Automated Screening Tool for Spinal Cancer in MRI

The aim of this project is to develop an automated system capable of detecting malignancy in spinal MRI

Area of work

Spinal cancer

Imaging modality

MRI

Size of data set

5000

image/svg+xml

Project lead

Dr Sarim Ather

Industry partners

Number of NHS partners

8

Unmet need

This project is focused on the automated detection of cancer in clinical MRI spinal scans with a focus on bone metastases and multiple myeloma. Detecting such diseases early is vital for informing decision-making in patient treatment; missed diagnoses have serious implications for prognosis and patient quality of life. However, such cases do occur with devastating impacts for patients as well litigation against clinicians and the NHS (Nishikawa et al., 2012). A potential solution to this problem is the use of automated methods for diagnosis and quantitative analysis. For diagnosis, this could act as an important second-opinion for clinicians, reducing the chances of missed disease as well as providing a form of triaging, so that radiologists can direct their attention to scans that most urgently require it. Furthermore, automated methods could provide a quick and easily reproducible method of quantitative analysis to aid clinical decision-making. In this project we aim to develop such a system using clinical whole-body MRI scans.

Project aims

The aim of this project is to develop an automated system capable of detecting malignancy in spinal magnetic resonance imaging (MRI).

This system will employ the latest methods in computer vision and machine learning and is focused particularly on

  • the automatic detection and quantification of bone metastases, where cancer cells migrate from a primary tumour to the bone and
  • myeloma, a form of bone marrow cancer.

Finding these diseases early and effectively can result in a significant difference in patient quality of life and allow for more effective treatment and pain management. An automated system will give clinicians a vital second opinion, reducing chances that important signs of this disease will be missed and allowing radiologists to direct their attention to scans that need it the most.

Automated methods also allow for quick, accurate and reproducible methods of quantifying disease burden, allowing for large-scale studies comparing different forms of treatment in a principled manner.

Outcomes

The aim of this project is to develop an automated system capable of detecting malignancy in spinal magnetic resonance imaging (MRI). This system will employ the latest methods in computer vision and machine learning and is focused on

  • the automatic detection and quantification of bone metastases, where cancer cells migrate from a primary tumour to the bone and
  • myeloma, a form of bone marrow cancer.

No articles found