Skip to main content

Web Content Display Web Content Display

SOLARIS centre

Web Content Display Web Content Display

Scientific publications list

Web Content Display Web Content Display

Detection of early pancreatic cancer lesions using infrared and machine learning

Detection of early pancreatic cancer lesions using infrared and machine learning

A group of researchers from the CIRI beamline in their latest publication entitled Pancreatic intraepithelial neoplasia detection and duct pathology grading using FT-IR imaging and machine learning published in Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy presented the results of their PanIN classification method, which provides opportunities for early recognition of changes in the cells lining the pancreatic ducts using infrared and machine learning.

Pancreatic intraepithelial neoplasia (PanIN) manifests itself by changes in the cells lining the pancreatic ducts. It is an early pre-cancerous lesion divided into low-grade and high-grade PanIN. In particular, high-grade PanIN is a lesion that often leads to Pancreatic ductal adenocarcinoma (PDAC). In the case of pancreatic cancer, due to the lack of characteristic symptoms of the disease in its early stage, patient survival is low. The basic examination performed to diagnose the disease is to take a fine-needle biopsy from the patient. The most common method of treatment is to remove part of the tissue affected by cancer, which increases the patient's chances of survival, especially if it is done at an early stage of the disease. Therefore, it is so important to understand the biochemistry of lesions such as PanIN and their progression to cancer. In recent years, due to difficulties related to the histopathological differentiation of PanIN into grades I, II, and III, their classification has been changed to low- and high-grade PanIN. This shows how difficult it is to distinguish these changes based on the image of tissue histopathologically stained. In their latest publication, they measured 48 pancreatic tissues (from fine-needle biopsies) using infrared (IR) imaging and then used machine learning methods to detect and grade PanIN. In the analyzed tissues, they detected epithelial neoplasia and also managed to assess the nature of the pancreatic ducts, placing them on a scale from healthy to cancerous ducts, without any staining. Tissues containing ducts affected by PanIN were in the middle between tissues containing cancerous ducts and healthy ducts in terms of biochemical composition. Moreover, since IR imaging allows obtaining biochemical information in the form of spectra but also tissue images, the result of our method is a classified tissue image, allowing for the visual assessment of changes, similar to the case of histopathological staining. In their publication, they were the first to apply this approach to PanIN classification, and the models they provided can be developed by other researchers to further explore epithelial neoplasia.

 

 

Fig. 1. Scheme of sample collection (figure upper part), and FT-IR imaged TMA processing using: Random Forest classification (figure middle part), PLS Regression (figure bottom part).

 

Fig. 1. Scheme of sample collection (figure upper part), and FT-IR imaged TMA processing using: Random Forest classification (figure middle part), PLS Regression (figure bottom part).

 

Author: Danuta Liberda-Matyja

Link to the publication: D. Liberda-Matyja, P. Kozioł, T. Wrobel, Pancreatic intraepithelial neoplasia detection and duct pathology grading using FT-IR imaging and machine learning, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 309, 123756(2023), doi: 10.1016/j.saa.2023.123756

 


 

Recommended
Hercynite – the mysterious refractory spinel compound understood!

Hercynite – the mysterious refractory spinel compound understood!

New method of data quality improvement by noise removal

New method of data quality improvement by noise removal

Direct observation of the spin texture and the Rashba effect of a ferroelectric semiconductor

Direct observation of the spin texture and the Rashba effect of a ferroelectric semiconductor

2D boron layer as a nanoscopic diffraction grating

2D boron layer as a nanoscopic diffraction grating