Multiview hyperspectral topography of tissue structural and functional characteristics

Abstract

Accurate and in vivo characterization of structural, functional, and molecular characteristics of biological tissue will facilitate quantitative diagnosis, therapeutic guidance, and outcome assessment in many clinical applications, such as wound healing, cancer surgery, and organ transplantation. We introduced and tested a multiview hyperspectral imaging technique for noninvasive topographic imaging of cutaneous wound oxygenation. The technique integrated a multiview module and a hyperspectral module in a single portable unit. Four plane mirrors were cohered to form a multiview reflective mirror set with a rectangular cross section. The mirror set was placed between a hyperspectral camera and the target biological tissue. For a single image acquisition task, a hyperspectral data cube with five views was obtained. The five-view hyperspectral image consisted of a main objective image and four reflective images. Three-dimensional (3-D) topography of the scene was achieved by correlating the matching pixels between the objective image and the reflective images. 3-D mapping of tissue oxygenation was achieved using a hyperspectral oxygenation algorithm. The multiview hyperspectral imaging technique was validated in a wound model, a tissue-simulating blood phantom, and in vivo biological tissue. The experimental results demonstrated the technical feasibility of using multiview hyperspectral imaging for 3-D topography of tissue functional properties.

Optical biopsy of head and neck cancer using hyperspectral imaging and convolutional neural networks

“…Preliminary research demonstrates that HSI has potential for providing diagnostic information for a myriad of diseases, including anemia, hypoxia, cancer detection, skin lesion and ulcer identification, urinary stone analysis, enhanced endoscopy, and many potential others in development. [11][12][13][14][15][16][17][18][19][20][21][22] Supervised machine learning and artificial intelligence algorithms have demonstrated the ability to classify images after being allowed to learn features from training or example images. One such method, convolutional neural networks (CNNs), has demonstrated astounding performance at image classification tasks due to their capacity for robust handling of training sample variance and ability to extract features from large training data sizes.…”

Hyperspectral imaging in wound care: A systematic review

“…Zhang et al 70 reported a multimodal imaging device designed to integrate HSI, laser speckle, and thermographic imaging. Liu et al 77 developed a 3D topography multiview/HSI integrated system. Chang et al 72 proposed a multimodal system (3D wound size measurements, HSI, thermo, and chemical vapour sensing analysis) for pressure injury assessment.…”

“…In most cases (n = 19) the multi‐patient studies were reported 1,2,5,7,54‐68 . However, in some cases (n = 12), the case studies were reported 8,50,69‐78 …”

“…Liu et al 77 used an occlusion test (180 mmHg period of 2 minutes, and a reactive hyperaemia period of 4 minutes) to validate Multiview/HSI integrated set up. Measurements were compared with simultaneous laser Doppler perfusion monitor and TcPO2 electrode measurements of blood perfusion and PO2.…”

New perspectives of hyperspectral imaging for clinical research

“…92–94 So far, HSI provides diagnostic information for anemia, hypoxia, cancer detection, skin lesions and ulcer identification, and urinary stone analysis. 12–14,95–106 Additional applications range from in vivo to ex vivo measurements including image-guided surgery, 12,106–108 endoscopy, 78,109 dermatology, 110 macroscopic investigations of ex-vivo tissue specimens 111–113 and histology. 13,14,114…”