Related Experiment Videos

Intra- and inter-observer agreement of brain MRI lesion volume measurements in multiple sclerosis. A comparison of

M Filippi1, M A Horsfield, S Bressi

  • 1Department of Neurology, Scientific Institute Ospedale San Raffaele, University of Milan, Italy.

Insights

A new semi-automated MRI technique significantly improves agreement for measuring multiple sclerosis lesion load. This method offers higher accuracy than manual or arbitrary scoring, crucial for multicenter studies and treatment monitoring.

Area of Science:

  • Neurology
  • Medical Imaging
  • Biostatistics

Background:

  • Magnetic Resonance Imaging (MRI) lesion load is vital for tracking multiple sclerosis (MS) progression and treatment effectiveness.
  • Accurate and reproducible lesion measurement is essential, especially in multicenter studies involving multiple observers or serial scans.
  • Current methods like arbitrary scoring systems (ASS) and manual tracing methods (MTM) may lack sufficient inter- and intra-observer agreement.

Purpose of the Study:

  • To evaluate the intra- and inter-observer agreement of a semi-automated lesion volume measurement technique using thresholding in multiple sclerosis (MS).
  • To compare the agreement of the semi-automated technique with ASS and MTM.
  • To assess the reliability of the semi-automated technique for quantifying MS lesion volumes.

Main Methods:

  • Brain MRIs from 20 patients with clinically definite multiple sclerosis were independently analyzed by three observers.
  • Lesion volumes were measured using an arbitrary scoring system (ASS), a manual tracing method (MTM), and a semi-automated thresholding technique.
  • Intra- and inter-observer agreements were calculated for each method, with a consensus threshold evaluation for the semi-automated technique.

Main Results:

  • The semi-automated technique demonstrated higher median intra-observer agreement (96.3%) and inter-observer agreement (93.7%) compared to ASS and MTM.
  • When a consensus threshold was used, the semi-automated technique achieved even higher agreements: 98.5% (intra-observer) and 96.1% (inter-observer).
  • The semi-automated method showed significantly greater agreement than both ASS and MTM, with significantly lower intra-observer variability (P < 0.0001).

Conclusions:

  • Semi-automated thresholding is a reliable technique for quantifying lesion volumes in multiple sclerosis (MS) with high intra- and inter-observer agreements.
  • This technique offers superior reproducibility compared to arbitrary scoring and manual tracing methods.
  • The semi-automated approach is well-suited for multicenter studies, though a single observer is still recommended for optimal results.

Related Concept Videos