G Soo1, K K Lau, T Yik, P Kutschera. 1. Diagnostic Imaging, Southern Health, Monash Medical Centre, Clayton, Victoria, Australia. soogeoffrey@gmail.com
Abstract
AIM: To evaluate the impact of different reconstructed section thicknesses on liver lesion detection using multidetector computed tomography (CT). METHODS: Fifty-three patients were examined using a 16-section CT machine with axial reconstructions provided at 2.5, 5, 7.5, and 10mm section thicknesses. Images of different reconstructed section thicknesses from different patients were presented in random order to three independent, blinded radiologists for review at multiple sessions. All images were then reviewed by three radiologists in a common session. Consensus was reached following review of the previous interpretation results and results of follow-up imaging regarding the number of true liver lesions (n=101) for comparison. RESULTS: Mean detection rates were as follows: 93/101 lesions detected with the 2.5mm section thickness, 98/101 lesions detected at the 5mm section thickness, 78/101 lesions detected at the 7.5mm section thickness, and 54/101 lesions detected at the 10mm section thickness. Lesions missed at the 2.5mm section thickness were due to masking by image noise. There was particular difficulty detecting subcapsular lesions and lesions adjacent to fissures or the gall bladder at the 7.5mm and 10mm section thicknesses. CONCLUSION: The optimal reconstructed section thickness for lesion detection in the liver was 5mm. Copyright (c) 2010 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
AIM: To evaluate the impact of different reconstructed section thicknesses on liver lesion detection using multidetector computed tomography (CT). METHODS: Fifty-three patients were examined using a 16-section CT machine with axial reconstructions provided at 2.5, 5, 7.5, and 10mm section thicknesses. Images of different reconstructed section thicknesses from different patients were presented in random order to three independent, blinded radiologists for review at multiple sessions. All images were then reviewed by three radiologists in a common session. Consensus was reached following review of the previous interpretation results and results of follow-up imaging regarding the number of true liver lesions (n=101) for comparison. RESULTS: Mean detection rates were as follows: 93/101 lesions detected with the 2.5mm section thickness, 98/101 lesions detected at the 5mm section thickness, 78/101 lesions detected at the 7.5mm section thickness, and 54/101 lesions detected at the 10mm section thickness. Lesions missed at the 2.5mm section thickness were due to masking by image noise. There was particular difficulty detecting subcapsular lesions and lesions adjacent to fissures or the gall bladder at the 7.5mm and 10mm section thicknesses. CONCLUSION: The optimal reconstructed section thickness for lesion detection in the liver was 5mm. Copyright (c) 2010 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.