Validation and iteration of CT perfusion defined malignant profile thresholds for acute ischemic stroke

Kevin J. Keenan, Soren Christensen, Manabu Inoue, Michael Mlynash, Gregory W. Albers, Wade S. Smith

Research output: Contribution to journalArticle

Abstract

Background: Malignant profile computed tomography perfusion (CTP) lesions are associated with poor outcomes after administration of intravenous tissue-plasminogen activator (IV-tPA) for ischemic stroke. Aims: To determine whether published CTP-based lesion thresholds predictive of poor outcomes in a predominantly 8 cm of CTP anatomic coverage cohort would predict poor outcomes in an independent 4 cm of CTP anatomic coverage cohort and to generate optimized 4 cm CTP thresholds. Methods: Ischemic stroke patients with baseline CTP imaging with 4 cm of anatomic coverage before receiving IV-tPA at a single institution were retrospectively studied. Perfusion lesion time to maximum of tissue residue function (Tmax) and cerebral blood flow (CBF) volumes were determined using RAPID automated software. Fisher's exact tests assessed associations between lesion thresholds and outcomes. Receiver operating characteristic (ROC) curves generated optimized thresholds for 4 cm of CTP coverage. Results: Sixty-three patients were included. Poor outcomes were associated with published thresholds of Tmax >6 s > 103 mL, Tmax > 8 s > 86 mL, and Tmax > 10 s > 78 mL but not CBF core >53 mL. Thresholds optimized for 4 cm of CTP coverage and associated with poor outcomes were Tmax > 6 s > 100 mL, Tmax > 8 s > 65 mL, Tmax >10 s > 46 mL, and CBF core >39 mL. Conclusions: We validated the ability of published CTP Tmax lesion volume thresholds to predict poor outcomes despite IV-tPA in an independent cohort using only 4 cm of CTP anatomical coverage. A CBF > 39 mL threshold, rather than the predominantly 8 cm CTP coverage derived CBF threshold of >53 mL, was associated with poor outcomes in this 4 cm CTP coverage cohort.

Original languageEnglish (US)
JournalInternational Journal of Stroke
DOIs
StateAccepted/In press - Jan 1 2019
Externally publishedYes

Fingerprint

Perfusion
Stroke
Tomography
Cerebrovascular Circulation
Tissue Plasminogen Activator
Perfusion Imaging
ROC Curve
Intravenous Administration
Software

Keywords

  • Acute stroke imaging
  • computed tomography perfusion
  • CT scan
  • ischemic stroke
  • malignant profile
  • perfusion imaging

ASJC Scopus subject areas

  • Neurology

Cite this

Validation and iteration of CT perfusion defined malignant profile thresholds for acute ischemic stroke. / Keenan, Kevin J.; Christensen, Soren; Inoue, Manabu; Mlynash, Michael; Albers, Gregory W.; Smith, Wade S.

In: International Journal of Stroke, 01.01.2019.

Research output: Contribution to journalArticle

Keenan, Kevin J. ; Christensen, Soren ; Inoue, Manabu ; Mlynash, Michael ; Albers, Gregory W. ; Smith, Wade S. / Validation and iteration of CT perfusion defined malignant profile thresholds for acute ischemic stroke. In: International Journal of Stroke. 2019.
@article{bc49d6ae38884d9185b633c8cb5a94b4,
title = "Validation and iteration of CT perfusion defined malignant profile thresholds for acute ischemic stroke",
abstract = "Background: Malignant profile computed tomography perfusion (CTP) lesions are associated with poor outcomes after administration of intravenous tissue-plasminogen activator (IV-tPA) for ischemic stroke. Aims: To determine whether published CTP-based lesion thresholds predictive of poor outcomes in a predominantly 8 cm of CTP anatomic coverage cohort would predict poor outcomes in an independent 4 cm of CTP anatomic coverage cohort and to generate optimized 4 cm CTP thresholds. Methods: Ischemic stroke patients with baseline CTP imaging with 4 cm of anatomic coverage before receiving IV-tPA at a single institution were retrospectively studied. Perfusion lesion time to maximum of tissue residue function (Tmax) and cerebral blood flow (CBF) volumes were determined using RAPID automated software. Fisher's exact tests assessed associations between lesion thresholds and outcomes. Receiver operating characteristic (ROC) curves generated optimized thresholds for 4 cm of CTP coverage. Results: Sixty-three patients were included. Poor outcomes were associated with published thresholds of Tmax >6 s > 103 mL, Tmax > 8 s > 86 mL, and Tmax > 10 s > 78 mL but not CBF core >53 mL. Thresholds optimized for 4 cm of CTP coverage and associated with poor outcomes were Tmax > 6 s > 100 mL, Tmax > 8 s > 65 mL, Tmax >10 s > 46 mL, and CBF core >39 mL. Conclusions: We validated the ability of published CTP Tmax lesion volume thresholds to predict poor outcomes despite IV-tPA in an independent cohort using only 4 cm of CTP anatomical coverage. A CBF > 39 mL threshold, rather than the predominantly 8 cm CTP coverage derived CBF threshold of >53 mL, was associated with poor outcomes in this 4 cm CTP coverage cohort.",
keywords = "Acute stroke imaging, computed tomography perfusion, CT scan, ischemic stroke, malignant profile, perfusion imaging",
author = "Keenan, {Kevin J.} and Soren Christensen and Manabu Inoue and Michael Mlynash and Albers, {Gregory W.} and Smith, {Wade S.}",
year = "2019",
month = "1",
day = "1",
doi = "10.1177/1747493019832987",
language = "English (US)",
journal = "International Journal of Stroke",
issn = "1747-4930",
publisher = "Wiley-Blackwell",

