Details

Autor(en) / Beteiligte

• Janjua, Nazli
• Smolak, Alex
• Bulic, Sebina
• Diaz, Kardelina
• Tan, Kay Sin
• Linden, Craig

Titel

Abstract TP32: ‘ABC/2’ Method for Stroke Measurement has Excellent Intra-Class Correlation for Different Modalities and Raters

Ist Teil von

• Stroke (1970), 2013-02, Vol.44 (suppl_1)

Erscheinungsjahr

2013

Quelle

Alma/SFX Local Collection

Beschreibungen/Notizen

• Abstract only BACKGROUND: The ‘Revascularization in Stroke Based on Clinical-Diffusion Mismatch’ study will evaluate the safety of endovascular therapy for stroke patients beyond 8 hours from onset. The inclusion criteria include diffusion weighted magnetic resonance imaging (DW-MRI) stroke lesion volume </=25 mL3. To apply this criterion in practice, measurement techniques easily performed by any clinician, such as the ‘ABC/2’ formula for intracerebral hemorrhage (ICH), are needed. This has been validated against planimetry for middle cerebral artery (MCA) infarction on DW-MRI. We sought to evaluate the inter-rater reliability between raters and the correlation between different ABC/2 formulas for lesions on DWI-MRI and CT. METHODS: CTs and DWI-MRIs of stroke and hemorrhage patients were screened to select a sample of lesions in the cerebral cortex, subcortex, basal ganglia, brainstem, and cerebellum for each group (CT hemorrhage, CT stroke, DW-MRI stroke). Three raters then measured lesion volume as follows: (AxBxC)/2, where A= largest linear lesion diameter in centimeters (cm) on a single slice; B=largest linear diameter perpendicular to A (cm); C=counts between lesion slices multiplied by slice thickness in cm. The ‘adjusted ABC/2’ formula for ICH weights each scan slice by proportion of hemorrhage to the index slice. To evaluate differences between these techniques, a single rater evaluated each lesion using each of the formulas. Measurements were examined through the intraclass correlation (ICC) statistic, to give a composite score of intra-observer and inter-observer variability. The ICC was tested for being significantly different from 0 yielding the reported p-value (1 indicates good ICC). RESULTS: CONCLUSION: The ABC/2 method has excellent ICC on DW-MRI and CT for different examiners, lesions, and locations as well as minimal variability between different formulas. This may be an alternative to computer planimetry for triage of stroke patients in future studies.