1887
banner image
No data available.
Please log in to see this content.
You have no subscription access to this content.
No metrics data to plot.
The attempt to load metrics for this article has failed.
The attempt to plot a graph for these metrics has failed.
CT image construction of a totally deflated lung using deformable model extrapolation
Rent:
Rent this article for
USD
10.1118/1.3531985
    + View Affiliations - Hide Affiliations
    Affiliations:
    1 Department of Electrical and Computer Engineering, The University of Western Ontario, London, Ontario N6A 5B9, Canada; Imaging Research Laboratories, Robarts Research Institute (RRI), London, Ontario N6A 5K8, Canada; and Lawson Health Research Institute (LHRI), London, Ontario N6A 4V2, Canada
    2 Department of Medical Biophysics, The University of Western Ontario, London, Ontario N6A 5C1, Canada; Imaging Research Laboratories, Robarts Research Institute (RRI), London, Ontario N6A 5K8, Canada; and Lawson Health Research Institute (LHRI), London, Ontario N6A 4V2, Canada
    3 Imaging Research Laboratories, Robarts Research Institute (RRI), London, Ontario N6A 5K8, Canada; Lawson Health Research Institute (LHRI), London, Ontario N6A 4V2, Canada; Departments of Medical Biophysics, The University of Western Ontario, London, Ontario N6A 5C1, Canada; Departments of Diagnostic Radiology and Nuclear Medicine, The University of Western Ontario, London, Ontario N6A 3K7, Canada
    4 Department of Electrical and Computer Engineering, The University of Western Ontario, London, Ontario N6A 5B9, Canada; Department of Surgery, The University of Western Ontario, London, Ontario N6A 3K7, Canada; and Lawson Health Research Institute (LHRI), London, Ontario N6A 4V2, Canada
    5 Department of Medical Biophysics, The University of Western Ontario, London, Ontario N6A 5C1, Canada; Department of Electrical and Computer Engineering, The University of Western Ontario, London, Ontario N6A 5B9, Canada; and Imaging Research Laboratories, Robarts Research Institute (RRI), London, Ontario N6A 5K8, Canada
    a) Electronic mail: asamani@uwo.ca
    Med. Phys. 38, 872 (2011); http://dx.doi.org/10.1118/1.3531985
/content/aapm/journal/medphys/38/2/10.1118/1.3531985
http://aip.metastore.ingenta.com/content/aapm/journal/medphys/38/2/10.1118/1.3531985
View: Figures

Figures

Image of FIG. 1.
FIG. 1.

A typical combined sequence histogram for a respiratory sequence consisting of three CT images acquired at different phases of respiration. The figure has been zoomed in to focus on the region of interest within the original combined histogram; the convergence points are indicated by arrows.

Image of FIG. 2.
FIG. 2.

CT image construction pipeline. Input: preoperative images; the right branch performs registration: first and second blocks find the deformable registration’s global and local transformation parameters to register a lung reference image with each preoperative lung image; the branch underneath estimates the lung’s air volume: first block finds the optimum lower and upper thresholds to segment the lung’s air content, while the second block uses these thresholds to estimate the lung’s air volume in each image. The extrapolation block extrapolates these parameters and passes them to the next block where the required image is constructed using the extrapolated parameters and the lung’s CT image acquired at the reference volume. Output: Lung image in a totally deflated state.

Image of FIG. 3.
FIG. 3.

[(a)–(t)] Preoperative free-breathing 4D-CT image sequence of a lung slice passing through a tumor acquired while the lung was being respired continuously; w: the same lung slice acquired while the lung was totally deflated (required).

Image of FIG. 4.
FIG. 4.

Combined sequence histogram for the 20 CT images of a 4D-CT respiratory sequence acquired while the lung was being respired continuously. The region of interest (black rectangle) has been zoomed in; optimal lower and upper segmentation thresholds are indicated by two arrows.

Image of FIG. 5.
FIG. 5.

[(a)–(t)] One lung image slice passing through a tumor from the respiratory 4D-CT image sequence acquired while the lung was being respired continuously; the air inside the lung was segmented using the determined optimal lower and upper threshold values. The bright and dark regions show the air and soft tissue with the background, respectively. The estimated air volumes calculated based on the segmentations are (a) 920 ml, (b) 826 ml, (c) 782 ml, (d) 714 ml, (e) 642 ml, (f) 573 ml, (g) 501 ml, (h) 441 ml, (i) 436 ml, (j) 359 ml, (k) 347 ml, (l) 375 ml, (m) 445 ml, (n) 532 ml, (o) 728 ml, (p) 802 ml, (q) 878 ml, (r) 949 ml, (s) 982 ml, and (t) 989 ml, respectively.

Image of FIG. 6.
FIG. 6.

(a) One difference lung image slice passing through a tumor. It shows the difference between the CT image acquired at the least and most inflated phases in the respiratory image sequence. The tumors and the lung boundaries from the two images are clearly visible in the difference image. (b) The same difference image slice after the deformable registration. The anatomy and the tumors coincide reasonably accurately. The coincided tumor locations are pointed by an arrow.

Image of FIG. 7.
FIG. 7.

Linear logarithmic curve (solid line) fitted to available values of one of the registration parameters (shown by “”) obtained by registering the reference image with images within the respiratory CT sequence. The curve is extrapolated (dashed line) in order to estimate the required parameter value (“”) to construct the CT image of the lung in its totally deflated state. The shown linear variation is actually nonlinear due to using a logarithmic scale. This function empowers the FFD registration technique to capture highly complex deformations expected in the lung because several thousands of parameters are typically involved.

Image of FIG. 8.
FIG. 8.

One tumor passing slice of the (a): lung’s reference CT image (989 ml), (b) totally deflated lung CT image's (required), (d): constructed CT image of the deflated lung; the center of the tumor and anatomical feature physical coordinates is given in parenthesis. The physical coordinate of the anatomical feature in the reference image is (207.9, 100.7, 104.8) mm. (c) Extrapolated deformation field (arrows) fused with the two overlaid CT images of the lung in its most inflated (reference) and totally deflated (required) states; the locations of the tumor in the overlaid images are highlighted with circles; the extrapolated arrows clearly point to the directions of local deformations used to construct the CT image of the lung in its totally deflated state based on the reference image. (e) Difference image of the lung’s reference CT image and the totally deflated lung’s CT image (required) and (f) difference image of the totally deflated lung’s constructed CT image and the totally deflated lung’s original CT image (required).

Image of FIG. 9.
FIG. 9.

Effects of using a subset of 20 preoperative 4D-CT images in the CT image construction process with figures (a), (b), (c), and (d) showing using 2, 3, 6, and 10 preoperative images, respectively. The center of the tumor physical coordinates is given in parenthesis. The original CT image of the totally deflated lung (required) is shown in Fig. 8(b).

Loading

Article metrics loading...

/content/aapm/journal/medphys/38/2/10.1118/1.3531985
2011-01-21
2014-04-19
Loading

Full text loading...

This is a required field
Please enter a valid email address
752b84549af89a08dbdd7fdb8b9568b5 journal.articlezxybnytfddd
Scitation: CT image construction of a totally deflated lung using deformable model extrapolation
http://aip.metastore.ingenta.com/content/aapm/journal/medphys/38/2/10.1118/1.3531985
10.1118/1.3531985
SEARCH_EXPAND_ITEM