Visualization of Normal Pulmonary Fissures on Sagittal Multiplanar Reconstruction MDCT

Transcription

1 Normal Pulmonary Fissures on MDCT Chest Imaging Original Research C D E M N E U T R Y L I M C I G O F I N G Koji Takahashi 1,2 rad Thompson 2 William Stanford 2 Yutaka Sato 2 Kenichi Nagasawa 1 Hiroaki Sato 1 Makoto Kubota 1 yako Kashiba 1 Hiroyuki Sugimori 1 Takahashi K, Thompson, Stanford W, et al. Keywords: chest, lung, lung cancer, MDCT, pleura, pulmonary nodules DOI: /JR Received January 28, 2005; accepted after revision March 23, Department of Radiology, sahikawa Medical College and Hospital, Midorigaowa-higashi, sahikawa, Japan. ddress correspondence to K. Takahashi (taka1019@asahikawa-med.ac.jp). 2 Department of Radiology, University of Iowa College of Medicine, Iowa City, I JR 2006; 187: X/06/ merican Roentgen Ray Society Visualization of Normal Pulmonary Fissures on Sagittal Multiplanar Reconstruction MDCT OJECTIVE. Delineation of the interlobar fissures on multiplanar reconstruction (MPR) images is useful to assess masses at the fissures for invasion into adjacent lobes. We performed this study to determine the appropriate MDCT protocol to visualize the interlobar fissures on sagittal MPR images. MTERILS ND METHODS. For the phantom studies, radiographic film was used to replicate the interlobar fissures. For the clinical studies, we obtained MDCT scans of 130 patients with normal interlobar fissures. Visualization of the interlobar fissures on sagittal MPR was assessed using the following scanning parameters: scan collimations of 0.5, 1, 2, and 3 mm with helical pitches of 1 and 1.5 for the phantom studies; and scan collimations of 0.5, 1, 2, and 3 mm with a helical pitch of 1.5 and a scan collimation of 2 mm with a helical pitch of 1 for the clinical studies. RESULTS. To visualize fissures as a sharp line, a 0.5- or 1-mm collimation was required for the major fissure and 0.5 mm for the minor fissure in the phantom studies. In the clinical studies, 0.5-mm-collimation MPR images depicted interlobar fissures as a sharp line in all cases. Fissures on MPR images using 1-, 2-, and 3-mm collimations appeared as a sharp line in %, %, and 0% of cases, respectively. CONCLUSION. Volume data obtained using a 1-mm collimation are required to visualize all the interlobar fissures as a sharp line on sagittal MPR images except the minor fissure and superior portion of the right major fissure, for which a 0.5-mm collimation is required. wo-dimensional multiplanar reconstruction (MPR) is the most T frequently used reconstruction technique for helical CT scans. The benefits of using MPR have been reported in the assessment of thoracic diseases, including central airway lesions [1, 2], central pulmonary embolism [3], mediastinal lymphadenopathy [4], and diaphragmatic rupture [5, 6]. Reports indicate that high-resolution MPR images obtained with a 0.5- or 1-mm collimation depict normal anatomy and lung disease in a manner similar to axial or direct coronal high-resolution CT images [7 9]. The interlobar fissure is an important landmark for assessing the location and distribution of pulmonary disease. In patients with lung tumors, assessment of tumor extension across the interlobar fissure is critical in staging the disease, assessing the feasibility of surgical resection, and deciding whether lobectomy or pneumonectomy will be required [10]. Quint et al. [10] reported poor sensitivities in detecting transfissural tumor extension on contiguous axial CT images obtained with a 10-mm slice thickness. However, with a collimation of less than 2 mm, the interlobar fissures appear as a hyperattenuating line [11, 12], and some investigators have indicated that the use of thin-section CT scans could improve accuracy in identifying tumor extension across the major fissure [10, 11, 13]. For identification of the interlobar fissures, we postulated that using sagittal images is more accurate than using axial images to assess the major fissures along their entire craniocaudal length. We believe the same hypothesis to be true for identification of the minor fissure because it lies perpendicular to the longitudinal axis of the thorax. However, researchers have not previously reported, to our knowledge, the appropriate MDCT protocol with which to visualize the interlobar fissures on sagittal MPR images. Materials and Methods Phantom Studies Initially, a phantom study was performed to assess the visibility of interlobar fissures on sagittal MPR images. To replicate the interlobar fis- JR:187, ugust