}

TY - JOUR

T1 - Validation and iteration of CT perfusion defined malignant profile thresholds for acute ischemic stroke

AU - Keenan, Kevin J.

AU - Christensen, Soren

AU - Inoue, Manabu

AU - Mlynash, Michael

AU - Albers, Gregory W.

AU - Smith, Wade S.

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Background: Malignant profile computed tomography perfusion (CTP) lesions are associated with poor outcomes after administration of intravenous tissue-plasminogen activator (IV-tPA) for ischemic stroke. Aims: To determine whether published CTP-based lesion thresholds predictive of poor outcomes in a predominantly 8 cm of CTP anatomic coverage cohort would predict poor outcomes in an independent 4 cm of CTP anatomic coverage cohort and to generate optimized 4 cm CTP thresholds. Methods: Ischemic stroke patients with baseline CTP imaging with 4 cm of anatomic coverage before receiving IV-tPA at a single institution were retrospectively studied. Perfusion lesion time to maximum of tissue residue function (Tmax) and cerebral blood flow (CBF) volumes were determined using RAPID automated software. Fisher's exact tests assessed associations between lesion thresholds and outcomes. Receiver operating characteristic (ROC) curves generated optimized thresholds for 4 cm of CTP coverage. Results: Sixty-three patients were included. Poor outcomes were associated with published thresholds of Tmax >6 s > 103 mL, Tmax > 8 s > 86 mL, and Tmax > 10 s > 78 mL but not CBF core >53 mL. Thresholds optimized for 4 cm of CTP coverage and associated with poor outcomes were Tmax > 6 s > 100 mL, Tmax > 8 s > 65 mL, Tmax >10 s > 46 mL, and CBF core >39 mL. Conclusions: We validated the ability of published CTP Tmax lesion volume thresholds to predict poor outcomes despite IV-tPA in an independent cohort using only 4 cm of CTP anatomical coverage. A CBF > 39 mL threshold, rather than the predominantly 8 cm CTP coverage derived CBF threshold of >53 mL, was associated with poor outcomes in this 4 cm CTP coverage cohort.

AB - Background: Malignant profile computed tomography perfusion (CTP) lesions are associated with poor outcomes after administration of intravenous tissue-plasminogen activator (IV-tPA) for ischemic stroke. Aims: To determine whether published CTP-based lesion thresholds predictive of poor outcomes in a predominantly 8 cm of CTP anatomic coverage cohort would predict poor outcomes in an independent 4 cm of CTP anatomic coverage cohort and to generate optimized 4 cm CTP thresholds. Methods: Ischemic stroke patients with baseline CTP imaging with 4 cm of anatomic coverage before receiving IV-tPA at a single institution were retrospectively studied. Perfusion lesion time to maximum of tissue residue function (Tmax) and cerebral blood flow (CBF) volumes were determined using RAPID automated software. Fisher's exact tests assessed associations between lesion thresholds and outcomes. Receiver operating characteristic (ROC) curves generated optimized thresholds for 4 cm of CTP coverage. Results: Sixty-three patients were included. Poor outcomes were associated with published thresholds of Tmax >6 s > 103 mL, Tmax > 8 s > 86 mL, and Tmax > 10 s > 78 mL but not CBF core >53 mL. Thresholds optimized for 4 cm of CTP coverage and associated with poor outcomes were Tmax > 6 s > 100 mL, Tmax > 8 s > 65 mL, Tmax >10 s > 46 mL, and CBF core >39 mL. Conclusions: We validated the ability of published CTP Tmax lesion volume thresholds to predict poor outcomes despite IV-tPA in an independent cohort using only 4 cm of CTP anatomical coverage. A CBF > 39 mL threshold, rather than the predominantly 8 cm CTP coverage derived CBF threshold of >53 mL, was associated with poor outcomes in this 4 cm CTP coverage cohort.

KW - Acute stroke imaging

KW - computed tomography perfusion

KW - CT scan

KW - ischemic stroke

KW - malignant profile

KW - perfusion imaging

UR - http://www.scopus.com/inward/record.url?scp=85062020015&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85062020015&partnerID=8YFLogxK

U2 - 10.1177/1747493019832987

DO - 10.1177/1747493019832987

M3 - Article

AN - SCOPUS:85062020015

JO - International Journal of Stroke

JF - International Journal of Stroke

SN - 1747-4930

ER -