Skip to content

Lesion Analysis

LesionAnalysisPipeline

__init__

Lesion filling pipeline.

Parameters:

Name Type Description Default
subject object

Subject object

 required
session object

Session object

 required
output_path str

Output path

 required
use_which_t1w str

Which T1w file to use. Defaults to None.

 None
use_which_lesion_mask str

Which lesion mask (T1w space) to use. Defaults to None. Seeks for files in <bids_dir>/derivatives/lesion_mask/sub-<subject_id>/ses-<session_id>

 None
lesion_fill bool

Whether to perform lesion filling. Defaults to False. Need to be True if out_contra_mask is True.

 False
lesion_fill_method str

Lesion filling method. Defaults to 'LIT'. can be 'left-right', 'sym_MNI', or 'LIT'.

 'LIT'
out_contra_mask bool

Whether to save contralateral lesion mask when using 'sym_MNI' method. Defaults to False.

 False
lesion_size_analysis bool

Whether to perform lesion size analysis (MUST be one cluster). Defaults to True.

 True
normalize bool

Whether to normalize lesion mask to MNI space. Defaults to False.

 False
extract_from str

Folder name to extract results from. Defaults to None.

 None

A more detailed description:

Lesion Mask

A basic consideration is to use T1w structural MRI for the basic processing of lesions. In previous studies, we have tried using DWI for lesion delineation (as acute ischemic stroke lesions are more apparent on DWI), but subsequent neuroimaging analysis often involves image registration, and using low-resolution (and distorted) DWI for registration can introduce significant inaccuracies. Therefore, we currently prefer to manually delineate lesions in T1w space for subsequent analysis.

  • if output contralateral lesions is desired (out_contra_mask: true): left-right flipped in 'MNI152NLin2009aSym' space and transformed to native space.

  • if normalization is desired (normalize: true): the lesion mask is normalized to the standard 'MNI152NLin6ASym' space.

Lesion Filling

We currently use the Lesion Inpainting Tool (LIT) developed by the Deep-MI team for lesion filling. This functionality can be enabled by setting lesion_fill: true and lesion_fill_method: 'LIT' (requires downloading the docker image: deepmi/lit:0.5.0).

Considerations for other lesion filling methods:

  • KUL_VBG: This method involves ANTs-based registration. It requires a longer time for processing compared to LIT.
  • FSL Lesion Filling: Taking this as a representative example, we found that many other lesion filling methods are based on WMH lesions in MS patients and are not suitable for the brain infarction lesions studied in our research.

Lesion Size

Three metrics are commonly used to quantify lesion size (only applicable for single-cluster lesion):

  1. Lesion Volume: The total volume of the lesion in cubic millimeters (mm³).

  2. Max Diameter (Axial): The maximum diameter of the lesion in the axial plane, measured in millimeters (mm).

  3. Max Diameter (3D): The maximum diameter of the lesion in three-dimensional space, measured in millimeters (mm).