2 sures, we used radiographic film (200 µm in thickness) tilted 30 to the z-axis to represent the major fissure and placed another film nearly perpendicular to the scanning z-axis to represent the minor fissure. The phantom images were obtained on a 4-MDCT scanner (quilion, Toshiba) using the following protocol: 0.5-mm collimation with a 0.5-mm reconstruction interval; 1- mm collimation with a 0.5-mm reconstruction interval; 2-mm collimation with a 1-mm reconstruction interval; and 3-mm collimation with a 1.5-mm reconstruction interval. To evaluate the visibility of the major fissure, scans were obtained at pitches of 1 and 1.5, a 0.5- second gantry rotation, 120 kv and 50 m, a highfrequency algorithm, a 30-cm field of view, and a matrix. For assessment of the minor fissure, the same parameters were used except a 15- cm field of view and matrix. Once scanning had been performed, sagittal MPR images with a 1.0-mm slice thickness were generated from the scan data. The MPR images were photographed separately on hard copies with a window width of 1,600 H and a window level of 600 H. C D E F G H Fig. 1 Sagittal multiplanar reconstruction images of phantom major fissure. D, Phantom major fissure is visualized as sharp line with no stairstep artifact on images obtained using 0.5-mm collimation ( and ) and 1-mm collimation (C and D) and pitch of 1 or 1.5. E and F, Phantom major fissure is visualized as slightly thick line with no and mild stairstep artifact on images obtained using 2-mm collimation at pitch of 1 (E) and 1.5 (F). G and H, Phantom major fissure is visualized as significantly thick line with mild and severe stairstep artifact on images obtained using 3-mm collimation at pitch of 1 (G) and 1.5 (H). 390 JR:187, ugust 2006

3 Normal Pulmonary Fissures on MDCT C E G Fig. 2 Sagittal multiplanar reconstruction (MPR) images of phantom minor fissure. On MPR images with 1-, 2-, and 3-mm collimation, stairstep artifact was mild at pitch of 1 and severe at pitch of 1.5. and, Phantom minor fissure is visualized as thin line with no and mild stairstep artifact on sagittal images using 0.5-mm collimation at pitch of 1 () and 1.5 (). C and D, Phantom minor fissure is visualized as slightly thick line on images obtained using 1-mm collimation at pitch of 1 (C) and 1.5 (D). E H, Phantom minor fissure is visualized as significantly thick line on images obtained with 2-mm collimation at ptich of 1 (E) and 1.5 (F) and 3-mm collimation at pitch of 1 (G) and 1.5 (H). We assessed the thickness of the visualized phantom fissures and the presence and degree of stairstep artifact on these images using the following criteria: thickness was graded as a thin line, a slightly thick line, or a significantly thick line; and stairstep artifact was graded as none, mild, or severe. The thickness of the fissures was graded in comparison with the phantom fissure on conventional axial images obtained using a 0.5-mm collimation; the grades were defined as follows: thin when the thickness of the fissures was equal to that on the axial phantom images, slightly thick when it was slightly thicker than on the axial phantom images, and significantly thick when it was considerably thicker than on the axial phantom images. The images were initially assessed independently by two pulmonary radiologists, and the final decision was made by consensus in cases for which there was interobserver disagreement. Clinical Studies The study was approved by the institutional review board, and informed consent was not required or obtained to review the CT studies. To determine the appropriate MDCT scanning parameters with which to visualize normal interlobar fissures on sagittal MPR images, we performed a retrospective review of CT studies of 130 patients obtained for clinical reasons during the period from January 2003 to June We used a 4-MDCT scanner (quilion, Toshiba) and obtained sagittal MPR images of the interlobar fissures in five groups of subjects using different scan collimations and helical pitches: 0.5-mm collimation with a 0.5-mm reconstruction interval at a pitch of 1.5 in 20 patients, 1-mm collimation with a 0.5-mm reconstruction interval at a pitch of 1.5 in 20 patients, 2-mm collimation with a 1-mm reconstruction interval at a pitch of 1 in 30 patients, 2-mm collimation with a 1-mm re- D F H JR:187, ugust

