Hyperspectral imaging for cancer applications

Abstract

Hyperspectral imaging (HSI) is a well-established technique in remote sensing that has been successfully translated into the medical field, especially for diagnosis and guided surgery. HS images are composed of spatial and spectral information, which are merged together into a three-dimensional array. Hence, the spectral information makes it possible to discriminate, at pixel level, the materials present in the captured scene according to their reflectance. Furthermore, one key advantage is that HIS is a contactless, non-ionizing, non-labeled, and non-invasive technology. In the medical area, HSI has been proposed for the detection of various types of tumor tissues in cancer diagnosis. Thus, this chapter presents a review of HSI for detecting and evaluating various types of cancer. First, the main types of HSI technologies and instrumentation in the medical field are described. The analysis algorithms based on spectral unmixing and artificial intelligence are discussed. Finally, the state of the art in applications of HSI for skin, brain, gastrointestinal, head & neck cancer and in histological samples for tumor detection are presented in this work.