Multifactorial estimation of clinical outcome in HPV-associated oropharyngeal squamous cell carcinoma via automated image analysis of routine diagnostic H&E slides and neural network modelling.

OBJECTIVE: Routine haematoxylin and eosin (H&E) photomicrographs from human papillomavirus-associated oropharyngeal squamous cell carcinomas (HPV + OpSCC) contain a wealth of prognostic information. In this study, we developed a high content image analysis (HCIA) workflow to quantify features of H&E images from HPV + OpSCC patients to identify prognostic features and predict patient outcomes. METHODS: First, we have developed an open-source HCIA tool for single-cell segmentation and classification of H&E images. Subsequently, we have used our HCIA tool to analyse a set of 889 images from diagnostic H&E slides in a retrospective cohort of HPV + OpSCC patients with favourable (FO, n = 60) or unfavourable (UO, n = 30) outcomes. We have identified and measured 31 prognostic features which were quantified in each sample and used to train a neural network (NN) model to predict patient outcomes. RESULTS: Univariate and multivariate statistical analyses revealed significant differences between FO and UO patients in 31 and 17 variables, respectively (P < 0.05). At the single-image level, the NN model had an overall accuracy of 72.5% and 71.2% in recognising FO and UO patients when applied to test or validation sets, respectively. When considering 10 images per patient, the accuracy of the NN model increased to 86.7% in the test set. CONCLUSION: Our open-source H&E analysis workflow and predictive models confirm previously reported prognostic features and identifies novel factors which predict HPV + OpSCC outcomes with promising accuracy. Our work supports the use of machine learning in digital pathology to exploit clinically relevant features in routine diagnostic pathology without additional biomarkers.

Continuously perfusable, customisable, and matrix-free vasculature on a chip platform.

Creating vascularised cellular environments in vitro is a current challenge in tissue engineering and a bottleneck towards developing functional stem cell-derived microtissues for regenerative medicine and basic investigations. Here we have developed a new workflow to manufacture vasculature on chip (VoC) systems efficiently, quickly, and inexpensively. We have employed 3D printing for fast-prototyping of bespoke VoC and coupled them with a refined organotypic culture system (OVAA) to grow patent capillaries in vitro using tissue-specific endothelial and stromal cells. Furthermore, we have designed and implemented a pocket-size flow driver to establish physiologic perfusive flow throughout our VoC-OVAA with minimal medium use and waste. Using our platform, we have created vascularised microtissues and perfused them at physiologic flow rates for extended time (>2 weeks) observing flow-dependent vascular remodelling. Overall, we present for the first time a scalable and customisable system to grow vascularised and perfusable microtissues, a key initial step to grow mature and functional tissues in vitro. We envision that this technology will empower fast prototyping and validation of increasingly biomimetic in vitro systems, including interconnected multi-tissue systems.

High-content image analysis to study phenotypic heterogeneity in endothelial cell monolayers.

Endothelial cells (ECs) are heterogeneous across and within tissues, reflecting distinct, specialised functions. EC heterogeneity has been proposed to underpin EC plasticity independently from vessel microenvironments. However, heterogeneity driven by contact-dependent or short-range cell-cell crosstalk cannot be evaluated with single cell transcriptomic approaches, as spatial and contextual information is lost. Nonetheless, quantification of EC heterogeneity and understanding of its molecular drivers is key to developing novel therapeutics for cancer, cardiovascular diseases and for revascularisation in regenerative medicine. Here, we developed an EC profiling tool (ECPT) to examine individual cells within intact monolayers. We used ECPT to characterise different phenotypes in arterial, venous and microvascular EC populations. In line with other studies, we measured heterogeneity in terms of cell cycle, proliferation, and junction organisation. ECPT uncovered a previously under-appreciated single-cell heterogeneity in NOTCH activation. We correlated cell proliferation with different NOTCH activation states at the single-cell and population levels. The positional and relational information extracted with our novel approach is key to elucidating the molecular mechanisms underpinning EC heterogeneity.