Personalized liver lesion detection in follow-up exams

• Automatic detection of liver lesions in follow-up MR scan(s), for better monitoring of disease progress and treatment response.
• Improved performance by inclusion of patient-specific image data (a fine-tuning training step of a deep learning network based on prior patient data).

Input(s):

• Dynamic contrastenhanced MR sequence of liver.
• Diffusion-weighted MR.
• Both acquired at min. two time points.

Main feature(s):

• Automatic detection of liver lesions in follow-up scan(s).
• Improved performance by inclusion of patient-specific image data.

Output(s):

• Accurate, automatic annotations of lesions in liver MRI.

Stand-alone package.

• Medical imaging researcher.
• Radiologist.

Code is research SW and can be made available upon request.