4 construction interval at a pitch of 1.5 in 30 patients, and 3-mm collimation with a 1.5-mm reconstruction interval at a pitch of 1.5 in 30 patients. Images were obtained using 120 kv and 150 ms, 0.5-second gantry rotation, 30-cm field of view, matrix, and a high-frequency reconstruction algorithm. Sagittal MPR images with a 3-, 2-, and 1-mm collimation were obtained from clinically indicated standard chest CT studies and those with a 0.5-mm collimation, from patients in whom high-resolution CT was clinically indicated. We excluded patients with incomplete fissures; diffuse lung disease; focal lesions adjacent to the interlobar fissures; pleural effusions; pleural abnormalities; cardiomegaly; and thoracic deformity, including scoliosis and pectus excavatum. The patient population included 74 men and 56 women who ranged in age from 34 to 71 years (mean, 48 years). There was no significant difference in sex and age among the five groups with different scan collimations and helical pitches. We obtained bilateral sagittal MPR images of the interlobar fissures of 1-mm thickness with 10-mm intervals and photographed both on hard copies with a window width of 1,600 H and a window level of 600 H. We assessed the visibility of the interlobar fissures on sagittal MPR images using the following criteria: sharp line, blurred line, and inadequate visualization. We judged the fissure to be a sharp line when it appeared similar to that on conventional axial images with a 1-mm collimation in 10 healthy subjects as a control group. We judged the fissure to be blurred when it appeared as a thick line with marginal blurring or when there was a stairstep artifact, and we judged the fissure to be inadequately visualized when it appeared as a hypo- or hyperattenuating band devoid of vascularity. We reviewed sagittal MPR images at the midportion of the thorax on the right side and mm lateral to the cardiac margin on the left. ecause the angle of the major fissure along its longitudinal axis differs between levels above and below the hilum, we assessed its conspicuity at regions both superior and inferior to the hila. We also assessed the visibility of the right minor fissure. ll images were independently reviewed by two pulmonary radiologists who were blinded to the collimation and pitch of the images. Interobserver variation of the score between the two reviewers was assessed using kappa statistics. Kappa values were defined as the following: , poor; , fair; , moderate; , good; and , excellent. Results Phantom Studies The phantom major fissure was visualized as a thin line with no stairstep artifact on sagittal MPR images obtained using 0.5- and 1- mm collimation at pitches of 1 and 1.5; as a slightly thick line with no and mild stairstep artifact on MPR images using a 2-mm collimation at pitches of 1 and 1.5, respectively; and as a significantly thick line with mild and severe stairstep artifact on MPR images obtained using a 3-mm collimation at pitches of 1 and 1.5, respectively (Fig. 1). The phantom minor fissure was visualized as a thin line with no stairstep artifact on the sagittal MPR images using a 0.5-mm collimation at pitches of 1 and 1.5. It was visualized as a slightly thick line on MPR images obtained with a 1-mm collimation and as a significantly thick line on images obtained with 2- and 3-mm collimation. Stairstep artifact on MPR images obtained using a 1-, 2-, and 3- mm collimation was mild at pitch of 1 and severe at pitch of 1.5 (Fig. 2). Clinical Studies Interobserver agreement in the quality criteria of the interlobar fissures on MPR images was excellent in all portions of the interlobar fissures: κ = 0.96 in the right superior major fissure, κ = 0.90 in the right inferior major fissure, κ = 0.88 in the left superior major fissure, κ = 0.90 in the left inferior fissure, and κ = 0.86 in the minor fissure. The mean percentages of the two reviewers for each criterion that is, sharp line, blurred line, and inadequate visualization in the assessment of the right superior and inferior major fissures, the left superior