Specificity and sensitivity of magnetic resonance imaging findings in the diagnosis of progressive supranuclear palsy. Issue 1 (22nd April 2017)

• Record Type: Journal Article
• Authors: Bacchi, Stephen; Chim, Ivana; Patel, Sandy
• Is Part Of: Journal of medical imaging and radiation oncology. Volume 62:Issue 1(2018:Feb.)
• Page Start: 21
• Page End: 31
• Publication Date: 2017-04-22
• Subjects: magnetic resonance imaging -- multiple system atrophy -- Parkinsonian disorders -- progressive supranuclear palsy -- support vector machine; Radiology, Medical -- Periodicals; Radiology, Medical -- Australasia -- Periodicals; 616.0757
• DOI: 10.1111/1754-9485.12613
• Languages: English
• ISSNs: 1754-9477

Abstract

Summary: Progressive supranuclear palsy (PSP) is a neurodegenerative condition that can only be diagnosed conclusively on pathological examination. Currently, the diagnosis is based upon the National Institute of Neurological Disorders and Stroke and the Society for PSP criteria. These criteria consist of purely clinical findings. Elements of brain MRI that are being investigated for this role include identifying structural features on conventional MRI, volume changes, signal abnormalities and diffusion changes. The aim of this study is to conduct a systematic search to identify which MRI findings have evidence to support their sensitivity/specificity/accuracy in the diagnosis of PSP. A search was conducted of Pubmed and Medline on July 5th–6th 2016 using the medical subject headings progressive supranuclear palsy and MRI . Seventy articles were identified which assessed the sensitivity/specificity/accuracy of MRI signs for the diagnosis of PSP. There were 13 studies that identified MRI features that had ≥95% sensitivity and specificity for the diagnosis of PSP. Four of these studies identified the magnetic resonance parkinsonism index as highly sensitive and specific. There were only four studies which assessed how effective given MRI features are at predicting the pathological diagnosis of PSP. Several markers, such as the magnetic resonance parkinsonism index, have been demonstrated to be both specific and sensitive for PSP. However, many studies assessing these markers have common weaknesses including small sample size and lacking autopsy correlation.