City University of Hong Kong

Haojun Hua

Wei Li

Bee Luan Khoo

The present disclosure provides a platform and method for predicting disease prognosis based on label-free liquid biopsy, in particular, to establish a disease model in vitro on a microfluidic device and provide a trained deep learning neural network for automatically extracting morphological features from cell clusters formed by the label-free liquid biopsy of a test subject in order to predict different disease stages and evaluate efficacies of a corresponding treatment regime to the test subject. The present disclosure also provides a method for training the neural network based on different cell cluster datasets obtained from the disease model corresponding to different deep learning models of the neural network.

Growing community of inventors

Hong Kong, China

Haojun Hua