Clinical study
CT perfusion source images improve identification of early ischaemic change in hyperacute stroke

https://doi.org/10.1016/j.jocn.2005.03.030Get rights and content

Abstract

CT perfusion scanning produces dynamic contrast-enhanced brain images, but to generate and interpret the colour maps of cerebral perfusion from these source images requires specialist neuroimaging knowledge. We hypothesized that physicians without such training might still utilise the CT perfusion source images (CTPSI) to diagnose early ischaemic change. Fifteen patients had cerebral non-contrast CT (NCCT) and perfusion CT within 6 hours of hemispheric stroke onset. We tested 15 non-stroke clinicians and radiology trainees, plus three experts, in assessing the presence and extent of early ischaemic change on NCCT versus CTPSI. Day 5–7 CT or MRI was used as the gold standard. Agreement with follow-up imaging was poor for both detection, and extent of early ischaemic change on NCCT (κ = 0.01–0.11). There was a marked improvement in agreement for both the presence and extent of early ischaemic change on CTPSI (κ = 0.67–0.83). CTPSI were much more accurate than NCCT in identifying acute ischaemic change. ‘Less expert’ users accurately identified major early ischaemic change on acute CTPSI. These findings suggest that such physicians might utilise CTPSI to screen potential thrombolysis candidates.

Introduction

Thrombolytic therapy with tissue plasminogen activator (tPA) improves outcome in selected acute stroke patients.1 Avoidance of tPA has been recommended in patients with non-contrast CT evidence of major early ischaemic change.2 However, the significance of major early ischaemia on NCCT is controversial.[3], [4] This may, in part, relate to the fact that the changes of acute ischaemia on NCCT are often too subtle for many observers to identify accurately.[5], [6], [7]

Perfusion CT imaging (CTP) using new generation multi-slice scanners is a relatively recent technique.[8], [9], [10] CTP relies on the rapid speed of modern helical scanners to track the passage of a bolus of intravenous ionic contrast.10 Colour maps of cerebral perfusion can be subsequently generated from the dynamic contrast-enhanced brain source images.[9], [11] The colour perfusion maps have great potential, but specialised expertise in reliably generating and interpreting these maps from the contrast source images is required. However, the contrast-enhanced CTP source images (CTPSI) are generated automatically during the CTP sequence, and do not require any post-processing. Medical staff that regularly assess acute stroke patients, but who do not have specialist neuroimaging knowledge, can immediately access the CTPSI at the CT workstation.

We have observed that it is much easier to appreciate early ischaemic change on the CTP source images compared to NCCT (Fig. 1). This may reflect lack of contrast enhancement in tissue with disordered perfusion. It is possible that such changes are similar to those seen with the source images used for CT angiography (CTA-SI).12 However, the ability of CTP source images to identify early ischaemia, and indeed whether these CTPSI lesions represent irreversible and/or reversible ischaemic change, has not been previously assessed. Therefore, the aims of this study were two-fold:

  • 1.

    To test the accuracy of observers with varying levels of experience in diagnosing the presence and extent of early ischaemic change on CTPSI compared to NCCT.

  • 2.

    To test the mechanisms and validity of CTP source images in identifying early ischaemic change by comparing them with the derived acute colour CTP maps, acute CT angiography source images, and the outcome infarct.

Section snippets

Patients

We prospectively studied 15 consecutive patients (eight women, mean age 68 ± 13 years; median National Institutes of Health Stroke Scale (NIHSS) = 15, inter-quartile range, 11–20) with suspected anterior circulation infarction, between July and December 2003. Patients had cerebral non-contrast CT, perfusion CT, and CT angiography within 6 hours of onset of stroke symptoms (mean, 4.1 ± 1.8 hours).13 Four patients received intravenous thrombolysis with tPA. Follow-up CT (eight patients) or MRI was

Assessment of early ischaemic change by different observers

On follow-up imaging, 14 of the 15 patients had evidence of recent infarction. None of the observer groups, including the experts, correctly identified more than seven of these 14 on acute NCCT (Table 1a). The inter-observer variability for any early ischaemic change on acute NCCT compared to follow-up imaging was poor. In comparison to acute NCCT, all groups performed much better at identifying the presence of early ischaemic change on CTPSI (Table 1a). The inter-observer variability for all

Discussion

In this study CTPSI were much more accurate than NCCT in identifying the presence and extent of acute ischaemic change by observers of all levels of experience. The specificity of a truly normal NCCT for excluding major ischaemic change should not be underestimated, as this may identify patients with the most to gain from thrombolytic therapy.[3], [15] However, CTPSI was more robust than NCCT in truly identifying patients with extensive early ischaemic change, and at least as accurate as NCCT

Acknowledgements

Supported in part by grants from the Australian Brain Foundation (2003; Dr Parsons) and Cardiovascular Lipid Grant (Pfizer, 2004: Dr Parsons).

