Optimization of a breath-hold magnetic resonance gradient echo technique for the detection of interstitial lung disease

Invest Radiol. 1995 Dec;30(12):730-7. doi: 10.1097/00004424-199512000-00007.

Abstract

Rationale and objectives: A rapid gradient echo technique was optimized to allow breath-held imaging of the lung parenchyma. The ability of this sequence to detect interstitial lung disease was then compared with high resolution computed tomography.

Methods: A Turbo-FLASH gradient echo sequence (repetition time [TR] 4.7, echo time [TE]2) was optimized using four volunteers. The parameters varied, including flip angle, inversion time, slice thickness, and number of measurements per unit time. The effect on signal to noise ratio of lung inflation and position within the lung were also studied. The optimized sequence was then compared with high resolution computed tomography in five patients being investigated for interstitial lung disease.

Results: Greatest signal was achieved using no inversion prepulse. This allowed up to six measurements, each with two acquisitions, during a single breath-hold. Optimum signal to noise ratio was achieved using a flip angle of 10 degrees with a 20-mm slice thickness, although a 10-mm slice was felt to be preferable for clinical use. Greatest signal to noise ratio was generated in the posterior aspect of the lung in full expiration. Comparison of the optimized magnetic resonance imaging technique with high resolution computed tomography showed no significant difference (P = 0.17) when assessing the proportion of diseased to normal lung.

Conclusions: Using standard magnetic resonance software and hardware, it has been possible to use a breath-hold sequence that allows detection and quantification of interstitial lung disease.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Echo-Planar Imaging / methods*
  • Humans
  • Image Enhancement
  • Lung / pathology*
  • Pulmonary Fibrosis / diagnosis*
  • Pulmonary Ventilation / physiology*
  • Reference Values