Application of Artificial Intelligence to Enhance the Fluorescence Microscopy of Circulation Tumor Cells Captured by MEF Chips
Project Description

Cancer is one of the top killers in Hong Kong and the other world. Conventional cancer diagnostics using invasive biopsy and/or positron emission tomography in hospitals is expensive and time-consuming. Circulation tumor cells (CTCs) in human blood has been shown to be one new liquid biopsy for cancer diagnostics. The Microfluidic Elasto-Filtration (MEF) Chips have been used for personalized detection of CTCs in the clinical study in the collaborative hospital, Sun-Yat Sun University Cancer Center in Guangzhou, China. However, the fluorescence microscopy of CTCs captured by MEF chips is labor-intensive and time-consuming. This project is to apply Artificial Intelligence, Deep Convolutional Neural Network (DCNN) algorithm, to automatically learned features for fluorescence microscopy of captured CTC.

Supervisor
LEE Yi-Kuen
Quota
4
Course type
UROP1000
UROP1100
UROP2100
UROP3100
UROP4100
Applicant's Roles

The student(s) will develop a computer program to analyze the fluorescence micrographs obtained from the multi-color fluorescence microscopy of microfluidic elasto-filtration (MEF) CTC chips for cancer diagnosis. The MEF CTC chips will be fabricated at HKUST Nanosystem Fabrication Facility. The program will perform the following tasks: counting of white blood cells (WBC), generate the histogram of WBC size; counting of CTC and plot the histogram of CTC size and then apply the open-source Deep Convolutional Neural Network (DCNN) to process the fluorescence micrographs of CTCs to achieve the new automatic detection of CTC based on FDA-approved immunofluorescence assay and the nuclear-cytoplasmic ratio(NC ratio).

Applicant's Learning Objectives

Develop an understanding of cancer and CTCs

Develop an understanding of multi-color fluorescence microscopy for CTCs

Develop computational skills to design and analyze the digital fluorescence micrographs from MEF chips.

Develop the skill to apply open-source Deep Convolutional Neural Network (DCNN) program to process the fluorescence micrographs of CTCs

Develop and practice technical communication skill

Complexity of the project
Moderate