The authors would like to thank the emergency, internal medicine, and radiology doctors who participated in this study. We also thank our stroke research nurses and CT radiographers.

References (17)

  • Tissue plasminogen activator for acute ischaemic stroke. The national institute of neurological disorders and stroke...
  • Practice advisory: Thrombolytic therapy for acute ischaemic stroke–summary statement. Report of the quality standards...
  • R. von Kummer et al.

    Acute stroke: Usefulness of early CT findings before thrombolytic therapy

    Radiology

    (1997)
  • S.C. Patel et al.

    Lack of clinical significance of early ischaemic changes on computed tomography in acute stroke

    JAMA

    (2001)
  • S.H. Horowitz et al.

    Computed tomographic-angiographic findings within the first five hours of cerebral infarction

    Stroke

    (1991)
  • R. von Kummer et al.

    Interobserver agreement in assessing early CT signs of middle cerebral artery infarction

    AJNR

    (1996)
  • J.C. Grotta et al.

    Agreement and variability in the interpretation of early CT changes in stroke patients qualifying for intravenous rtpa therapy

    Stroke

    (1999)
  • K. Kudo et al.

    Quantitative cerebral blood flow measurement with dynamic perfusion ct using the vascular-pixel elimination method: Comparison with H2(15)O positron emission tomography

    AJNR Am J Neuroradiol

    (2003)
There are more references available in the full text version of this article.

Cited by (30)

  • Brain CT perfusion improves intracranial vessel occlusion detection on CT angiography

    2019, Journal of Neuroradiology
    Citation Excerpt :

    Yet, CTP derived parameters cannot reliably identify patients who will benefit from IAT [16]. It has, however, been reported that less experienced observers more accurately identify major ischemic changes on CTP source images than on NCCT [12], and that 4D-CTA reformatting of CTP increases the diagnostic accuracy and certainty in proximal vessel occlusion detection [17]. In this study we investigate whether brain CTP aids in the detection of intracranial vessel occlusion on CTA.

  • Mechanical thrombectomy in acute ischemic stroke

    2017, Revue Neurologique
    Citation Excerpt :

    MRI can provide an estimate of the salvageable tissue as can CT (e.g. Tmax threshold > 6s) [20–22]. CT perfusion source imaging might also offer a better core estimate than non-contrast CT [23]. The ASPECTS is easier to use, but exhibits a mediocre correlation with absolute lesion volumes, especially for intermediary scores (ASPECTS 4–6) [24,25].

  • The selections of acute stroke patients for catheter based intervention

    2016, Cor et Vasa
    Citation Excerpt :

    Ischemic stroke results from impaired cranial perfusion due to total blockade or slowing of cranial blood flow in at least one of the cerebral vessels. Together with impaired cerebral blood flow, an area referred to as the “core zone” surrounded by another zone, i.e. “penumbra” develop [1,2]. Core is characterized by tissue necrosis occurring due to decline in blood flow below a critical threshold, and even successful reperfusion does not lead to tissue repair in this area.

  • Added value of CT perfusion and CT angiography in detection of early ischemic stroke: Comparison with noncontrast CT

    2015, Egyptian Journal of Radiology and Nuclear Medicine
    Citation Excerpt :

    After additional analysis of CTP, the sensitivity significantly increased to 100%. This was in agreement with previous studies which demonstrated increased diagnostic sensitivity using CTP compared with non-contrast CT (NCCT) (10–12). Kloska et al. (7) found a sensitivity for CTP of 76%, whereas Koenig et al. (9,13) reported 89% and 100% in small patient cohorts.

  • Clinical use of computed tomographic perfusion for the diagnosis and prediction of lesion growth in acute ischemic stroke

    2014, Journal of Stroke and Cerebrovascular Diseases
    Citation Excerpt :

    The failure of CBV in the detection of the ischemic core and matching the DWI lesions was not completely unexpected. CBV has been shown to be a good surrogate of the ischemic core only in a few studies using large MCA territorial strokes, and research-grade postprocessing software is not available to clinicians.5,26,27 In addition, CBV lesions can be reversible and appear to be time-dependent.28

  • Prevention and Treatment of Stroke

    2013, Vascular Medicine: A Companion to Braunwald's Heart Disease: Second Edition
View all citing articles on Scopus
View full text