| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 All authors: Department of Radiology, University Hospitals of Cleveland, Case Western Reserve University, 11100 Euclid Ave., Cleveland, OH 44106-5056.
Received January 19, 2004;
accepted after revision April 12, 2004.
Address correspondence to D. T. Boll
(boll{at}uhrad.com).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Seventy-three small noncalcified pulmonary nodules were identified in 30 patients who underwent ECG-gated MDCT. The volume of each nodule was assessed throughout the cardiac cycle using computer-aided automatic segmentation algorithms, and the assessment was repeated three times. To ensure the validity of the subtle changes in volume that were detected, we determined the volume and signal attenuation in phantom data sets and patient nodules without temporal or spatial differentiation. Subsequently, nodules were assigned to pulmonary segments, and volume changes were correlated to cardiac phases, nodular location, and mean nodular size. Statistical multivariate tests were performed to evaluate significant patterns.
RESULTS. The validity of significant measurements was proven in evaluated phantom data sets with a general tendency toward overestimating nodular volume (p = 0.492). Statistical evaluation of nodular signal attenuation confirmed true deformation and compression of nodules rather than partial volume effects as the reason for volume variations (p = 0.874). Differentiating pulmonary nodules in cardiac phases, pulmonary locations, and mean nodular volumes, we found that one single effect did not determine the amount of cardiovascular motion conveyed to pulmonary parenchyma and subsequently led to nodule deformation. Multivariate testing revealed statistically significant measures identifying patterns correlating variation in nodular volume with cardiac phase (p < 0.001), nodular location (p = 0.007), and mean nodular size (p < 0.001).
CONCLUSION. Cardiovascular motion was disproportionately conveyed to various pulmonary segments and led to changes in the volume of pulmonary nodules, especially in small pulmonary nodules. A precise volumetric assessment was therefore possible only by identifying the underlying cardiac phase.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Whereas larger pulmonary nodules, those more than 10 mm in greatest dimension, have a higher likelihood of malignancy than small nodules [4], subcentimeter nodules are more difficult to characterize and are more commonly benign but do not preclude malignancy [5]. After the detection of a small pulmonary nodule, nodule growth rate—rather than morphologic CT characteristics, such as the presence of marginal calcifications [6]—has been proposed to differentiate malignant from benign nodules [7]. Although malignant nodules typically double in volume in 30 days-14 months, doubling times of less than 30 days are usually found in patients with inflammatory or infectious diseases, whereas doubling times of more than 14 months have been associated with benign nodules such as hamartomas [8].
The importance of determining the exact dimensions of a nodule is not reflected in traditional assessments by measuring transverse 2D diameters with inherent and significant intraand interobserver variabilities [9]. However, the introduction of multidetector technology has transformed CT from a transaxial cross-sectional technique into a true 3D imaging technique. The increased number of parallel detector arrays combined with ECG gating not only has accelerated the scanning speed but also has led to a simultaneous increase in spatial resolution, thereby allowing image reconstruction throughout the entire cardiac cycle to visualize morphologic cardiac-dependent changes [10]. In one study, the combination of high-resolution MDCT data sets and computer-aided quantitative volume measurement algorithms allowed a volumetric assessment of small pulmonary nodules with minimized variation and decreased measurement error [11]. Another study proved that prospective axial ECG triggering of chest CT examinations enhances the detectability of small pulmonary nodules by offsetting the effects of cardiac motion [12].
In our study, we sought to assess the volume of small pulmonary nodules throughout the cardiac cycle on ECG-gated MDCT. This study was performed to test the hypothesis that cardiovascular motion disproportionately conveyed to various pulmonary segments leads to volumetric changes in small pulmonary nodules, thereby impacting precise determination of nodule growth rates.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Image Acquisition
All CT examinations were performed on a multidetector scanner (Mx8000IDT,
Philips Medical Systems) with a parallel arrangement of 16 detector arrays
after ECG electrodes were placed on the patient's chest and were subsequently
attached to the CT unit for ECG gating. Helical CT studies covered a field of
view from the lateral diaphragmatic recesses to the apices of the lungs. The
helical scanning protocol consisted of a detector collimation of 16 x
0.75 mm, a detector pitch of 0.24, a gantry rotation period of 0.42 sec, and a
matrix size of 512 x 512 pixels applying an X-ray voltage of 140 kV and
current of 150 mAs. The duration of an entire examination was 21-27 sec.
Contrast-enhanced imaging commenced after 100 mL of a low-osmolar iodinated
contrast agent (Optiray 300 [ioversol], Mallinckrodt) was administered via an
antecubital vein with a 4.0 mL/sec flow rate using bolus-tracking techniques
with the threshold region of interest placed in the left ventricle.
