Longitudinal Lesion Analysis

This workflow takes two T2-FLAIR images from different timepoints, extracts a mask containing MS lesions to each of them, and then it aligns the two images, computes lesion matching between the two timepoints and it generates a report analysing the progression of the lesions.

The workflow includes

  1. DICOM files to NIfTI conversion and reorients to standard. The file filter sends the dicom or nifti files required for the workflow to run and the box converts them.
  2. Apply a lesion prediction algorithm to create a lesion segmentation mask of both input images .
  3. Registration of the first timepoint to the second timepoint to enable their proper comparison.
  4. Analyzing the progression of the segmented lesions between the two images and collecting the most relevant information in a structured report.

Required inputs:

  • T2 timepoint 1:

    • Can be either a NIfTI or DICOM file (DICOM files will be converted before radiomics extraction).
    • Accepted modalities: T2-FLAIR (must be labeled as flair).
  • T2 timepoint 2:

    • Can be either a NIfTI or DICOM file (DICOM files will be converted before radiomics extraction).
    • Accepted modalities: T2-FLAIR (must be labeled as flair).

Minimum input requirements:

  • For optimal result reliability, an isotropic resolution is highly recommended.
    • Recommended resolution: 1mm isotropic.

Settings

DCM2NII:

  • Preferred DICOM to NIfTI conversion tool (drop-down selection):
    • DCM2niix (Default)
    • Mrtrix
    • DCM2nii
    • diffunpack
    • MRIConvert

      The selected tool will be tried first to convert DICOM to NIfTI. If the conversion fails, the other options will be tried sequentially until a successful conversion.

LST:

  • Binarize threshold (decimal):
    • Default: 0

      This value is used as threshold to compute the lesion mask from the probability map

Output files:

  • DCM2NII (DICOM to NIfTI conversion):

    • T2-FLAIR image in NIfTI format.
    • Report of the conversion for quality control purposes.
  • LST (Lesion segmentation tool):

    • T2-FLAIR image in NIfTI format.
    • Mask with segmented lesions (single-label and multi-label).
    • Probability map of the segmented lesions.
    • csv files with volumetric information about the lesions.
    • Histogram of the volume of the lesions.
    • pdf report summarizing all the results.
  • Longitudinal Registration:

    • T2-FLAIR of timepoint 2 image in NIfTI format.
    • Mask with segmented lesions of timepoint 1 aligned to timepoint 2.
  • Lesion progression analysis:

    • csv file with information about the progression fo the lesions.
    • Report containing an informative table and visual results of the most relevant changes.
    • Html report to be displayed in the platform.

References

 


Create free account now!

Sign Up