Current liver fibrosis scoring by computer-assisted image analytics is not fully automated as it requires manual preprocessing (segmentation and feature extraction) typically based on domain knowledge in liver pathology. Deep learning-based algorithms can potentially classify these images without the need for preprocessing through learning from a large dataset of images. We investigated the performance of classification models built using a deep learning-based algorithm pre-trained using multiple sources of images to score liver fibrosis and compared them against conventional non-deep learning-based algorithms - artificial neural networks (ANN), multinomial logistic regression (MLR), support vector machines (SVM) and random forests (RF). Automated feature classification and fibrosis scoring were achieved by using a transfer learning-based deep learning network, AlexNet-Convolutional Neural Networks (CNN), with balanced area under receiver operating characteristic (AUROC) values of up to 0.85–0.95 versus ANN (AUROC of up to 0.87–1.00), MLR (AUROC of up to 0.73–1.00), SVM (AUROC of up to 0.69–0.99) and RF (AUROC of up to 0.94–0.99). Results indicate that a deep learning-based algorithm with transfer learning enables the construction of a fully automated and accurate prediction model for scoring liver fibrosis stages that is comparable to other conventional non-deep learning-based algorithms that are not fully automated.

Bile canaliculi expand and contract in response to the amount of secreted bile, and resistance from the surrounding actin bundles. Further expansion due to bile duct blockade leads to the formation of inward blebs, which carry away excess bile to prevent bile build up in the canaliculi.

Hepatocyte spheroids are useful models for mimicking liver phenotypes in vitro because of their three-dimensionality. However, the lack of a biomaterial platform which allows the facile manipulation of spheroid cultures on a large scale severely limits their application in automated high-throughput drug safety testing. In addition, there is not yet a robust way of controlling spheroid size, homogeneity and integrity during extended culture. This work addresses these bottlenecks to the automation of hepatocyte spheroid culture by tethering 3D hepatocyte spheroids directly onto surface-modified polystyrene (PS) multi-well plates. However, polystyrene surfaces are inert toward functionalization, and this makes the uniform conjugation of bioactive ligands very challenging. Surface modification of polystyrene well plates is achieved herein using a three-step sequence, resulting in a homogeneous distribution of bioactive RGD and galactose ligands required for spheroid tethering and formation. Importantly, treatment of polystyrene tethered spheroids with vehicle and paradigm hepatotoxicant (chlorpromazine) treatment using an automated liquid handling platform shows low signal deviation, intact 3D spheroidal morphology and Z’ values above 0.5, and hence confirming their amenability to high-throughput automation. Functional analyses performance (i.e. urea and albumin production, cytochrome P450 activity and induction studies) of the polystyrene tethered spheroids reveal significant improvements over hepatocytes cultured as collagen monolayers. This is the first demonstration of automated hepatotoxicant treatment on functional 3D hepatocyte spheroids tethered directly on polystyrene multi-well plates, and will serve as an important advancement in the application of 3D tethered spheroid models to high throughput drug screening.

No abstract

A wide range of liver diseases manifest as biliary obstruction, or cholestasis.However, the sequence of molecular events triggered as part of the early hepatocellular homeostatic response to abnormal elevations in biliary pressure remains poorly elucidated. Bile canaliculi are dynamic luminal structures that undergo actomyosin-mediated periodic contractions to propel secreted bile.Additionally, pericanalicular actin is accumulated during obstructive cholestasis. Therefore, we hypothesize that the pericanalicular actin cortex undergoes significant remodeling as a regulatory response against increased biliary pressure. Here, we report that, actomyosin contractility induces transient deformations along the canalicular membrane, a process we have termed inward blebbing. We show that these membrane intrusions are initiated by local ruptures in the pericanalicular actin cortex, and they typically retract following repair by actin polymerization and actomyosin contraction. However, above a certain osmotic pressure threshold, these inward blebs pinch away from the canalicular membrane into the hepatocyte cytoplasm as large vesicles (2-8 m). Importantly, we show that these vesicles aid in the regurgitation of bile from the canalicular system. Conclusion: Actomyosin contractility induces the formation of bile-regurgitative vesicles, thus serving as an early homeostatic mechanism against increased biliary pressure during cholestasis.

In vitro tumor models have played vital roles in enhancing the understanding of the cellular and molecular composition of tumors, as well as their biochemical and biophysical characteristics. Advances in technology have enabled the evolution of tumor models from two-dimensional cell cultures to three-dimensional printed tumor models with increased levels of complexity and diverse output parameters. With the increase in complexity, the new generation of models is able to replicate the architecture and heterogeneity of the tumor microenvironment more realistically than their predecessors. In recent years, artificial intelligence (AI) has been used extensively in healthcare and research, and AI-based tools have also been applied to the precise development of tumor models. The incorporation of AI facilitates the use of high-throughput systems for real-time monitoring of tumorigenesis and biophysical tumor properties, raising the possibility of using AI alongside tumor modeling for personalized medicine. Here, the integration of AI tools within tumor modeling is reviewed, including microfluidic devices and cancer-on-chip models.

Gallbladder carcinoma (GBC) is a vicious and invasive disease. The major challenge in the clinical treatment of GBC is the lack of a suitable prognosis method. Chemokine receptors such as CXCR3, CXCR4 and CXCR7 play vital roles in the process of tumour progression and metastasis. Their expression levels and distribution are proven to be indicative of the progression of GBC, but are hard to be decoded by conventional pathological methods, and therefore, not commonly used in the prognosis of GBC. In this study, we developed a computer‐aided image analysis method, which we used to quantitatively measure the expression levels of CXCR3, CXCR4 and CXCR7 in the nuclei and cytoplasm of glandular and interstitial cells from a cohort of 55 GBC patients. We found that CXCR3, CXCR4 and CXCR7 expressions are associated with the clinicopathological variables of GBC. Cytoplasmic CXCR3, nuclear CXCR7 and cytoplasmic CXCR7 were significant predictive factors of histology invasion, whereas cytoplasmic CXCR4 and nuclear CXCR4 were significantly correlated with T and N stage and were associated with the overall survival and disease‐free survival. These results suggest that the quantification and localisation of CXCR3, CXCR4 and CXCR7 expressions in different cell types should be considered using computer‐aided assessment to improve the accuracy of prognosis in GBC.

Growing evidence supports the critical role of tumour microenvironment (TME) in tumour progression, metastases, and treatment response. However, the in-situ interplay among various TME components, particularly between immune and tumour cells, are largely unknown, hindering our understanding of how tumour progresses and responds to treatment. While mainstream single-cell omics techniques allow deep, single-cell phenotyping, they lack crucial spatial information for in-situ cell-cell interaction analysis. On the other hand, tissue-based approaches such as hematoxylin and eosin and chromogenic immunohistochemistry staining can preserve the spatial information of TME components but are limited by their low-content staining. High-content spatial profiling technologies, termed spatial omics, have greatly advanced in the past decades to overcome these limitations. These technologies continue to emerge to include more molecular features (RNAs and/or proteins) and to enhance spatial resolution, opening new opportunities for discovering novel biological knowledge, biomarkers, and therapeutic targets. These advancements also spur the need for novel computational methods to mine useful TME insights from the increasing data complexity confounded by high molecular features and spatial resolution. In this review, we present state-of-the-art spatial omics technologies, their applications, major strengths, and limitations as well as the role of artificial intelligence (AI) in TME studies.