Article Text

Download PDFPDF
25 A study of the different methodologies used in calculation of extra-cellular volume by CMR imaging
  1. Ravi Vijapurapu,
  2. Francesca Hawkins,
  3. Boyang Liu,
  4. Nicola Edwards,
  5. Richard Steeds
  1. Department of Cardiology, Queen Elizabeth Hospital, Birmingham, UK


Introduction Tissue characterisation with MRI derived extra-cellular volume (ECV) measurement using T1 mapping is a well-recognised tool guiding clinical diagnosis and treatment. Concerns regarding blood-pool artefact have led to multiple methodologies with no consistency in assessment. This study investigates how ECV values and inter-observer reproducibility varies depending on technique.

Methods Patients with primary mitral regurgitation (MR) underwent multiparametric CMR imaging (1.5 T Siemens Avanto). ECV was calculated using MOLLI-derived T1 mapping (mid- and basal-LV, CircleCvi42) by two methods: global segmental analyses with variable endocardial:epicardial percentage offsets (20:20, 0:50, 10:50, 20:50) and region of interest (ROI). A second observer independently analysed a subset for inter-observer reproducibility, with independent exclusion of artefactual segments.

Results Thirty patients were studied (LVEF 72%±7%, LVEDVi 101±19 ml/m2, LVESVi 30±11 ml/m2, MR fraction 41%±16%, age 66±11 years). ECV values were lowest using ROI (septum 26.0%±2.5%, anterior 26.0%±0.26%, posterior 26.3%±0.39%), followed by 20:20 offset (28.0%±3.3%), and were highest with 0:50 offset (29.9%±3.4%). A 20:20 offset approach was most reproducible following 8% segmental exclusion (mean bias 0.21%±0.12%, intraclass correlation (ICC) 0.994), followed by 20:50, 10:50, and 0:50 offsets (mean bias 0.39%±0.24%, ICC 0.992). Anterior ROI was more reproducible than septal (mean bias 0.14%±0.72%, ICC 0.96 vs 0.48%±0.74%, ICC 0.929).

Conclusion There is considerable variability in ECV derived using ROI compared with a segmental approach, highlighting the importance of consistency within analyses. ECV calculation using septal ROI is commonly reported in the literature, but we have shown that a global segmental approach using 20:20 offsets have superior reproducibility, despite variable exclusion of artefactual segments.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.