A dedicated reconstruction for ECG-gated MDCT was used to provide a well-defined slice sensitivity profile, resulting in a 0.6-mm-interval reconstruction. Therefore, if patient heart rate was less than 65 beats per minute, one subsegment of the consecutive MDCT data from the same heart period was used, resulting in a temporal resolution of 210 msec. In patients with higher heart rates, data can be combined from multiple consecutive cardiac cycles, resulting in a temporal resolution varying between 53 and 210 msec, because heart cycle and gantry rotation periods must be desynchronized [13]. Five complete sets of 3D near-isotropic image data sets covering the entire lung at 0%, 25%, 50%, 75%, and 83% offset between two adjacent ECG R waves were subsequently reconstructed using a dedicated pulmonary kernel.
Image Analysis
All five image data sets acquired throughout the cardiac cycle of each of
the 30 patients were evaluated semiautomatically three times with an interval
of more than 1 week between each measurement repetition using commercially
available software for computer-aided volumetric measurements of pulmonary
nodules (MxView, Philips Medical Systems). The automatic segmentation
algorithms used two approaches synergistically to determine the precise
volumes of predefined pulmonary nodules
[14,
15].
The density-based approach with locally adaptive thresholding and region growing considers the fact that pulmonary nodules have relatively higher densities than lung parenchyma. After a nodule was manually defined, its mean density was calculated in a volume with a radius of 3 pixels around this defined center. Mean pulmonary density was determined by evaluating the outer margins of a 20-mm maximum-intensity-projection box with the selected nodule as its center. The model-based approach uses a priori knowledge of the morphology of small pulmonary nodules. In particular, their relatively compact and sphere-shaped appearance allowed calculations of compactness as a measure of how closely the detected nodule resembled a sphere.
The computer-aided quantitative volumetric assessment calculates volume, maximal 3D area, maximal transverse diameter, and the mean density in Hounsfield units of each pulmonary nodule.
The software application was calibrated before patient measurements in phantom studies. A spherical phantom of defined volume (60.0 mm3) was imaged according to the CT examination protocol. All five-phase data sets were evaluated five times using the computer-aided quantitative volumetric technique, and measurement validity was then statistically evaluated.
Statistical Evaluation
To determine measurement validity, we evaluated the analysis results of the
phantom studies with defined volumes using one-way analysis of variance
(ANOVA) with the Student-Newman-Keuls test.
After confirming the validity of the phantom measurements, we statistically evaluated the attenuation of every pulmonary nodule throughout the entire cardiac cycle of the patients in the study group. One-way ANOVA with the Student-Newman-Keuls test was used to confirm a true deformation and compression of nodules rather than partial volume effects as the reason for variation in the volumes.
Thereafter, nodular volume was compared with the determined maximal cross-sectional area and the transverse diameter of the pulmonary nodule. By calculating Spearman's rank correlation coefficient, we evaluated whether the transverse diameter and area showed comparable dynamic variations as determined by measuring the nodular volume.
Finally, all lung nodules were assigned to their corresponding pulmonary segments in a consensus decision by two experienced radiologists. Multivariate tests calculating Pillai-Bartlett trace and Wilks lambda analyzed whether any changes in nodular volume were related to the cardiac phase. If significance was detected, multivariate tests were used to further evaluate whether the magnitude of volume change, expressed as the coefficient of variance, was also dependent on pulmonary location and was related to nodular size. The mean nodular size and the magnitude of volume change were then analyzed by performing a curve estimation regression to identify the specific type of regression.
All statistical analysis was performed using SPSS software (version 11.5, Statistical Package for the Social Sciences); significance level was defined as a p value of less than 0.05.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The analysis of nodular attenuation throughout the entire cardiac cycle of every pulmonary nodule of patients using one-way ANOVA with the Student-Newman-Keuls test revealed an F distribution of 0.354 with a significance level of 0.878. The small noncalcified nodules presented a mean attenuation of 17.7 ± 4.8 H. Statistically significant homogeneity was detected by analyzing the attenuations of each pulmonary nodule throughout the entire cardiac cycle.
The 73 pulmonary nodules we detected ranged in volume from 0.2 to 399 mm3, with an overall average volume of 25.3 mm3. Comparing the dynamically changing volumes of each pulmonary nodule with its measured maximal cross-sectional area, we determined a mean Spearman's rank correlation coefficient of 0.671. The varying volumes of the pulmonary nodules were correlated with their corresponding maximal transverse diameters, and a mean Spearman's rank coefficient of 0.192 was obtained.
