Automatic Liver Viability Scoring with Deep Learning and Hyperspectral Imaging.

Eric Felli,Didier Mutter,Michele Diana,Jacques Marescaux,Bernard Geny,Mahdi Al-Taher,Andrea Baiocchini,Giuseppe Maria Ettorre,Catherine Schuster,Sylvain Gioux,Richard Nkusi,Alexandre Hostettler,Emanuele Felli,Veronique Lindner,Jordi Gracia-Sancho,Toby Collins,Cindy Vincent,Manuel Barberio

doi:10.3390/diagnostics11091527

Abstract

Hyperspectral imaging (HSI) is a non-invasive imaging modality already applied to evaluate hepatic oxygenation and to discriminate different models of hepatic ischemia. Nevertheless, the ability of HSI to detect and predict the reperfusion damage intraoperatively was not yet assessed. Hypoxia caused by hepatic artery occlusion (HAO) in the liver brings about dreadful vascular complications known as ischemia-reperfusion injury (IRI). Here, we show the evaluation of liver viability in an HAO model with an artificial intelligence-based analysis of HSI. We have combined the potential of HSI to extract quantitative optical tissue properties with a deep learning-based model using convolutional neural networks. The artificial intelligence (AI) score of liver viability showed a significant correlation with capillary lactate from the liver surface (r = −0.78, p = 0.0320) and Suzuki’s score (r = −0.96, p = 0.0012). CD31 immunostaining confirmed the microvascular damage accordingly with the AI score. Our results ultimately show the potential of an HSI-AI-based analysis to predict liver viability, thereby prompting for intraoperative tool development to explore its application in a clinical setting.

Highlights

During the last decade, hyperspectral imaging (HSI) has gained importance in the biomedical field [1,2,3]
The aim of the study was to predict liver viability through the analysis of hyperspectral images using artificial intelligence based on convolutional neural networks (CNNs) to (i) discriminate the liver from the rest of the tissues, (ii) recognize perfused from the non-perfused liver, (iii) predict the level of liver perfusion during the reperfusion phase, and (iv) predict biological data (Figure 1a–c)
After 90 min of ischemia, the parenchyma was characterized by weaker staining in some of the hepatocytes’ cytoplasm, which appeared pale with a reduced volume and hyperchromatic nuclei

Summary

Introduction

Hyperspectral imaging (HSI) has gained importance in the biomedical field [1,2,3]. HSI systems aim to build images based on the computational analysis of light-tissue interactions through the detection of relative reflectance. This allows for the quantification of organic compounds such as oxygenated and deoxygenated hemoglobin at different depths in a wide large field of view [4]. Hepatic artery occlusion (HAO) is a dreadful vascular event, which can occur in different clinical scenarios such as hepatic artery thrombosis, hepatic artery ligation during liver surgery, emboli, arterial abnormalities, eclampsia, and sickle cell crisis. Oxygen deprivation is the starting event that induces parenchymal damage, which is further aggravated once blood circulation has been reestablished by the ischemia-reperfusion injury mechanism (IRI) [12,13]. IRI is associated with a massive injury to hepatocytes even when diagnosed in the early phase

Objectives

Methods

Results

Discussion

Conclusion