Segmenting lymph nodes from thorax CT

Segmenting and charcterizing lymph nodes from thorax CT scans using machine learning

Funding:

Norwegian National Advisory Unit for Ultrasound and image-guided therapy (www.usigt.org)

Publications:

https://link.springer.com/article/10.1007/s11548-019-01948-8

https://arxiv.org/abs/2102.06515

Samarbeidspartnere

SINTEF

NTNU

Formål

Accurate lung cancer diagnosis is crucial to select the best course of action for treating the patient. From a thorax CT volume, it is necessary to identify whether the cancer has spread to nearby lymph nodes or not. It is equally important to know precisely where each malignant lymph node is with respect to the surrounding anatomical structures and the airways. In this project, we introduce a new data-set containing annotations of fifteen different anatomical structures in the mediastinal area, including lymph nodes of varying sizes. We have developed a 2D pipeline for semantic segmentation and instance detection of anatomical structures and potentially malignant lymph nodes in the mediastinal area.

Effekt/kvalitetsmål

Improved success rate, improved quality of life. Better leveraging of the 3D information and station mapping for the detected lymph nodes are the next steps.

Problemstilling

See Formål/Purpose.

Ekstern lenke til prosjektet

https://www.sintef.no/en/projects/2011/usigt/