Lung nodules are reliably detectable on ultra-low-dose CT utilising model-based iterative reconstruction with radiation equivalent to plain radiography

A. R. Miller, D. Jackson, C. Hui, S. Deshpande, E. Kuo, G. S. Hamilton, K. K. Lau

Research output: Contribution to journalArticleResearchpeer-review

18 Citations (Scopus)

Abstract

AIM: To determine if ultra-low-dose (ULD) computed tomography (CT) utilising model-based iterative reconstruction (MBIR) with radiation equivalent to plain radiography allows the detection of lung nodules. MATERIALS AND METHODS: Ninety-nine individuals undergoing surveillance of solid pulmonary nodules undertook a low-dose (LD) and ULD CT during the same sitting. Image pairs were read blinded, in random order, and independently by two experienced thoracic radiologists. With LD-CT as the reference standard, the number, size, and location of nodules was compared, and inter-rater agreement was established. RESULTS: There was very good inter-rater agreement with regards nodules ≥4mm for both the LD- (k=0.931) and ULD-CT (k=0.869). One hundred and ninety-nine nodules were reported on the LD-CT by both radiologists and 196 reported on the ULD-CT, with no nodules reported only on the ULD-CT. This gives a sensitivity of 98.5% and specificity of 100% for ULD-CT with MBIR. The effective dose of radiation was significantly different between the two scans (p<0.0001), 1.67 mSv for the LD-CT and 0.13 mSv for the ULD-CT. CONCLUSION: ULD-CT utilising MBIR and delivering radiation equivalent to plain radiography, allows detection of lung nodules with high sensitivity. The attendant 10-fold reduction in radiation may allow for dramatic reductions in cumulative radiation exposure.

Original languageEnglish
Pages (from-to)409.e17-409.e22
Number of pages6
JournalClinical Radiology
Volume74
Issue number5
DOIs
Publication statusPublished - May 2019

Keywords

  • Computed tomography
  • Multiple Pulmonary Nodules
  • Radiation dosage
  • Solitary Pulmonary Nodule

Cite this