Machine and deep learning for improving photoacoustic imaging for CAR T-cell cancer therapy

这个由EPSRC iCASE资助的项目建立在萨里大学(UoS)、癌症研究所(ICR)和国家物理实验室(NPL)之间现有的合作基础上，开发光声成像以帮助优化CAR(嵌合抗原受体)T细胞癌症治疗。这是一种新兴的治疗技术，于2018年被批准用于患者，预计将在个性化癌症药物方面取得重大进展。它包括从血液中移除患者的T细胞，对其进行基因改造，然后将细胞重新注入患者体内。该项目将开发利用光声成像来确定CAR-T细胞在体内的生物分布。光声成像是一种重要的生物医学成像方法，它使用超声扫描仪和脉冲激光相结合的方法来生成组织和细胞中分子的光学吸收光谱的高分辨率3D图像。该项目的目标包括开发方法学，包括应用机器和深度学习，使用光谱和时间光声信号的组合来识别背景组织中的CAR T细胞，并量化CAR-T细胞的数量。这些方法将在模拟组织的模体中和肿瘤或组织切片后的体内细胞的光声显微镜下进行验证。如果有时间和兴趣，将有机会探索CAR-T细胞穿透肿瘤时的动态量化。学生将以萨里大学为基地，可能会在ICR和NPL花费一些时间使用商业光声成像仪进行数据采集。

This EPSRC iCase funded project builds on an existing collaboration between the University of Surrey (UoS), the Institute of Cancer Research (ICR) and the National Physical Laboratory (NPL), to develop photoacoustic imaging to assist the optimisation of CAR (chimeric antigen receptor) T-cell cancer therapy. This is an emerging treatment technique, approved for patients in 2018 and expected to enable major advancements in personalised cancer medicine. It involves removal of the patient's T cells from the blood, their genetic modification, and re-infusion of the cells back into the patient. The project will develop the use of photoacoustic imaging for determining the biodistribution of the CAR-T cells in vivo. Photoacoustic imaging is an important biomedical imaging method which employs an ultrasound scanner in combination with a pulsed laser to create high-resolution 3D images of optical absorption spectra of molecules in tissues and cells. The project objectives involve the development of methodologies, including the application of machine and deep learning, to use combinations of spectral and temporal photoacoustic signatures to recognise CAR T cells against background tissue and to quantify CAR-T cell numbers. Methods will be validated against photoacoustic microscopy of cells in tissue-mimicking phantoms and in vivo after tumour or tissue sectioning. Opportunities will be available, given time and interest, to explore dynamic quantification of CAR-T cell numbers as they penetrate tumours. The student will be based at the University of Surrey, and some time may be spent at the ICR for data acquisition using a commercial photoacoustic imager and at the NPL.