| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Innovation |
1
Department of Radiology, University Hospital Graz, Auenbruggerplatz 9, A-8036
Graz, Austria.
2
Institute of Medical Informatics, Statistics and Documentation, University
Hospital Graz, A-8036 Graz, Austria.
3
Department of Otorhinolaryngology, University Hospital Graz, A-8036 Graz,
Austria.
Received April 30, 2001;
accepted after revision July 18, 2001.
Address correspondence to R. Groell.
Introduction
 Top Introduction Materials and MethodsResultsDiscussionReferences |
|---|
For exact staging of tumors in these patients and to determine possible therapeutic regimes, in particular potential resectability, assessment of the tumor-to-vessel relationship is crucial. Therefore, images are desired on which optimal tumor delineation is combined with sufficient vessel opacification. Harris et al. [1] have recommended administration of a second bolus of contrast material during the delayed scanning to achieve this purpose. However, this technique would substantially increase the total amount of contrast material—and, therefore, the costs of the CT examination. In this study of patients with squamous cell carcinomas of the head and neck, we tried to achieve simultaneous enhancement of cervical vessels and tumors without a second bolus of contrast material by using digital image fusion of early and delayed CT scans.
|
TopIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CT was performed using one of two CT scanners in our department (Somatom Plus 4, Siemens, Erlangen, Germany; or LightSpeed Qx/i, General Electric Medical Systems, Milwaukee, WI). All studies were performed in the helical mode with a slice thickness of 3 mm (n = 2, Siemens CT scanner) or 3.75 mm (n = 6, General Electric CT scanner), using a pitch factor of 1.5; all images were reconstructed every 3 mm or 3.75 mm, respectively. The scanning volume covered the area from the skull base to the thoracic inlet in the cephalocaudal direction using axial slices. All patients were examined in the supine position.
Each patient received a total of 100 mL of non-ionic contrast material (Ultravist, iopromide 300 mg I/mL; Schering, Berlin, Germany) at a uniphasic flow rate of 2 mL/sec, administered by power injector (MCT; Medrad, Pittsburgh, PA) through an IV cannula in an antecubital vein. In all patients, the early scans started 30 sec after the injection was begun. The CT examinations were performed according to one of two CT protocols that were routinely used in our department: the delayed phase started either 180 sec (n = 7) or 300 sec (n = 1) after commencement of injection, using the same image acquisition parameters as were used for the early scans. No additional contrast material was administered for the delayed scans. During scanning, the patients were instructed to breathe normally but not to swallow.
The images were reviewed on a digital image workstation (Sienet Magic View; Siemens) and transferred to a PC connected to the local picture archiving and communication system. In those regions of interest that showed the tumor, the corresponding image pairs were manually selected at the PC for further image fusion.
The software for image fusion was developed with the interactive data language software tool (IDL 5.2; Research Systems, Boulder, CO). In a first step, the vessels and bones were extracted from the early scans using a single threshold algorithm. These thresholded images were digitally fused with the delayed scans (Fig. 1). Image fusion was realized by following logical operation: if the attenuation value of a pixel in the early scan was above the threshold, then this pixel value was inserted into the fused image; if the attenuation value of a pixel in the early scan was below the threshold, then the corresponding attenuation value of the delayed image was inserted. If the early images were not thresholded, but simply added to the delayed images, then the lower tumor-to-background contrast of the early image would also contribute to the fused image.
|
Patient motion resulted in parallel shifting of bony and soft-tissue structures between the early and delayed scans. Such possible patient motion between the early and delayed scans was recognized by shifting of bony structures that were clearly visible on the thresholded and on the delayed scans. This shifting was corrected manually by allowing a linear pixel shifting in the x- or y-axis, respectively. (Automatic adjustment of the images was possible.) As the bones were aligned, other structures, such as tumors and vessels, were aligned simultaneously. No notable motion occurred in the z-axis.
Two radiologists reviewed the images independently. On both early and delayed scans, as well as on the fused images, the quality of vessel opacification and tumor delineation was determined separately according to a three-step scoring system (1 = good, 2 = medium, 3 = poor quality). The results were compared using a paired Student's t test.
|
TopIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
TopIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
When fusion is performed between different imaging modalities, several technical and methodological problems must be solved. Parameters such as image format, image resolution, field of view, pixel size, zooming factors, or patient orientation usually vary between the modalities. These factors must be aligned before fusion. Such alignment can be performed, in part, automatically, but some parameters require manual, or at least partially manual, adaptation, in particular, the alignment of different body positions and orientations in all three body axes [8]. Therefore, topographic landmarks are necessary that are visible on both modalities; these landmarks may be difficult to define, especially in radionuclide studies. In general, such alignment is time-consuming and bears the risk of wrong alignment.
In this study, CT scans that were obtained consecutively using identical image acquisition and reconstruction parameters were fused. The only factor that had to be accounted for was possible body movement during the time span between the early and the delayed scans (180-300 sec). However, this movement proved to be minimal and its effects were corrected by linear shifting in the x- and y-axes of less than 3 mm in each patient. No important motion occurred in the z-axis. The bony structures were used as the topographic landmarks and were clearly visible on both early and delayed scans.
Image fusion of the early and delayed scans resulted in simultaneous opacification of vessels and tumors in all patients. The results of the image scoring confirmed the synergetic effects of image fusion. In general, image fusion enabled optimal assessment of the tumor-to-vessel relationships in all of the patients.
Radiologists are trained to merge different images mentally, either to compare different modalities or to compare follow-up studies of the same modality. However, such mental image fusion may be difficult for physicians who are not trained on digital images but who still need the information of the tumor-to-vessel relationship to plan the therapeutic regimen: for them, such fused images might be particularly helpful.
Our study was performed with a selected patient group in whom we expected synergetic effects from image fusion. This is certainly not the case in every patient who undergoes multiphase CT examination. However, this study shows that image fusion of different phases of helical CT examinations is easy to perform and may be of value to physicians reviewing the studies.
|
TopIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  K. Lemola, M. Sneider, B. Desjardins, I. Case, J. Han, E. Good, K. Tamirisa, A. Tsemo, A. Chugh, F. Bogun, et al. Computed Tomographic Analysis of the Anatomy of the Left Atrium and the Esophagus: Implications for Left Atrial Catheter Ablation Circulation, December 14, 2004; 110(24): 3655 - 3660. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
