With the emergence of multiple-detector computed tomography (CT), which results in voxel isotropy, spatial resolution is not affected by reconstruction along different axes. Multiplanar reformation (MPR) of thin-slice images achieves clear demonstration of the anatomic positioning and variation in interlobar fissures, providing important diagnostic information for visualization of the location, steric, variation, and dysplasia of interlobar fissures.1,2
The entire lung can be captured by thin-slice images in one 64-detector CT scanning, allowing MPR, maximum intensity projection (MIP), volume rendering, and virtual endoscopy reconstruction. Using 64-detector spiral CT and reconstruction, we noted that interlobar fissures appeared as multiple, parallel-arranged linear opacities on axial MIP imaging under certain conditions: 5-line signs. It was also found that when the oblique reconstruction ran in parallel to a certain segment of the interlobar fissures, the fissures appeared as multiple neatly arranged linear opacities in resultant thin-slice MPR images. The objectives of this study were to introduce conditions for using MIP or MPR for analyzing fissures, to compare the difference between 2 types of images, and to explain the mechanism underlying the phenomenon of an interlobar fissure as a multiline shadow in MIP and MPR images based on published literature.
MATERIALS AND METHODS
This study was approved by the institutional review boards (Haikou People’s Hospital and Affiliated Hospital of Hainan Medical College) with a waiver of informed consent.
Fifty patients (32 males and 18 females aged 16–76 years) who presented with chest pain, chest tightness, or trauma and who had normal chest CT scan were included in the study.
A GE Lightspeed VCT scanner with an AW 4.3 workstation was used with the following scanning parameters: 120 kV; 89 to 301 mA (Auto mA); collimation width, 32 × 1.25 mm; rotation speed, 0.6 second/rotation; bed motion, 55 mm/rotation; pitch, 1.375; reconstructive slice thickness, 1.25 mm; interval, 1.25 mm; and slab thickness, 6.25 mm for axial MIP using a standard algorithm and a lung window width setting of 1000 Hounsfield units (HU) and a level setting of −700 HU. Sagittally, the interlobar fissure was averaged into upper, middle, and lower segments (the middle segment of the right interlobar fissure was taken from the minor fissure). The upper, middle, and lower segments of the right lung were coded RU, RM, and RL, respectively, whereas those of the left lung were coded LU, LG (lingula), and LL, respectively. Maximum intensity projection reconstruction was performed for each segment to allow further observation. When a subtle interlobar fissure was found, the MIP image was properly amplified on the workstation (AW, version 4.3).
MPR on Thin-Slice Images of Interlobar Fissures
On coronal or sagittal thin-slice images in the workstation, the oblique reconstruction was rotated so that it ran in parallel to a certain segment of the major or minor fissure (Fig. 1) to obtain an oblique-section MPR (OS-MPR) parallel to that segment of the major or minor fissure.
Measurement of the Transversal Diameter of Vessels on Axial MIP Imaging
In 6.25-mm axial MIP images, beaded (segmented) vessels were identified in both the upper hilar regions (at the opening of the anterior segments of both upper lobes), lower hilar regions (at the opening of the bronchia of the basal segments of both lower lobes), and middle hilar regions (between the upper and lower hilar regions). The transversal diameter was measured for segmented vessels with 4 or more segments.
Obtained data were statistically analyzed using the Statistical Package for the Social Sciences version 16.0 software package. A χ2 test was used to evaluate the differences in the improvement of linear clarity between sagittal OS-MPR and coronal OS-MPR, and using axial MIP plus OS-MPR, as a supplemental method, compared with single axial MIP; significance was tested at the level of α = 0.05.
With axial MIP reconstruction, normal interlobar fissures appeared as multiple parallel-arranged linear opacities, whereas the number of these linear opacities was equal to the ratio of the MIP slab thickness to the original slice thickness. For example, when the original slice thickness was 1.25 mm, and the MIP slab thickness was 6.25 mm, normal interlobar fissures appeared as 5 linear opacities, which resembled a musical staff in axial MIP images and was called a “5-line” sign (Fig. 2). With increasing MIP slab thickness, the number of the lines was increased, whereas the density was decreased until the lines vanished (Fig. 3).
