CHICAGO - Performing maximum intensity projection (MIP) reformatting of MDCT studies adds significant utility for finding pulmonary nodules less than 10 mm in size, according to research from the University Medical Center Hamburg-Eppendorf in Hamburg, Germany.
"MIP reformations on the basis of 64-slice MDCT datasets offer significantly higher detection rates in the detection of small pulmonary nodules," said Dr. Michael Groth. He presented the team's research during a scientific session at the 2008 RSNA meeting in Chicago.
The German research team sought to prospectively evaluate axial standard reconstructions (SR), MIP, and volume rendering (VR) in 25 consecutive patients with pulmonary nodules. SR was performed with a slice thickness of 5 mm and an increment of 5 mm, while MIP and VR-rendered images were created with a slice thickness of 10 mm and an increment of 5 mm.
Two junior observers, with one and two years of experience, respectively, and one senior observer with seven years of experience separately interpreted the studies and provided their diagnostic confidence level on a three-point grading scale (1 = definitely positive, 2 = probably positive, and 3 = finding probably related to a nodule).
Reviewing time was also assessed, and consensus readout using 1 mm reconstructions served as the reference standard, Groth said. Statistical analysis was performed using Wilcoxon's matched-pairs test.
Of the 25 patients, 115 nodules were found. Fifty-seven were sized from 1 mm to 4 mm, and 45 were 5 mm to 9 mm. Thirteen were greater than 10 mm.
Detection rates for small nodules were significantly higher from the MIP images than with SR or VR images, irrespective of reader's experience, Groth said. The difference was statistically significant (p < 0.05).
The experienced reader was able to detect more nodules with the VR images than with the SR images (p = 0.03). However, junior reviewers detected fewer nodules reading the VR images.
Junior readers realized a significant increase in average diagnostic confidence in the small nodules with MIP (1.21), compared with SR (1.35, p = 0.027) and VR images (1.32, p = 0.03). Average reading times for all readers were 91 seconds with MIP, 102 seconds for SR, and 113 seconds for the VR images.
MIP reformations allow for higher diagnostic confidence and shorter reviewing times, Groth said.
"VR is not appropriate for small nodule detection, especially in the hands of the [inexperienced] reader," Groth said.
By Erik L. Ridley
AuntMinnie.com staff writer
December 3, 2008
Related Reading
Nodule measurement streamlined with space-scale analysis, November 25, 2008
CT databases aimed at lung imaging research, August 8, 2008
Ground-glass nodule features on CT reveal malignancy risk, April 29, 2008
CAD shines in detecting large nonsolid pulmonary nodules, February 27, 2008
Measuring lung nodule growth beyond variability, November 5, 2007
Copyright © 2008 AuntMinnie.com
