Tumor cell identification and classification in esophageal adenocarcinoma specimens by hyperspectral imaging

Abstract

Esophageal cancer is the sixth leading cause of cancer-related death worldwide. Histopathological confirmation is a key step in tumor diagnosis. Therefore, simplification in decision-making by discrimination between malignant and non-malignant cells of histological specimens can be provided by combination of new imaging technology and artificial intelligence (AI). In this work, hyperspectral imaging (HSI) data from 95 patients were used to classify three different histopathological features (squamous epithelium cells, esophageal adenocarcinoma (EAC) cells, and tumor stroma cells), based on a multi-layer perceptron with two hidden layers. We achieved an accuracy of 78% for EAC and stroma cells, and 80% for squamous epithelium. HSI combined with machine learning algorithms is a promising and innovative technique, which allows image acquisition beyond Red–Green–Blue (RGB) images. Further method validation and standardization will be necessary, before automated tumor cell identification algorithms can be used in daily clinical practice.

Hyperspectral imaging for early diagnosis of diseases: A review

“…In yet another study, Nathan et al, 2018 designed a cancer detection system for distinguishing different types of cancers using HSI and combining MLP with an SVM model, which obtained much better results in terms of speed when compared with detection via biopsy. In another study by Maktabi et al, 2022, a MLP with two hidden layers was utilized, and hyperspectral imaging (HSI) data from 95 patients was analysed to stratify three different histopathological features (squamous epithelium cells, oesophageal adenocarcinoma cells, and tumour stroma cells).…”

“…In yet another study,Nathan et al, 2018 designed a cancer detection system for distinguishing different types of cancers using HSI and combining MLP with an SVM model, which obtained much better results in terms of speed when compared with detection via biopsy. In another study byMaktabi et al, 2022, a MLP with two hidden layers was utilized, and hyperspectral imaging (HSI) data from 95 patients was analysed to stratify three different histopathological features (squamous epithelium cells, oesophageal adenocarcinoma cells, and tumour stroma cells).T A B L E 2 (Continued) Z. Liu et al, 2011 Tongue tumour SVM, SR, RVM Further scope of experiment can be to study further the spectrochemical properties of tumour tissues for the purpose of building a more robust and time efficient algorithm Some experiments made use of transfer learning utilizing pre-trained models for detection. Manni et al, 2020 propose a local feature detection using a pre-trained spinal surgery feature detection model of a CNN-based DELF network.…”

Evaluation of Spectrum-Aided Visual Enhancer (SAVE) in Esophageal Cancer Detection Using YOLO Frameworks

“…Fang et al [ 32 ] used semantic segmentation with NBI and HSI to label early-stage EC, utilizing U-Net as the primary neural network and complementing it with ResNet for precise classification and cancer location prediction. Maktabi et al [ 33 ] utilized HSI data and a multilayer perceptron to discern histopathological features in EC, highlighting HSI’s potential with machine learning for advanced tumor diagnosis. Wu et al [ 34 ] proposed a method for early EC identification, combining endoscopy and hyperspectral endoscopic imaging to detect early cancerous lesions, showcasing potential applications in capsule endoscopy and telemedicine.…”

Contribution of hyperspectral imaging in interventional environment: application to orthopedic surgery

“…While already widely used in underwater imaging and plant biology, hyperspectral imaging has also been applied in the medical field, particularly in oncology for tissue differentiation between healthy and cancerous tissue. [3][4][5] In this work, we propose a system based on hyperspectral imaging for the automatic recognition of different biological organs and tissues, showcasing its potential as a valuable tool for guiding surgical procedures. By integrating hyperspectral imaging into the operating room, we hope to contribute to the development of computer-assisted orthopedic surgery that is minimally invasive, precise, and safe, ultimately improving patient outcomes and quality of life.…”