Based on clarity, a 5-line sign of a normal interlobar fissure could be classified as clear (with 5 clear and well-defined lines), barely clear (the 5 lines were of low density and poor contrast with a blurry interspace but still identifiable), and unclear (a single linear opacity was observed in the original thin-slice image, but no multiple linear opacities were displayed in MIP images; a blurry band was noted instead (Fig. 4). Three hundred segments of interlobar fissures were reviewed. The results suggested that the 5-line sign was barely clear in most cases for different segments of interlobar fissures of the lungs. Clear or barely clear 5-line signs were observed in up to 93% of all cases. Right upper oblique fissure had the highest percentage of clear 5-line signs, whereas right middle interlobar fissures (right minor fissures) had the highest percentage of unclear 5-line signs, followed by left middle segment fissures. Table 1 shows the appearance of 5-line signs in axial MIP images for 50 individuals with normal pulmonary fissures.
When a 5-line sign was not clearly shown in axial MIP images, coronal or sagittal OS-MPR could be used supplementally. In these reconstructive images using both methods, interlobar fissures were also observed as multiple parallel-arranged linear opacities. Table 2 shows the results of coronal or sagittal OS-MPR for barely clear or unclear 5-line signs in axial MIP images.
As shown in Table 2, using sagittal OS-MPR, the unclear sign became clear (2 segments) or barely clear (12 segments), accounting for 66.67% (14/21). Although 7 segments remained unclear, among these, the middle and lower segments of the left oblique fissure were sagittally identified as dysplastic in 2 cases; except for these 4 segments, only 3 segments remained unclear, accounting for 14.29% (3/21) and representing 1% of all 300 reviewed segments.
Using coronal OS-MPR, only 19.05% (4/21) became clear (1 segment) or barely clear (3 segments). The improvement was poorer than that using sagittal OS-MPR (χ2 = 9.722, P = 0.002).
In addition, 56 segments with a barely clear 5-line sign in axial MIP images became clear, and 2 unclear segments became clear in sagittal OS-MPR images. Thus, using axial MIP plus sagittal OS-MPR, the percentage of clear linear opacities for interlobar fissures was remarkably improved compared with single axial MIP (122/300 vs 64/300, respectively; χ2 = 26.212; P <C; 0.001). Supplemental coronal OS-MPR barely achieved a clear sign (74/300 vs 64/300, respectively; P > 0.05).
With OS-MPR, interlobar fissures did not show a typical 5-line sign but multiple linear opacities. The number of lines depended on the length of the pleural section captured coronally or sagittally by the oblique reconstruction: The longer the section captured, the greater the number of lines observed. Additionally, multiple linear opacities can also be classified as clear, barely clear, and unclear. With a curved pleural section, thin-slice OS-MPR images mostly displayed missing lines. In such cases, using MIP reconstruction partially restored the missing lines with increased MIP slab thickness (Fig. 5). Unlike in axial MIP images, OS-MPR linear opacities showed no overlapping with adjacent structures, and the number of lines increased without affecting image clarity (Fig. 4). During OS-MPR + MIP imaging, other signs appeared as well (Figs. 6, 7).
The appearance of pulmonary vessels in axial MIP images for the 50 individuals with normal pulmonary images was also observed. It was found that most of the beaded vessels (72%–90%) were distributed in the peripheral region of the upper, middle, and lower segments of both sides of the lungs. The vessels were approximately 1.5 mm (range, 1.424–1.563 mm) in diameter. Vessels showed a higher density compared with that ofthe linear opacity of the fissure, and the interspace between beaded vessels was significantly narrower and more blurry compared with that between the lines in a 5-line sign (Fig. 8).
