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.

Nanoparticles (NPs) have found a wide range of applications in clinical therapeutic and diagnostic fields. However, currently most NPs are still in the preclinical evaluation phase with few approved for clinical use. Microfluidic systems can simulate dynamic fluid flows, chemical gradients, partitioning of multi-organs as well as local microenvironment controls, offering an efficient and cost-effective opportunity to fast screen NPs in physiologically relevant conditions. Here, in this review, we are focusing on summarizing key microfluidic platforms promising to mimic in vivo situations and test the performance of fabricated nanoparticles. Firstly, we summarize the key evaluation parameters of NPs which can affect their delivery efficacy, followed by highlighting the importance of microfluidic-based NP evaluation. Next, we will summarize main microfluidic systems effective in evaluating NP haemocompatibility, transport, uptake and toxicity, targeted accumulation and general efficacy respectively, and discuss the future directions for NP evaluation in microfluidic systems. The combination of nanoparticles and microfluidic technologies could greatly facilitate the development of drug delivery strategies and provide novel treatments and diagnostic techniques for clinically challenging diseases.

Exposure to teratogenic chemicals during pregnancy may cause severe birth defects. Due to high inter-species variation of drug responses as well as financial and ethical burdens, despite the widely use of in vivo animal tests, it’s crucial to develop highly predictive human pluripotent stem cell (hPSC)-based in vitro assays to identify potential teratogens. Previously we have shown that the morphological disruption of mesoendoderm patterns formed by geometrically-confined cell differentiation and migration using hPSCs could potentially serve as a sensitive morphological marker in teratogen detection. Here, a micropatterned human pluripotent stem cell test (µP-hPST) assay was developed using 30 pharmaceutical compounds. A simplified morphometric readout was developed to quantify the mesoendoderm pattern changes and a two-step classification rule was generated to identify teratogens. The optimized µP-hPST could classify the 30 compounds with 97% accuracy, 100% specificity and 93% sensitivity. Compared with metabolic biomarker-based hPSC assay by Stemina, the µP-hPST could successfully identify misclassified drugs Bosentan, Diphenylhydantoin and Lovastatin, and show a higher accuracy and sensitivity. This scalable µP-hPST may serve as either an independent assay or a complement assay for existing assays to reduce animal use, accelerate early discovery-phase drug screening and help general chemical screening of human teratogens.