and inferior major fissures, and the minor fissure are shown in Table 1. Discussion For visualization of the interlobar fissures on axial CT images, the most significant factors affecting their appearance and conspicuity are partial volume averaging from the adjacent lung tissue; the thinness of the fissures; and their anatomic course, which is oblique to axial scanning planes. To overcome problems with volume averaging and to visualize the interlobar fissures as a hyperattenuating line, a thin collimation (generally < 2 mm) is required on conventional axial scanning [11, 12]. However, to assess the entire interlobar fissures, the use of thin-collimation scans results in an excessive number of axial images. lternatively, a 25 cranially tilted axial scan can provide the same benefit in visualization of the major fissure, which appears as a linear or bandlike density, and can do so even with a 5- to 10- mm collimation [13]. However, this method TLE 1: Visualization of the Normal Pulmonary Fissures on Sagittal Multiplanar Reconstruction Images Variable Right Superior Major Fissure Right Inferior Major Fissure Left Superior Major Fissure Left Inferior Major Fissure Minor Fissure Imaging parameter Collimation (mm) Pitch ppearance of fissure Sharp lurred Inadequate JR:187, ugust 2006

5 Normal Pulmonary Fissures on MDCT Fig. 3 Sagittal multiplanar reconstruction images obtained with each of scanning parameters show typical appearance of interlobar fissures in five patients. and, Upper and lower portions of right major fissure and minor fissure are visualized as sharp line on images obtained using 0.5-mm collimation in 57-year-old woman () and 1-mm collimation in 62-year-old man () at pitch of 1.5. C and D, Upper and lower portions of right major fissure and minor fissure are visualized as blurred line (arrows) on images obtained using 2-mm collimation at pitch of 1 in 55-year-old man (C) and at pitch of 1.5 in 68-year-old woman (D). (Fig. 3 continues on next page) C is generally impractical and undesirable for routine scanning. With the recent advent of MDCT and especially with the increases in the number of data acquisition channels and with the thinner detector collimation, MDCT can now readily provide high-resolution MPR images of the lung from nearly isovoxel volume data. Honda et al. [7] reported that coronal MPR images obtained using a 0.5-mm collimation can provide image quality similar to that of direct coronal thin-section CT scans for the evaluation of normal pulmonary structures, including the interlobar fissures, on autopsy lung specimens. D JR:187, ugust

6 Fig. 3 (continued) Sagittal multiplanar reconstruction images obtained with each of scanning parameters show typical appearance of interlobar fissures in five patients. E, Upper and lower portions of right major fissure and minor fissure in 51-year-old man are visualized as hyperattenuating band (arrows) on image obtained using 3-mm collimation at pitch of 1.5. Recently, the authors of two studies reported that coronal MPR images obtained from helical scan data with a thin collimation (0.5 1 mm) and axial high-resolution CT images were comparable for the evaluation of lung parenchymal abnormalities [8, 9]. In both studies, the authors indicated that MPR images depict the features and distribution of parenchymal abnormalities as accurately as axial high-resolution CT images. rakawa et al. [8] reported that MPR images provided additional information to that provided by axial high-resolution images in 22% of cases, especially for the presence and distribution of ground-glass opacities, intralobular reticular opacities, and bronchial dilatation. Remy-Jardin et al. [9] suggested that the significantly reduced numbers of images provide additional benefit as a diagnostic algorithm when using MPR. On the basis of these studies, we thought that sagittal MPR images obtained from thin-collimation volume data would be suitable for assessment of the interlobar fissures (Figs. 3 and 4). Storto et al. [14] assessed the value of MPR images for the detection of neoplastic extension across the major and minor fissures. They obtained MPR images from single-detector helical scans with a 2-mm collimation and pitches of 1 and 1.5. In their study, the visibility of the interlobar fissures on MPR images was