Maximum intensity projection was first used for observation of blood vessels. With the advantage of maximum display of the high-density linear vascular continuum in the lungs, MIP has proven to be an excellent approach to identify vessel section and intrapulmonary micronodules in thin-slice images.3,4 However, care must be taken to differentiate actual penvascular micronodules from the beaded appearance of MIP artifacts due to partial volume averaging. This phenomenon of beaded vessels, which was first described as “stair-step vessels” by Napel,5 results when smaller vessels, particularly those with dimensions that approximate the section width, pass obliquely through the scan plane because of partial volume averaging. In the present study, beaded vessels were mostly distributed in peripheral lungs in axial MIP images with a mean diameter of approximately 1.5 mm. This finding was generally consistent with that described previously.6,7 By analogy, if linear intrapulmonary structures were observed to be linearly arranged spots in MIP images, then planar intrapulmonary structures might be mostly displayed as linear opacities, such as interlobar fissures. The authors believed that multiple linear opacities and beaded vessels used the same mechanism and should be classified as the same type of artifact. Although normal interlobar fissures had a lower density compared with that in blood vessels, normal interlobar fissures showed clear interspaces between lines, a result that should facilitate observation. The interlobar fissures appeared as 5 linear opacities by axial MIP imaging, resembling a musical staff and were thus named 5-line signs.
Beaded vessels were best displayed when the diameters of the vessels were similar to the scanning slice thickness. In the current study, using an original slice thickness of 1.25 mm, both beaded vessels and multiple linear opacities of interlobar fissures were demonstrated. This result suggested that the slice thickness was similar to the width of the oblique section for an interlobar fissure. The width of the oblique section was related to the angle between the oblique and interlobar fissures. If the interlobar fissure ran horizontally and when the oblique and interlobar fissures formed an acute angle, the lower the acute angle was, the wider the linear opacity of the interlobar fissure was. Thus, linear opacities observed at the right horizontal fissure were mostly wide lines in the MIP images (Fig. 4). A slab thickness of 6.25 mm (5 × 1.25 mm) was chosen for the axial MIP reconstruction; compared with other conditions, these provided a better effect to indicate clear or barely clear signs for most segments of interlobar fissures with proper line thicknesses and clear interspaces. If an original slice thickness of 0.625 mm, a slice interval of 0.625 mm, and a slab thickness of 3.125 mm (5 × 0.625 mm) were chosen for the axial MIP reconstruction, most of the lines (except lines of right minor fissures) were thinner with blurry interspaces. Even with amplification of the workstation, images were poorer compared with those using a slice thickness of 1.25 mm. Therefore, the authors proposed that when the GE 64-row multiple-detector CT is used for MIP reconstruction, an original scanning slice thickness of 1.25 mm and an axial MIP slab thickness of 6.25 mm should be used as a general protocol to reveal 5-line signs for normal interlobar fissures.
Axial MIP images could reveal an entire segment of an interlobar fissure on a transversal section. According to the natural journey of the fissure segment, a 5-line sign might appear as different patterns, including a stave, waves, and a growth ring. In our sample, a typical 5-line sign was observed for 93% of interlobar fissures. Using the same reconstruction parameters, the same number of lines resulted. If the number of the lines was reduced, it might be due to a narrower section. With increased slab thickness, the number of the lines increased, and adjacent structures overlapped with them; linear opacities gradually became blurred and even vanished. The cause might be that the brightest pixel was chosen in the MIP reconstruction algorithm, leading to attenuation of low-density structures.3,5 Remy-Jardin suggested that an 8-mm MIP slab thickness was not suitable for visualization of low-density and poorly defined micronodules with a diameter of less than 3 mm and that a 5-mm MIP slab thickness was preferred for observation of such pulmonary micronodules with low density.3 They also reported that increased MIP slab thickness attenuated intrapulmonary microlesions. In addition to reconstruction thickness and interval, as well as slab thickness of the MIP, the window width, window level, and algorithm were also significant factors for observation of microstructures on MIP images.1,3 Furthermore, a suitable density and thickness of the interlobar fissure itself were required to produce a 5-line sign in axial MIP reconstruction.
