Multispectral Imaging for Automated Tissue Identification of Normal Human Surgical Specimens

Abstract

Objective Safe surgery requires the accurate discrimination of tissue intraoperatively. We assess the feasibility of using multispectral imaging and deep learning to enhance surgical vision by automated identification of normal human head and neck tissues. Study Design Construction and feasibility testing of novel multispectral imaging system for surgery. Setting Academic university hospital. Subjects and Methods Multispectral images of fresh-preserved human cadaveric tissues were captured with our adapted digital operating microscope. Eleven tissue types were sampled, each sequentially exposed to 6 lighting conditions. Two convolutional neural network machine learning models were developed to classify tissues based on multispectral and white-light color images (ARRInet-M and ARRInet-W, respectively). Blinded otolaryngology residents were asked to identify tissue specimens from white-light color images, and their performance was compared with that of the ARRInet models. Results A novel multispectral imaging system was developed with minimal adaptation to an existing digital operating microscope. With 81.8% accuracy in tissue identification of full-size images, the multispectral ARRInet-M classifier outperformed the white-light-only ARRInet-W model (45.5%) and surgical residents (69.7%). Challenges with discrimination occurred with parotid vs fat and blood vessels vs nerve. Conclusions A deep learning model using multispectral imaging outperformed a similar model and surgical residents using traditional white-light imaging at the task of classifying normal human head and neck tissue ex vivo. These results suggest that multispectral imaging can enhance surgical vision and augment surgeons’ ability to identify tissues during a procedure.

Deep learning classification of inverted papilloma malignant transformation using 3D convolutional neural networks and magnetic resonance imaging

“…8 In the field of otolaryngology, CNNs have advanced research in automated analysis of imaging, including detection of ostiomeatal complex occlusion on sinus CT scans, 9 thyroid lesions on ultrasound, 10 endolymphatic hydrops on optical coherence tomography, 11 and head and neck tissue types on multispectral imaging. 12 In contrast to expert derived image features, CNNs are trained in an end-to-end manner using ground truth labels to "learn" the set of relevant image features for the computer vision task. This makes CNNs well suited to the task of classifying MRIs with IP and IP-SCC, where the full set of MRI features that are relevant for classification is unknown.…”

“…CNNs are a set of deep learning architectures that have demonstrated state‐of‐the‐art performance in multiple computer vision tasks in neuroradiology and other subareas of medicine 8 . In the field of otolaryngology, CNNs have advanced research in automated analysis of imaging, including detection of ostiomeatal complex occlusion on sinus CT scans, 9 thyroid lesions on ultrasound, 10 endolymphatic hydrops on optical coherence tomography, 11 and head and neck tissue types on multispectral imaging 12 . In contrast to expert derived image features, CNNs are trained in an end‐to‐end manner using ground truth labels to “learn” the set of relevant image features for the computer vision task.…”

Multispectral Image under Tissue Classification Algorithm in Screening of Cervical Cancer

“…erefore, the tissue classification algorithm was used to classify the spectra of different tissues in this study. e tissue classification algorithm is to use the band characteristics or spatial characteristics of the multispectral image as the boundary between normal and diseased, so that the elements on the spectrum are divided into different levels or categories, thereby realizing alternative subjective analysis [11]. e tissue classification algorithm used in this study was mainly to adopt the K-value clustering algorithm to summarize and analyze the physiological and pathological regional features of the cervical cancer site and then apply the corresponding regional features to analyze the cervical tissues of 50 patients with no lesions.…”

Texture synthesis of ecological plant protection image based on convolution neural network

“…In recent years, texture synthesis technology, as an important image-based realistic rendering technology, has always received the attention of researchers and has high practical application value ( Jassim and Harte, 2020 ; Shenson et al., 2021 ). Deep learning refers to interpreting data by simulating the thinking mode of human brain through different machine learning algorithms.…”