Image Fusion for Cancer Diagnosis

Published on

A standard approach for examining excised human tissues is to stain sections with haematoxylin and eosin (H&E) and examine them with optical light microscopy. This approach results in the preservation of morphology and spatial relationships of cells and tissues and highlights the nucleic acid and protein content of the tissues at blue and red visible wavelengths, respectively. Expanding the wavelength range of tissue images to the infrared (IR) adds considerable information on the chemistry of the tissue but the longer IR wavelengths significantly reduce the spatial resolution of the image from that obtained at visible wavelengths.

Recently Safaa Al Jedani, a research student in the SciaScan group, has developed a method of fusing optical and infrared spectral images of tissue in a way that combines the spatial resolution of the optical image with the spectral information present in the IR image. This insight into the chemistry of the tissue is particularly useful for histopathologist in the diagnosis of cancer. However, the technique has wider applications.

 

The image on the left is formed from fusing an H&E stained image of a tissue biopsy, showing a tumour, with the corresponding Fourier transform IR image in which the spectral contributions from three molecules, cytokeratin, collagen and DNA are highlighted in red, green and blue respectively. The rectangular section of the image is shown at higher spatial resolution on the right. The top image on the right us the usual H&E stained image. The bottom fused image allows the histopathologist to see the spatial distribution of these three important molecules in unprecedented detail.

 

Reference

"Tissue discrimination in head and neck cancer using image fusion of IR and optical microscopy," Safaa Al Jedani, Caroline I. Smith, James Ingham, Conor A. Whitley, Barnaby G. Ellis, Asterios Triantafyllou, Philip J. Gunning, Peter Gardner, Janet M. Risk, Peter Weightman and Steve D. Barrett. Analyst 148 4189 (2023)

Peter Weightman