AJR Get Involved! Join ARRS Today
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Strub, W. M.
Right arrow Articles by Sun, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Strub, W. M.
Right arrow Articles by Sun, D.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?
Hotlight (NEW!)
Right arrow
What's Hotlight?
DOI:10.2214/AJR.07.2244
AJR 2007; 189:W115-W116
© American Roentgen Ray Society

Hybrid Reconstruction Kernel: Optimized Chest CT

William M. Strub, Kenneth L. Weiss and Dongmei Sun

University of Cincinnati, Cincinnati, OH 45267



 
WEB—This is a Web exclusive article.

Note—K. L. Weiss has a proprietary interest in the techonology described in this letter.

In thoracic imaging, in which imaging of both the lungs and the mediastinum needs to be optimized, two separate data sets with different reconstruction algorithms are often created. The higher resolution algorithms, such as bone and lung, preserve the higher spatial frequencies at the expense of greater noise. On the other hand, softer algorithms, such as soft tissue, reduce the higher frequency contribution, decreasing the noise and the spatial resolution [1]. We have begun testing a hybrid CT algorithm to simultaneously optimize lung and soft-tissue characterization to limit the number of images that need to be generated and stored.


Figure 1
View larger version (75K):
[in this window]
[in a new window]
[as a PowerPoint slide]
  		Fig. 1A —53-year-old woman with known metastatic colon carcinoma of the lung. Axial CT images of chest obtained with lung (A), standard (B), and hybrid (C) kernel reconstruction algorithms. Lung and hybrid kernels were thought to be equivalent with regard to their ability to display lung nodules.

 


Figure 2
View larger version (67K):
[in this window]
[in a new window]
[as a PowerPoint slide]
  		Fig. 1B —53-year-old woman with known metastatic colon carcinoma of the lung. Axial CT images of chest obtained with lung (A), standard (B), and hybrid (C) kernel reconstruction algorithms. Lung and hybrid kernels were thought to be equivalent with regard to their ability to display lung nodules.

 


Figure 3
View larger version (76K):
[in this window]
[in a new window]
[as a PowerPoint slide]
  		Fig. 1C —53-year-old woman with known metastatic colon carcinoma of the lung. Axial CT images of chest obtained with lung (A), standard (B), and hybrid (C) kernel reconstruction algorithms. Lung and hybrid kernels were thought to be equivalent with regard to their ability to display lung nodules.

 


Figure 4
View larger version (111K):
[in this window]
[in a new window]
[as a PowerPoint slide]
  		Fig. 2A —50-year-old woman with known liver metastasis who has undergone prior open biopsy of liver. Axial images of upper abdomen from chest CT obtained in standard (A), lung (B), and hybrid (C) kernel reconstruction algorithms. Standard and hybrid kernels were thought to be equivalent in their ability to display soft-tissue abnormalities.

 


Figure 5
View larger version (129K):
[in this window]
[in a new window]
[as a PowerPoint slide]
  		Fig. 2B —50-year-old woman with known liver metastasis who has undergone prior open biopsy of liver. Axial images of upper abdomen from chest CT obtained in standard (A), lung (B), and hybrid (C) kernel reconstruction algorithms. Standard and hybrid kernels were thought to be equivalent in their ability to display soft-tissue abnormalities.

 


Figure 6
View larger version (109K):
[in this window]
[in a new window]
[as a PowerPoint slide]
  		Fig. 2C —50-year-old woman with known liver metastasis who has undergone prior open biopsy of liver. Axial images of upper abdomen from chest CT obtained in standard (A), lung (B), and hybrid (C) kernel reconstruction algorithms. Standard and hybrid kernels were thought to be equivalent in their ability to display soft-tissue abnormalities.

 
All CT examinations are performed on a 16-MDCT HiSpeed scanner (GE Healthcare) with our routine unenhanced clinical protocol of 120 kVp, 300 mA, 5-mm collimation, 0.8-second rotation time, and 1.375 helical pitch. All scans are reconstructed with the lung and standard reconstruction algorithms. CT images generated with separate lung and standard kernels are retrospectively combined so that soft-tissue algorithm pixels less than -150 H or greater than 150 H are substituted with corresponding lung kernel reconstructed pixels. Hybrid images are subsequently generated in Matlab (Math Works) and then reimported into an e-Film Workstation 2.0 (Merge Technologies) for the radiologist to view along with conventional images. For each case, corresponding image sections are simultaneously viewed in manufacturer-preset settings for lung (window, 1,500 H; level, -600 H) and mediastinum (window, 350 H; level, 40 H) with the ability to independently adjust the window and level settings.

For the depiction of lung in all cases, the hybrid-reconstructed images were noted to be equivalent to the lung kernel reconstructions but superior to the soft-tissue kernel (Fig. 1A, 1B, 1C). For depiction of the mediastinal soft-tissue structures, the hybrid kernel was rated equivalent to the soft-tissue kernel but superior to the lung kernel (Fig. 2A, 2B, 2C).

CT reconstruction algorithms differ principally in the choice of the reconstruction kernel, providing freedom to design kernels that suppress or enhance specific ranges of spatial frequencies to affect the visual properties of the reconstructed images [2], with the ultimate choice of reconstruction kernel affecting performance for lesion-detection tasks [2]. Judy and Swensson [3] showed that the detectability of small high-contrast lesions improved as the reconstruction kernel became smoother. Prevrhal et al. [4] showed that the accuracy of evaluating thin structures improved with the use of high-resolution kernels.

To maximize the benefits of both reconstruction kernels, we developed the hybrid technique. Although our experience is brief, we found the hybrid kernel equivalent to both the lung reconstruction algorithm, with its ability to detect lung parenchymal abnormalities, and the standard reconstruction algorithm to evaluate the soft tissues. A hybrid lung reconstruction kernel is a promising technique that optimizes lung and soft-tissue evaluation while significantly reducing the number of images needed to be transmitted, stored, and reviewed.


References
Top
 References
 

  1. Boedeker KL, McNitt-Gray MF, Rogers SR, et al. Emphysema: effect of reconstruction algorithm on CT imaging measures. Radiology 2004;232 : 295-301[Abstract/Free Full Text]
  2. Armato SG, Altman MB, La Riviere PJ. Automated detection of lung nodules in CT scans: effect of image reconstruction algorithm. Med Phys 2003; 30:461 -472[CrossRef][Medline]
  3. Judy P, Swensson R. Detection of small focal lesions in CT images: effects of reconstruction filters and visual display windows. Br J Radiol 1985; 58:137 -145[Abstract/Free Full Text]
  4. Prevrhal S, Engelke K, Kalender W. Accuracy limits for the determination of the cortical width and density: the influence of object size and CT imaging parameters. Phys Med Biol1999; 44:751 -764[CrossRef][Medline]

Add to CiteULike CiteULike   Add to Complore Complore   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us   Add to Digg Digg   Add to Reddit Reddit   Add to Technorati Technorati    What's this?



This Article
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Strub, W. M.
Right arrow Articles by Sun, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Strub, W. M.
Right arrow Articles by Sun, D.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?
Hotlight (NEW!)
Right arrow
What's Hotlight?

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS