« Back
Textural Analysis of Early Features of an Ischaemic stroke on the initial Non-contrast CT head using TexRad™
Poster Title: Textural Analysis of Early Features of an Ischaemic stroke on the initial Non-contrast CT head using TexRad™
Submitted on 07 Nov 2019
Author(s): Mark-Joel Clovis, Terence Jones, Charles Hutchinson
Affiliations: University of Warwick,University Hospital Coventry and Warwickshire, City Hospital Birmingham
This poster was presented at International College of Undergraduate research
Poster Views: 954
View poster »

Poster Information
Abstract: Patients with symptoms of stroke routinely receive a computed tomography radiograph of the head, however, this imaging modality is relatively insensitive to ischaemic changes. This makes it difficult to diagnose ischaemic strokes and therefore to differentiate them from stroke mimics. Research into ischaemic stroke detection by textural analysis has shown that it is possible to detect subacute ischaemic strokes but none have explicitly detected acute strokes. This study aims to use TexRad to separate ischaemic and non-ischaemic tissue based on textural changes earlier than can be determined by visual inspection. Follow up non-enhanced CT heads of 39 patients showing areas of infarction were used to identify corresponding regions of ischaemia on initial scans. TexRadTM was used to analyse textural parameters of the ischaemic and anatomically corresponding non-ischemic regions in initial and follow-up scans. The machine learning tool, Weka, was then trained to recognise patterns in the data and to classify regions as ischaemic or non-ischaemic. The radiology reports of initial scans were used to assess the accuracy of ischaemic stroke detection by radiologists. Use of machine learning algorithms has revealed that it is possible to identify ischaemic strokes with an accuracy of 60.2% (p=0.017) vs 38.4% accuracy by radiologists. Textural analysis has been demonstrated to identify ischaemic strokes with greater accuracy than a reporting radiologist. Conversely, the use of textural analysis may be able to better identify those who do not have ischaemic changes. These patients may undergo further investigations to more accurately diagnose and manage their conditions. Summary: Textural analysis of potential acute ischaemic strokes may provide a novel method to stratify patients who would vs those who may not benefit from thrombolysis. References: 1. WHO. Media centre The top 10 causes of death. Fact sheet N°310 (Updated May 2014) [Internet]. 2014;(June):2012–4. Available from:

2. Feigin VL, Forouzanfar MH, Krishnamurthi R, Mensah GA, Bennett DA, Moran AE, et al. From the Global Burden of Disease Study 2010. 2014;383(9913):245–54.

3. National Institute for Health and Clinical Excellence (NICE). Stroke and transient ischaemic attack in over 16s: diagnosis and initial management, Clinical guideline. 2008;(July). Available from:

4. Kane I, Whiteley WN, Sandercock PAG, Wardlaw JM. Availability of CT and MR for assessing patients with acute stroke. Cerebrovascular Disease. 2008;25(4):375–7.

5. Moulin T, Cattin F, Crepin-Leblond T, Tatu L, Chavot D, Piotin M, et al. Early CT signs in acute middle cerebral artery infarction. Neurology [Internet]. 1996 Aug 1;47(2):366 LP-375. Available from: http://n.neurolo
Report abuse »
Ask the author a question about this poster.
Ask a Question »

Creative Commons

Related Posters

A Unique Case of Hemochromatosis and Pancreatitis
Magana Herrera, Fernando D.O , Fieker, Aaron D.O , Sharma, Niha Haresh MD

Library Construction Service
Marcia Brady

Emaranhado de “red flags”
Maria Beatriz Morgado, Pedro Pacheco

Syndrome de Budd-Chiari sur un myélome multiple compliqué d’une cirrhose : à propos d’un cas
Nada Faquir, Fatima Zahra Belabbes,Maryame Ahnach, Sara Mounsif, Meryem Maizi, Fedoua Rouibaa

Algoritmo de abordagem: Fenómeno de Raynaud nos Cuidados Primários
Maria Beatriz Morgado, Pedro Pacheco