Evaluation of an optimized metal artifact reduction algorithm for flat-detector angiography compared to DSA imaging in follow-up after neurovascular procedures

BMC Med Imaging. 2019 Aug 14;19(1):66. doi: 10.1186/s12880-019-0352-2.

Abstract

Background: Flat detector CT - angiography (FDCTA) has become a valuable imaging tool in post- and peri-interventional imaging after neurovascular procedures. Metal artifacts produced by radiopaque implants like clips or coils still impair image quality.

Methods: FDCTA was performed in periprocedural or follow-up imaging of 21 patients, who had received neurovascular treatment. Raw data was sent to a dedicated workstation and subsequently a metal artifact reduction algorithm (MARA) was applied. Two neuroradiologists examined the images.

Results: Application of MARA improved image appearance and led to a significant reduction of metal artifacts. After application of MARA only 8 datasets (34% of the images) were rated as having many or extensive artifacts, before MARA 15 (65%) of the images had extensive or many artifacts. Twenty percent more cases of reperfusion were diagnosed after application of MARA, congruent to the results of digital subtraction angiography (DSA) imaging. Also 3 (13% of datasets) images, which could not be evaluated before application of MARA, could be analyzed after metal artifact reduction and reperfusion could be excluded.

Conclusion: Application of MARA improved image evaluation, reduced the extent of metal artifacts, and more cases of reperfusion could be detected or excluded, congruent to DSA imaging.

Keywords: Angiography; Brain/ brain stem; CT metal artifacts reduction; Coiling; Flat-detector CT angiography.

Publication types

  • Comparative Study

MeSH terms

  • Algorithms
  • Angiography, Digital Subtraction / methods*
  • Artifacts
  • Brain / blood supply
  • Brain / diagnostic imaging*
  • Computed Tomography Angiography / methods*
  • Humans
  • Perioperative Care
  • Postoperative Care
  • Radiographic Image Interpretation, Computer-Assisted / methods*