The changes in nodular volume statistically evaluated with multivariate tests indicated a highly significant dependency on the underlying cardiac phase; Pillai-Bartlett trace and Wilks lambda were less than 0.0001.
The detailed analysis focusing on the magnitude of these detected changes in relation to the pulmonary distribution of the small nodules, as summarized in Table 1, resulted in statistically significant measures of Pillai-Bartlett trace of 0.007 and Wilks lambda of less than 0.0001. Segmental distribution of the volume changes and cardiac phase with minimal volume variation, summarized in Table 2, presented statistically significant patterns.
|
|
Focusing on the right upper pulmonary lobe, we found that the variations in nodular volume decreased from segment 1, which was situated adjacent to an axis consisting of the ascending aorta and superior vena cava; to segment 2, located close to an axis of the ascending aorta and trachea; and, finally, to segment 3, with only a small caudal contact zone to the pulmonary artery (Figs. 1A and 1B). In the left upper lobe, the cardiovascular motion conveyed to the pulmonary parenchyma originated homogeneously from the aortic arch (Figs. 1C and 1D). The middle lobe presented small changes in nodular volume in segment 4, with only a small contact zone to the pulmonary artery and larger volumetric variations in segment 5, located adjacent to the right atrium and ventricle (Figs. 2A and 2C). In contrast, the left-sided lingula showed greater volumetric variations in segment 4 situated adjacent to the anterior cardiac silhouette, compared with segment 5, with a contact zone to the left ventricle (Figs. 2B and 2D). Both lower lobes showed variations in nodular volume, most significantly in segments 6 and 8, which were situated alongside the major pulmonary fissures (Figs. 3A, 3B, 3C, 3D). Additional significant variations in nodular volume were detected in the left segment 10 located directly adjacent to the thoracic aorta.
|
|
|
|
|
|
|
|
|
|
|
|
The statistical analysis focusing on the magnitude of the nodular changes related to the size of the small nodules resulted in statistically significant measures of Pillai-Bartlett trace and Wilks lambda of less than 0.001. Subsequently performed regression analysis confirmed a logarithmic regression with an intercept of 0.318 mm3, an associated p value of less than 0.001, and a slope of -0.084 and associated p of 0.029, thereby emphasizing that variations in nodular volumes were most evident in small nodules (Fig. 4).
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Nevertheless, determining growth rates of pulmonary nodules has proved difficult for many reasons. One- or two-dimensional perpendicular measurements to determine nodular diameters led to suboptimal results in one study because only one imaging plane could be visualized [18]. Furthermore, considerable inter- and intraobserver variability was found if the observers were not assisted by semiautomatic autocontour segmentation strategies [19]. Moreover, accurate 3D assessment of small solid pulmonary nodules relies on advances in CT detector technology. Data sets acquired with single-detector CT scanners showed significant measurement variability depending on the grade of spherical deformation of the nodules [9]. Also, selection of minimal slice thickness was of paramount importance in reducing the effects of partial volume averaging in nodule visualization—in particular, because automatic nodule segmentation is traditionally based on thresholding algorithms that have proven to be especially susceptible to partial volume effects [9].
Many prior limitations of assessing the volume of small pulmonary nodules with minimized variation and decreased measurement error were addressed by combining high-resolution 16-MDCT techniques with computer-aided density- and model-based algorithms for automatic nodule segmentation [11]. However, once we overcame the technical acquisition and postprocessing limitations, physiologic processes became visible, such as the influence of cardiovascular motion conveyed onto the pulmonary parenchyma [12]. In this study, we sought to assess whether cardiac motions—in particular, aortic and cardiac chamber motions—produce variability in the volume of small pulmonary nodules and whether the variability in volume is related to the phase of cardiac motion.
To ensure the validity and emphasize the significance of any detected subtle but rapid volumetric changes in small noncalcified pulmonary nodules throughout the cardiac cycle, we first analyzed phantom and patient nodules in their entirety without temporal or spatial differentiation. The phantom study emphasized that the volumetric analysis of the static sphere, resembling a volumetrically defined nodule, did not lead to any variation in volume throughout the cardiac phases, even though the spherical volume was constantly overestimated. In comparison with the variations in the measured volumes of pulmonary nodules, we concluded that pulmonary location and cardiac phase are decisive factors to explain these measured variations in volume. The analysis of attenuations of actual pulmonary nodules throughout the cardiac cycle resulted in no significant variation, thereby conclusively showing that any semiautomatically detected volume change is directly related to actual changes in the size and morphology of the measured nodule.
