Keywords:
Lung, Pulmonary vessels, Artificial Intelligence, CT-Angiography, CAD, CT, Computer Applications-Detection, diagnosis, Contrast agent-intravenous
Authors:
N. Hendriks1, J. J. Zigterman2, M. Roelofs1, J. Nijboer-Oosterveld1, E. De Boer1, M. F. Boomsma1; 1Zwolle/NL, 2Zwolle, Overijsel/NL
DOI:
10.26044/ecr2019/C-2541
Aims and objectives
Acute pulmonary embolism (PE) is a common pulmonary disease which could lead to health problems if not diagnosed in time.
The current reference standard the radiologist uses to diagnose PE is a Computer Tomography Pulmonary Angiogram (CTPA)1.
The sensitivity of the CTPA is estimated 83%-100% and its specificity between 89%-98%2-4.
Adequate use of Computer-Aided Detection (CAD) techniques may be a resource to support the radiologist in the detection of PE5-7.
Scan quality is proven to be of influence on the software performance.
Scan quality,
defined as signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR),
is affected by the use of different levels of Hybrid Iterative Reconstruction (HIR)8-11.
New reconstructors make it possible to reconstruct the CTPA data with Model-based Iterative Reconstruction (MBIR) software.
MBIR is the next generation of iterative reconstruction software with more noise reduction,
better spatial- and contrast resolution and better artifact reduction12.
This retrospective cohort study examines whether there is evidence of improvement in software performance on a CAD programme for the detection of peripheral pulmonary embolism,
while maintaining routine dose protocols using MBIR in comparison to the current used HIR.