Transformers for Ischemic Stroke Infarct Core Segmentation from Spatio-temporal CT Perfusion Scans

Lucas de Vries; Bart J. Emmer; Charles B. L. M. Majoie; H Marquering; Efstratios Gavves

Transformers for Ischemic Stroke Infarct Core Segmentation from Spatio-temporal CT Perfusion Scans

Lucas de Vries, Bart J. Emmer, Charles B. L. M. Majoie, H Marquering, Efstratios Gavves

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

Abstract

The infarct core size is a crucial biomarker for treatment selection for ischemic stroke patients. For this purpose, we present a novel approach to perform infarct core segmentation using CT perfusion (CTP) source data, without ordinary deconvolution analysis. We propose the use of transformers to encode sequential CTP scans in spatial attention maps, which we subsequently use for infarct core segmentation. We report new top results on the ISLES 2018 challenge test data set for infarct core segmentation. This work presents a primary benchmark for infarct core segmentation from CTP source data using transformers.

Original language	English
Title of host publication	Medical Imaging with Deep Learning Lübeck, 7 ‑ 9 July 2021
Publication status	Published - Jul 2020

Keywords

: CT Perfusion
Infarct core segmentation
Ischemic stroke
Transformers

Access to Document

https://pure.amc.nl/ws/files/21244750/transformers_for_ischemic_stro.pdf

Cite this

@inproceedings{819452cd07f44ae0a088143f47206d96,

title = "Transformers for Ischemic Stroke Infarct Core Segmentation from Spatio-temporal CT Perfusion Scans",

abstract = "The infarct core size is a crucial biomarker for treatment selection for ischemic stroke patients. For this purpose, we present a novel approach to perform infarct core segmentation using CT perfusion (CTP) source data, without ordinary deconvolution analysis. We propose the use of transformers to encode sequential CTP scans in spatial attention maps, which we subsequently use for infarct core segmentation. We report new top results on the ISLES 2018 challenge test data set for infarct core segmentation. This work presents a primary benchmark for infarct core segmentation from CTP source data using transformers.",

keywords = ": CT Perfusion, Infarct core segmentation, Ischemic stroke, Transformers",

author = "{de Vries}, Lucas and Emmer, {Bart J.} and Majoie, {Charles B. L. M.} and H Marquering and Efstratios Gavves",

year = "2020",

month = jul,

language = "English",

booktitle = "Medical Imaging with Deep Learning L{\"u}beck, 7 ‑ 9 July 2021",

}

TY - GEN

T1 - Transformers for Ischemic Stroke Infarct Core Segmentation from Spatio-temporal CT Perfusion Scans

AU - de Vries, Lucas

AU - Emmer, Bart J.

AU - Majoie, Charles B. L. M.

AU - Marquering, H

AU - Gavves, Efstratios

PY - 2020/7

Y1 - 2020/7

N2 - The infarct core size is a crucial biomarker for treatment selection for ischemic stroke patients. For this purpose, we present a novel approach to perform infarct core segmentation using CT perfusion (CTP) source data, without ordinary deconvolution analysis. We propose the use of transformers to encode sequential CTP scans in spatial attention maps, which we subsequently use for infarct core segmentation. We report new top results on the ISLES 2018 challenge test data set for infarct core segmentation. This work presents a primary benchmark for infarct core segmentation from CTP source data using transformers.

AB - The infarct core size is a crucial biomarker for treatment selection for ischemic stroke patients. For this purpose, we present a novel approach to perform infarct core segmentation using CT perfusion (CTP) source data, without ordinary deconvolution analysis. We propose the use of transformers to encode sequential CTP scans in spatial attention maps, which we subsequently use for infarct core segmentation. We report new top results on the ISLES 2018 challenge test data set for infarct core segmentation. This work presents a primary benchmark for infarct core segmentation from CTP source data using transformers.

KW - : CT Perfusion

KW - Infarct core segmentation

KW - Ischemic stroke

KW - Transformers

M3 - Conference contribution

BT - Medical Imaging with Deep Learning Lübeck, 7 ‑ 9 July 2021

ER -