Project ID BE-MI2023_17


Co Supervisor 1A School of Biomedical Engineering & Imaging SciencesWebsite

Co Supervisor 1B School of Biomedical Engineering & Imaging SciencesWebsite

Additional Supervisor Dr. Cheng Fang, Dr. Hasti Robbie

Development of Artificial Intelligence (AI) for Automated Detection of Pathological Mediastinal Lymph Nodes 

Lung cancer is the leading cause of cancer related death worldwide. The presence of malignant mediastinal lymph nodes (MLNs) is important in determining disease stage and appropriate treatment. Non-invasive screening methods using CT and/or PET have low sensitivity (Fréchet 2018), requiring biopsy for definitive diagnosis. Artificial intelligence (AI) techniques could potentially detect and classify MLNs (Kawaguchi 2021), reducing unnecessary biopsies and improving patient stratification.

This project aims to: Develop a computer algorithm to automatically detect and classify MLN.

Characterise algorithm performance using retrospective data with histologic ground truth.

The student undertaking this project will develop skills in Python programming and toolboxes including SciPy, PyTorch, and SimpleITK. They will learn image processing, AI/Deep Learning, and statistical validation techniques. The student will work in a collaborative interdisciplinary team, including interventional radiologist and biomedical engineers.

Year 1: In collaboration with Drs. Fang and Robbie curate a dataset of CT chest images containing biopsy-confirmed MLNs. Begin development of algorithm to detect MLNs.

Year 2: Complete detection algorithm. Begin development of classification algorithm, considering features extracted with convolutional layers.

Year 3: Complete the MLNs classification algorithm. Assess algorithm performance using curated dataset to determine precision and recall.

Works Cited

Fréchet, Benoît, et al. “Diagnostic accuracy of mediastinal lymph node staging techniques in the preoperative assessment of nonsmall cell lung cancer patients.” Journal of bronchology & interventional pulmonology 25.1 (2018): 17-24.

Kawaguchi, Yohei, et al. “The predictive power of artificial intelligence on mediastinal lymph node metastasis.” General Thoracic and Cardiovascular Surgery 69.12 (2021): 1545-1552.

Representative Publications

F. Pérez-García, R. Sparks, S. Ourselin. TorchIO: a Python library for efficient loading, preprocessing, augmentation and patch-based sampling of medical images in deep learning. Computer Methods and Programs in Biomedicine, 2021. doi: 10.1016/j.cmpb.2021.106236

A. Granados, Y. Han, O. Lucena, V. Vakharia, R. Rodionov, S. B. Vos, A. Miserocchi, A. W. McEvoy, J. S. Duncan, R. Sparks, S. Ourselin. Patient-specific prediction of SEEG electrode bending for stereotactic neurosurgical planning. International Journal of Computer Assisted Radiology and Surgery, 2021. doi:10.1007/s11548-021-02347-8