A scale to assess cerebrovascular disease on CT images in patients with cognitive impairment

L. Bresciani, C. Geroldi, A. Beltramello, G. Binetti, O. Zanetti, G. B. Frisoni

Research output: Contribution to journalArticlepeer-review


Objective. To test the validity of a newly developed rating scale to assess cerebrovascular disease (CVD) on CT images. The scale assumes that different types of lesions have different strength to cause cognitive impairment: leukoaraiosis is assumed to be the least efficient, white matter and subcortical lesions are intermediate, and cortical lesions are the most efficient. Within these 3 domains, severity was weighted in order to reflect this hierarchy. Methods. Our hierarchical scale (URL: www.unibs.it/∼grg/epineuro/form.htm) rates presence and severity (number and size, i.e. > 2 cm or <2 cm) of cortical, white matter (frontal, temporal, parietal, occipital), and subcortical (basal ganglia, internal capsule, and cerebellum) lesions, and of leukoaraiosis (score of 0 to 18) as assessed on CT images. CVD scores are:"0", no vascular lesions; 1, isolated mild leukoaraiosis; "2", moderate leukoaraiosis or 1 large or max 2 small lacunes; "3", severe leukoaraiosis or more/larger lacunes; and "4", severe leukoaraiosis with more/larger lacunes or cortical lesions. Known-group and convergent validity were tested in cognitively impaired patients with a MMSE of 18 and higher. Results. Intraclass correlation coefficients for inter-rater and test-retest reliability were between 0.85 and 0.92. CVD as graded with the hierarchical scale was significantly lower in AD than both mild cognitive impairment of the vascular type (MCI-VT) and vascular dementia (1.5±0.9, 1.7±1.3, 2.8±1.0, and 3.0±1.1; P

Original languageEnglish
JournalNeurological Sciences
Issue number4 SUPPL.
Publication statusPublished - 2000

ASJC Scopus subject areas

  • Neuroscience(all)
  • Clinical Neurology


Dive into the research topics of 'A scale to assess cerebrovascular disease on CT images in patients with cognitive impairment'. Together they form a unique fingerprint.

Cite this