Lesion Segmentation Workflow

This workflow performs automatic segmentation of T2-FLAIR hyperintensity lesions and uses this information to "fill" the white matter lesions in a corresponding T1 image with surrounding healthy WM values. This enables a fast an automatic process to obtain lesions mask in T1 space, offers the possibility for semi-automatic or manual QC and editing of this lesion mask, and produces a lesion-filled T1 that can be used for improved quantitative calculations in morphometry and connectivity workflows, avoiding incorrect tissue segmentation assignment of the T1 lesions due to lower intensity values.
The workflow includes the following steps:

Workflow scheme



Lesion Segmentation Tool output examples

T2 FLAIR input
Lesion probability mask
Lesion segmentation


Semi automatic lesion editing

 

Lesion Filling output example


 

Required inputs:

  • T1: anatomical 3D image to be lesion-filled.
    Isotropic resolution recommended
    Must be labelled as 'T1' modality
  • T2-FLAIR: anatomical 3D image with hyper-intense lesions
    Isotropic resolution recommended
    Must be labelled as 'T2' modality

Minimum input requirements:

  • For optimal result reliability, isotropic resolution is recommended
    • Recommended resolution: 1 mm isotropic.
    • Minimum reliable resolution: 2mm isotropic.

References:

 

Create free account now!

Sign Up