By correlating determined volumes with measured maximal transverse diameter during the cardiac cycle, we found—in concordance with studies performed without ECG-gating [9]—that the transverse nodule diameter did not reflect the dynamic variations of its actual morphology. However, good correlation was achieved when the maximal 3D area was chosen as the predictor of morphologic deformation.
When we differentiated the identified pulmonary nodules in the cardiac phases, pulmonary locations, and mean nodular volumes, it became evident that no single effect determined the amount of cardiovascular motion conveyed to pulmonary parenchyma or subsequently led to nodule deformation and compression.
The upper pulmonary lobe was predominantly deformed through structures located in the upper mediastinum—namely, the ascending aorta, aortic arch, superior vena cava, and trachea. Therefore, during early systole (which is characterized by closed mitral and aortic valves, no flow in the aorta, and reduced central venous flow), the least amount of cardiovascular motion was conveyed to the adjacent upper lobe structures. This study further showed that a pointed cardiovascular force, as seen on the right side originating from two axes consisting of the ascending aorta and superior vena cava and of the ascending aorta and trachea, induced greater pulmonary motion than the wider left-sided aortic arch-pulmonary interface. In the middle lobe and lingula, the maximal cardiovascular motion that conveyed to the anterior pulmonary segments originated from the ventricular apex. Although the right ventricle transferred minimum force to the middle lobe during early systole, the left, more agile ventricle shifted the lingula the least during late diastole. This minimum left ventricular force further shifted the dorsome-dial segments of the right lower lobe the least, whereas maximum pulmonary contortion was observed on segments alongside both major fissures.
In general, we observed that the relative magnitude of volume change was increased in the small pulmonary nodules compared with large ones. Therefore, a precise volumetric assessment of small nodules was complicated even further.
Our study had several limitations that must be addressed. By assigning the various pulmonary nodules solely to their pulmonary segments, proximity to neither the heart nor the chest wall was reflected in the spatial location of the nodules. Furthermore, the semiautomatic volume assessment software we used did not allow a close analysis of nodular deformation occurring during the cardiac cycle but reported only the descriptive parameters. In addition, this study did not follow up on any of the small noncalcified pulmonary nodules we detected to assess the impact of variation in nodular volume throughout the cardiac cycle over longer observation periods.
In conclusion, we accept the hypothesis that cardiovascular motion is disproportionately conveyed to various pulmonary segments and therefore leads to volumetric changes, especially in small pulmonary nodules. A precise volumetric assessment is therefore possible only by identifying the underlying cardiac phase.
The findings of this study emphasize that advances in imaging technology, such as further enhanced temporal and spatial resolution, will increasingly rely on individual physiologic parameters, such as cardiac phase, to correlate imaging accuracy with physiologic processes. Although threshold levels for an increase in nodular volume and interval of follow-up examinations continue to be debated, we conclude in accordance with the results in this study that the widely distributed MDCT units should use ECG gating for pulmonary imaging to allow visualization of pulmonary nodules in specific cardiac phases for reliable assessment of volume.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  K MARTEN, C DULLIN, W MACHANN, J S SCHMID, M DAS, K-P HERMANN, and C ENGELKE Comparison of flat-panel-detector-based CT and multidetector-row CT in automated volumetry of pulmonary nodules using an anthropomorphic chest phantom Br. J. Radiol., September 1, 2009; 82(981): 716 - 723. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O HONDA, M KAWAI, T GYOBU, Y KAWATA, T JOHKOH, J SEKIGUCHI, N TOMIYAMA, S YOSHIDA, H SUMIKAWA, A INOUE, et al. Reproducibility of temporal volume change in CT of lung cancer: comparison of computer software and manual assessment Br. J. Radiol., September 1, 2009; 82(981): 742 - 747. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Rampinelli, E. De Fiori, S. Raimondi, G. Veronesi, and M. Bellomi In Vivo Repeatability of Automated Volume Calculations of Small Pulmonary Nodules With CT Am. J. Roentgenol., June 1, 2009; 192(6): 1657 - 1661. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Girvin and J. P. Ko Pulmonary Nodules: Detection, Assessment, and CAD Am. J. Roentgenol., October 1, 2008; 191(4): 1057 - 1069. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Petrou, L. E. Quint, B. Nan, and L. H. Baker Pulmonary Nodule Volumetric Measurement Variability as a Function of CT Slice Thickness and Nodule Morphology Am. J. Roentgenol., February 1, 2007; 188(2): 306 - 312. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. R. Goodman, M. Gulsun, L. Washington, P. G. Nagy, and K. L. Piacsek Inherent Variability of CT Lung Nodule Measurements In Vivo Using Semiautomated Volumetric Measurements. Am. J. Roentgenol., April 1, 2006; 186(4): 989 - 994. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