good, represented as a sharp hyperattenuating line in 78.4% of patients; acceptable, as a bandlike opacity with marginal blurring, in 13.7%; and poor, as insufficient visualization for diagnosis, in 7.8%. They also assessed transfissural extension of tumors on MPR images using the following two criteria for potential fissural neoplastic invasion: the presence of a mass on both sides of the fissure and neoplastic lobulations extending across the fissural plane. They reported that the two criteria yielded a high diagnostic accuracy with a sensitivity and specificity of 100% and 92.8%, and 100% and 100% for the major and minor fissures, respectively [14]. However, the interlobar fissures on MPR images appeared indistinct with marginal blurring and were considered similar to the blurred line group from our criteria. Furthermore, the number of subjects in their analysis was small: seven cases with invasion and 14 cases without invasion in the assessment of the major fissure and eight with invasion and four without invasion in that of the minor fissure. We agree with their statement that sagittal MPR images are useful in evaluating a lung mass at the fissures for invasion into adjacent lobes. However, we believe that MPR images obtained using a 2-mm collimation would be suboptimal in detecting subtle changes reflecting transfissural tumor extension and that the use of a thinner collimation would provide a better morphologic representation of the interlobar fissures on MPR images and thereby augment the assessment of abnormalities of the fissure, adjacent lung, or both. In clinical studies, a 0.5-mm-collimation scan revealed both the major and minor fissures as a sharp line in all cases. 1-mm-collimation scan depicted fissures as a sharp line in more than 90% of cases except for the superior portion of the right major fissure and the minor fissure, for which only a sharp line was seen in 82.5% and 77.5% of the cases, respectively. This difference is thought to be due to the more horizontal orientation of these fissures, which results in significant partial volume averaging and stairstep artifact. MPR images obtained using 2- and 3-mm collimation were considered to be suboptimal for visualizing the fissures because of marginal blurring and frequent nonvisualization as a line, respectively. Limitations in our data are the lack of clinical study of the diagnostic accuracy of sagittal MPR images for assessing abnormalities of the interlobar fissures, including transfissural tumor extension. However, we have found that clear visualization of the interlobar fissures on high-resolution MPR images is beneficial in the evaluation of adjacent lesions (Figs. 5 and 6). nother clinical study is ongoing in an attempt to clarify the benefit of high-resolution MPR images. n additional limitation is that our data were obtained using a 4-MDCT scanner. Currently, 64-MDCT scanners are in use commercially. With this increase in the number of slices, CT images may deteriorate due to cone beam artifact, which is not compensated for by MDCT reconstruc- E 394 JR:187, ugust 2006

7 Normal Pulmonary Fissures on MDCT Fig. 4 Poor visualization of upper major fissure and minor fissure on sagittal multiplanar reconstruction images in three patients., Image obtained using 1-mm collimation at pitch of 1.5 shows that lower portion of right major fissure in 60-year-old man appears as sharp line, whereas its upper portion and minor fissure appear as blurred line (arrow)., Image obtained using 2-mm collimation at pitch of 1 shows that lower portion of right major fissure in 48-year-old woman appears as sharp line, whereas its upper portion and minor fissure appear as blurred line (arrows). C, Image obtained using 2-mm collimation at pitch of 1.5 shows that lower portion of right major fissure in 66-year-old man appears as blurred line, whereas its upper portion and minor fissure appear as hypoattenuating band (arrows). tion algorithms in 4-MDCT scanner systems that use relatively narrow cone beam angles. However, cone beam artifact can be minimized using a modified volumetric CT reconstruction technique in MDCT scanners with more than 16 tracks [15]. In future configurations, interpolation algorithms for MDCT will likely use z-axis deconvolution to further improve longitudinal spatial resolution [16]. In studies using MDCT scanners, the detector collimation should be selected on the basis of the coverage desired and taking into account the thinnest slice thickness needed for retrospective reconstruction. We think our data are useful in determining a scanning protocol for visualization of the interlobar fissures on sagittal MPR images. C JR:187, ugust

