Minho Lee PhD, Presenter: Nothing to Disclose

Namkug Kim PhD, Abstract Co-Author: Nothing to Disclose

Joon Beom Seo MD, PhD, Abstract Co-Author: Nothing to Disclose

Sang Young Oh MD, Abstract Co-Author: Nothing to Disclose

Sang Min Lee MD, Abstract Co-Author: Nothing to Disclose

Jae Seung Lee, Abstract Co-Author: Nothing to Disclose

Yeon-Mok Oh MD, PhD, Abstract Co-Author: Nothing to Disclose

Yongjun Chang, Abstract Co-Author: Nothing to Disclose

To propose a quantification method to classify emphysema clusters by size using length scale analysis in volumetric chest CT.

Volumetric CT scans of twenty patients with chronic obstructive pulmonary disease (COPD) were performed by a 16-multi detector row CT scanner (Siemens Sensation 16) with in 0.75mm collimation. Using thresholding by -950 HU, emphysema index (EI) of low attenuation area (LAA) mask was evaluated. Based on these LAA masks, a length scale analysis to estimate each emphysema cluster’s size was performed as follows. At first, Hole filling algorithm was performed on the emphysema mask and Gaussian low pass filter (LPF) with various size of kernel (<1.5, <7.5, <15, >15mm) was performed from large to small size, iteratively. Maximum density voxel in the each filtered volume was selected and dilated by the size of the kernel, which was regarded as the specific size emphysema mask.  In this way, emphysema cluster with specific size range was classified and evaluated from the LAA mask. The accuracy of this classification result was evaluated and compared by an expert thoracic radiologist with 10 scale visual evaluation to determine size classifying accuracy and to determine probabilities for incorrect estimation. In addition, an artificial phantom study for mimicking emphysema and a COPD patients study were performed to evaluate the accuracy of this algorithm.

In phantom study, in case of sphere-like the shape of emphysema with various sizes from 1mm to 15mm, the method shows exact estimation on every case. In the COPD patients, size based EI were 3.48±1.97%, 12.85±7.07%, 7.07±7.88%, and 4.11±8.22%, (size : <1.5, <7, <15, and >15mm), respectively. In addition, association study between blind visual evaluations of size based EI by an expert thoracic radiologist and our method showed all significant correlations r values : 0.499,0.725,0.768,0.939, respectively) and probabilities for incorrect estimation were 0.0±0.0%, 0.67±0.2%, 0.5±0.41%, and 1.17±0.26% (size : <1.5, <7, <15, and >15mm), respectively. Overall underestimation and overestimation probabilities are 1.17% and 1.17%, respectively. 

The methods proposed a robust emphysema clustering method, which could lead to new implication and progress of COPD

This method is especially useful in measuring size based emphysema analysis and could be possible to evaluate etiology and progress of COPD using 3D volumetric chest CT.

Lee, M, Kim, N, Seo, J, Oh, S, Lee, S, Lee, J, Oh, Y, Chang, Y, The Size-based Emphysema Quantification Using Length Scale Analysis in 3D Volumetric Chest CT.  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13027670.html