English
首页
科学研究
研究课题
RUSH
出版物
论文专著
软件著作权
成果奖励
团队成员
带头人
全部成员
讲座
联系方式
课题研究
首页
>
课题研究
>
通过表型成像的精确癌症诊断
通过表型成像的精确癌症诊断
## Precision cancer medicine by phenotypic imaging ### Abstract Precision medicine is the future for major human diseases such as cancer, which relies on high-throughput screening of therapeutics that matches the patient tumour phenotypic heterogeneity. Microfluidics, organ-on-a-chip and 3D bioprinting are the cluster of techniques that enables the rapid, high-throughput and precision modelling and long-term culturing of patient-derived disease models *in vitro*. Computational imaging and deep learning enables the precision information acquisition of patient conditions in diagnosis and prognosis. ### Highlights  ####Precision microfluids manipulation Droplet-based microfluidics has evolved as an important platform in many aspects of scientific research, due to its ability to isolate targets from their surroundings through w/o interfaces. The flow topology in moving microdroplets has a significant impact on the behaviour of encapsulated objects and hence on applications of the technology. The droplets moving in a rectangular microchannel, studied by means of micro-particle image velocimetry (μPIV), was found to have the well-known flow pattern characterised by a parabola-like profile in the droplet bulk-volume, surrounded by two counter rotating recirculation zones on either side of the droplet axis when the droplet system has lower inner-to-outer viscosity ratios. As the viscosity ratio between the two phases is increased, the flow pattern becomes more uniform, exhibiting low velocities in the droplet bulk-volume and higher-reversed velocities along the w/o interface. Shear rate magnitude values were found to be an order of magnitude lower than those in the channel and hence capable of reducing stresses in flow cytometry to far below reported values for cell damage. Our study highlights the complex, three-dimensional (3D) nature of the flow inside droplets in microchannels and demonstrates the ability to precisely control the droplet flow environment.  ####Precision disease modelling The behavior and function of cells are critically influenced by the mechanical and chemical properties of the microenvironment. In the emerging field of microscale tissue engineering materials are required that direct cell and tissue function by controlling the microenvironment in 3D. Physiologically relevant 3D microtumours have been explored as promising tools for preclinical drug evaluation for personalised cancer therapeutics, or to accelerate new drug development. Large scale 3D ordering of anisotropic gel objects, such as gel microrods, both rigid and soft, is in demand for the engineering of replica tissues but has not yet been achieved. Gel microrods will be useful for the production of printed tissues, which mimic intricate architectures found in Nature that cannot presently be attained. Droplet-based microfluidics offers an efficient approach to access monodisperse microtumours with tailorable compositions (e.g. stroma enclosing the hypoxia core, or co-culturing of a microtumour and another healthy microorgan), sizes and shapes in high-throughput and highly reproducibility. The integrated microfluidics-3D printing system holds promise to approach patient-derived tumour models for precision cancer therapy.  ####Precision therapeutic information extraction Cells in culture displaying diverse motility behaviours and morphological phenotypes may reflect the cell states and functions, providing the cues to discriminate the cell healthy conditions. Deep learning methods applied to a large corpus of patient tissue and cell phenotypes may provide a meaningful and more data-driven approach to patient categorization of disease (cancer) development and therapy. The application of high-content imaging in conjunction with multivariate clustering techniques enables phenotypic profiling of drug response. The integration of multi-modality data (e.g. phenotypic information across multi-scales and multi-tissue depths) extraction with deep learning based data processing has the potential to drive the development of a new generation of novel therapeutic screening for personalized precision (cancer) medicine.