8 Fig year-old man with lung cancer in superior segment of right lower lobe., Sagittal multiplanar reconstruction (MPR) image obtained using 0.5-mm collimation at pitch of 1.5 shows tumor abutting and retracting major fissure but not involving minor fissure., MPR image obtained using 2-mm collimation at pitch of 1.5 does not clearly show anatomic relationship between tumor and interlobar fissures. Fig year-old man with lung cancer in anteromedial basal segment of left lower lobe., Sagittal multiplanar reconstruction (MPR) image obtained using 1-mm collimation at pitch of 1.5 shows neoplastic lobulations extending fissural plane (arrows)., MPR image obtained using 2-mm collimation at pitch of 1.5 does not clearly show relationship between tumor margin and fissural plane. In conclusion, 1-mm-collimation volume data are required to visualize all the interlobar fissures as a sharp line on sagittal MPR images except the minor fissure and the superior portion of the right major fissure, for which a 0.5-mm collimation is required. References 1. Chooi WK, Matthews S, ull MJ, Morcos SK. Multislice helical CT: the value of multiplanar image re- 396 JR:187, ugust 2006

9 Normal Pulmonary Fissures on MDCT construction in assessment of the bronchi and small airways disease. r J Radiol 2003; 76: Quint LE, Whyte RI, Kazerooni E, et al. Stenosis of the central airways: evaluation by using helical CT with multiplanar reconstructions. Radiology 1995; 194: Remy-Jardin M, Remy J, Cauvain O, Petyt L, Wannebroucq J, eregi JP. Diagnosis of central pulmonary embolism with helical CT: role of two-dimensional multiplanar reformations. JR 1995; 165: Padhani, Fishman EK, Heitmiller RF, Wang KP, Wheeler JH, Kuhlman JE. Multiplanar display of spiral CT data of the pulmonary hila in patients with lung cancer. Clin Imaging 1995; 19: Killeen KL, Mirvis SE, Shanmuganathan K. Helical CT of diaphragmatic rupture caused by blunt trauma. JR 1999; 173: rink J, Heiken JP, Semenkovich J, Teefey S, McClennan L, Sagel SS. bnormalities of the diaphragm and adjacent structures: findings on multiplanar spiral CT scans. JR 1994; 163: Honda O, Johkoh T, Yamamoto S, et al. Comparison of quality of multiplanar reconstructions and direct coronal multidetector CT scans of the lung. JR 2002; 179: rakawa H, Sasaka K, Lu WM, Hirayanagi N, Nakajima Y. Comparison of axial high-resolution CT and thin-section multiplanar reformation (MPR) for diagnosis of diseases of the pulmonary parenchyma: preliminary study in 49 patients. J Thorac Imaging 2004; 19: Remy-Jardin MJ, Campistron P, mara, Mastora I, Delannoy V, Remy J. Workflow issue with multislice CT of the thorax: usefulness of multiplanar reformations in the diagnostic approach of infiltrative lung disease. (abstr) Radiology 2001; 221(P): Quint LE, Glazer GM, Orringer M. Central lung masses: prediction with CT of need for pneumonectomy versus lobectomy. Radiology 1987; 165: Glazer HS, nderson DJ, DiCroce JJ, et al. natomy of the major fissure: evaluation with standard and thin-section CT. Radiology 1991; 180: erkmen YM, uh YH, Davis SD, Kazam E. natomy of the minor fissure: evaluation with thin-section CT. Radiology 1989; 170: Sakai O, Takahashi K, Nakashima N, et al. CT visualization of the major pulmonary fissures: value of 25 degrees cranially tilted axial scans. JR 1993; 161: Storto ML, Ciccotosto C, Guidotti, Merlino, Patea RL, onomo L. Neoplastic extension across pulmonary fissures: value of spiral computed tomography and multiplanar reformations. J Thorac Imaging 1998; 13: Remy-Jardin M, Remy J, Mayo JR, Muller NL. cquisition, injection, and reconstruction techniques. In: Remy-Jardin M, Remy J, Mayo JR, Muller JR, eds. CT angiography of the chest. Philadelphia, P: Lippincott Williams & Wilkins, 2001: Wang G, Vannier MW. Longitudinal resolution in volumetric X-ray computerized tomography: analytical comparison between conventional and helical computerized tomography. Med Phys 1994; 21: JR:187, ugust