A diagnostic decision tree for adult cerebellar ataxia based on pontine magnetic resonance imaging

J Neurol Sci. 2018 Apr 15:387:187-195. doi: 10.1016/j.jns.2018.02.022. Epub 2018 Feb 12.

Abstract

Cerebellar ataxias (CAs) are heterogeneous conditions often require differential diagnosis. This study aimed to establish a diagnostic decision tree for differentiating CAs based on pontine MRI findings. Two-hundred and two consecutive ataxia patients were clinically classified into 4 groups: (1) spinocerebellar ataxia (SCA) with brainstem involvement (SCA-BSI), (2) Pure cerebellar SCA, (3) cerebellar dominant multiple system atrophy (MSA-c), and (4) Other CA. Signal intensity in pons was graded into 3 types: hot cross bun sign (HCBS), pontine midline linear T2-hyperintensity (PMH), or normal. The distance ratio of pontine base to tegmentum, named "BT-ratio", was measured. The presence of HCBS indicated either MSA-c with a specificity of 97.7%, or SCA2. When PMH was observed, a BT-ratio above 3.54 strongly indicated SCA-BSI, namely Machado-Joseph disease, SCA1, or dentatorubral-pallidoluysian atrophy, whereas a BT-ratio below 3.54 indicated MSA-c or SCA2. When the signal intensity was normal, a BT-ratio above 3.52 indicated SCA-BSI, whereas a BT-ratio below 3.52 suggested Pure cerebellar SCA or Other CA with pure cerebellar type. The decision tree was confirmed useful in a different 30 CA patients. We propose that differential diagnosis of CAs can be supported by combining pontine MRI signal intensity changes and BT-ratio.

Keywords: Magnetic resonance imaging; Multiple system atrophy; Pons; Spinocerebellar ataxia.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Aged
  • Cerebellar Ataxia / classification
  • Cerebellar Ataxia / diagnostic imaging*
  • Cerebellum / diagnostic imaging
  • Cohort Studies
  • Decision Trees*
  • Female
  • Humans
  • Image Processing, Computer-Assisted
  • Magnetic Resonance Imaging*
  • Male
  • Middle Aged
  • Pons / drug effects*
  • ROC Curve