Owing to orientations and variation in the anatomic location of pulmonary fissures, 5-line signs were not clearly visible in a few axial MIP reconstructive images, and MPR could be used as a supplementary method. Coronally or sagittally, when the oblique reconstruction ran in parallel to the interlobar fissure (180-degree angle), the fissure appeared as multiple stair-step lines on CT imaging, suggesting that multiple linear opacities could also be revealed directly on thin-slice oblique MPR images resembling 5-line signs found in axial MIP images. Because interlobar fissures were curved to a certain extent, OS-MPR could only capture a part of the pleura that ran in parallel to local interlobar fissures, and the number of resultant lines depended on the length of the interlobar fissures captured. Additionally, the shape and size of the captured local interlobar pleura varied according to the OS-MPR images, leading to the formation of a few 5-line signs. It is not suitable to use OS-MPR imaging before MIP imaging to observe 5-line signs. However, the advantage of OS-MPR imaging is that multiple linear opacities are not overlapped by adjacent structures, and the linear clarity is not affected by greater or fewer lines. As a supplemental method, the improvement of linear clarity using sagittal OS-MPR was much better than that by coronal OS-MPR.
Although a single and clear linear opacity of an interlobar fissure can be observed on thin-slice CT imaging, the variation in size, thickness, and homogeneity of a local planar pleural structure cannot be determined. These features could be extensively analyzed through 5-line signs shown in MIP images. Thus, 5-line signs could be used to differentiate accessory fissures and fissure hypoplasia from interstitial fibrosis and septal thickening, which are irregular and uneven in the lung.8–10
1. Takahashi K, Thompson B, Stanford W, et al. Visualization of normal pulmonary fissures on sagittal multiplanar reconstruction MDCT. AJR Am J Roentgenol
. 2006; 187: 389–397.
2. Ya-fei W, Shu-chun W. Visualization of variations in the pulmonary fissures on multi-planar reconstruction images of 64-slice spiral CT. Chinese Journal of Radiology
. 2008; 42: 60–63.
3. Remy-Jardin M, Remy J, Artaud D, et al. Diffuse infiltrative lung disease: clinical value of sliding-thin-slab maximum intensity projection CT scans in the detection of mild micronodular patterns. Radiology
. 1996; 200: 333–339.
4. Sakai M, Murayama S, Gibo M, et al. Can maximum intensity projection images with multidetector-row computed tomography help to differentiate between the micronodular distribution of focal and diffuse infiltrative lung diseases? J Comput Assist Tomogr
. 2005; 29: 588–591.
5. Napel S, Rubin GD, Jeffrey RB Jr. STS-MIP: a new reconstruction technique for CT of the chest. J Comput Assist Tomogr
. 1993; 17: 832–838.
6. Bhalla M, Naidich DP, McGuinness G, et al. Diffuse lung disease: assessment with helical CT--preliminary observations of the role of maximum and minimum intensity projection images. Radiology
. 1996; 200: 341–347.
7. Napel SA. Principles and techniques of 3D spiral CT angiography. In: , eds. Spiral CT
. New York, NY: Raven Press; 1995: 167–196.
8. Hansell DM, Bankier AA, MacMahon H, et al. Fleischner Society: glossary of terms for thoracic imaging. Radiology
. 2008; 246: 697–722.
9. Simon G. Principals of Chest X-ray Diagnosis
. 4th ed. London, UK: Butterworths; 1978.
10. Hill CA. “Tail” signs associated with pulmonary lesions: critical reappraisal. AJR Am J Roentgenol
. 1982; 139: 311–316.
Keywords:© 2014 by Lippincott Williams & Wilkins
pleura; image processing, computer-assisted; image reconstruction; computed tomography; x-ray