Keywords:
Thorax, Lung, CT, Image manipulation / Reconstruction, Computer Applications-General, Radiation safety, Artifacts, Dosimetric comparison
Authors:
S. Carey1, S. Kandel1, J. Kavanagh2, T. Chung1, C. Farrell1, P. Rogalla1; 1Toronto, ON/CA, 2Toronto/CA
DOI:
10.26044/ecr2019/C-3543
Aims and objectives
Conventional chest x-ray (CXR) is often used as a first-step diagnostic tool as it is cheaper,
quicker,
and has a reduced radiation dose as compared to standard computed tomography (CT).
Conversely,
CXR suffers from reduced sensitivity and carries an increased false-negative rate for many indications[1][2][3][4].
The primary advantage then of CXR with respect to patient health is the reduced radiation exposure.
The effective radiation dose of an Ultra Low-Dose (ULD) CT protocol can be lowered to a similar range as that of plain film radiography (~0.18 mSv),
however,
increased voxel noise is incurred as a tradeoff with reduced dose.
The increased noise can be mitigated by suitable post-processing; to this end we refine a projection algorithm introduced in a prior work[5] and apply it to ULDCT images to produce a new thick-slice image,
henceforth the thoracic tomogram.
The aim of the study was to compare thoracic tomograms with CXR images with respect to clinical interpretation characteristics,
time to read,
and reader confidence.