Automated detection and volume measurement of plexiform neurofibromas in neurofibromatosis 1 using magnetic resonance imaging

Comput Med Imaging Graph. 2004 Jul;28(5):257-65. doi: 10.1016/j.compmedimag.2004.03.002.

Abstract

An automated technique for segmentation and volumetric measurement of plexiform neurofibromas (PN) in neurofibromatosis 1 using short T1-inversion recovery magnetic resonance images is presented. The algorithm described implements heuristics derived from human-based recognition of lesions. This technique combines region-based with boundary-based segmentation. Two observers, who performed semi-automated volume calculations and manual tracings to estimate tumor volume, validated this method on 9 PNs of different size and location. This automated method was reproducible (coefficient of variation 0.6-5.6%), yielded similar results to manual tumor tracings (R = 0.999) and will likely improve the ability to measure PNs in ongoing clinical trials.

MeSH terms

  • Algorithms
  • Automation
  • Humans
  • Magnetic Resonance Imaging
  • Neurofibroma, Plexiform / pathology*
  • Neurofibromatosis 1 / pathology*
  • Reproducibility of Results
  • Sensitivity and Specificity
  • United States