluation of these as well. 

To read the full article, click here