Pulmonary Nodules: Contrast-Enhanced Volumetric Variation at Different CT Scan Delays
Abstract
OBJECTIVE. The purpose of this study was to assess the effects of IV contrast medium and different CT scan delays on volumetric measurements of pulmonary nodule.
MATERIALS AND METHODS. Automated volumes of 35 pulmonary nodules were calculated with two dedicated software packages (designated as software A and software B) for each unenhanced and contrast-enhanced CT scan at 30-, 60-, 120-, 180-, and 300-second delays (injection protocol, 2 mL/s and 2 mL/kg). Nodule attenuation was also determined. Differences between unenhanced and contrast-enhanced volumes were analyzed by Wilcoxon's signed rank test. Contrast-enhanced volume ratio was calculated as the ratio between contrast-enhanced and unenhanced nodule volume.
RESULTS. Contrast-enhanced volumes were significantly larger than unenhanced volumes (p < 0.05) for all the timing delays except at 30 seconds for software A, and no significant differences were found among volumes measured with both software programs at different contrast-enhanced delays. Median volume ratios between contrast-enhanced and unenhanced volumes were 1.04–1.07 for software A and 1.04–1.06 for software B, and median volume ratios within different contrast-enhanced delays were 0.99–1.03 for software A and 0.99–1.04 for software B. We did not find any significant association between contrast-enhanced volume ratio and nodule diameter, site, shape, unenhanced density, or contrast-enhanced density ratio (p > 0.05).
CONCLUSION. We recommend comparing volume of pulmonary nodules obtained from CT examinations only if they are all performed with or without contrast material, whereas nodule volumes obtained by use of enhanced CT performed with different scan delays are comparable.
Introduction
CT follow-up of lung nodules is commonly used to assess potential growth of undetermined lesions and to evaluate therapeutic response of pulmonary metastases. To reduce intra- and interobserver variability of manual measurements [1, 2], some authors [3, 4] recommend automated volume calculation as a reliable tool to evaluate potential nodule variation that, in clinical practice, could drastically influence the diagnosis and eventual therapeutic choices.
The accuracy of such a method is therefore crucial. Besides the physiologic and pathologic conditions [5–8] as well as the intrinsic nodule characteristics [6, 9], there are many scanning and reconstruction parameters affecting the calculation of automated volume, such as section thickness, algorithm, and reconstruction interval [10, 11].
The effect of contrast medium administration on volumetric analysis has been less investigated; to our knowledge, only one other study has evaluated lung nodule volume variability before and after contrast medium injection [12], but none has determined the influence of different CT scan delays on volumetric variation. Even though iodinated contrast medium injection is mandatory for several CT examinations, in clinical practice we sometimes compare CT examinations performed with and without contrast medium or CT examinations with different timing delays even if the same CT protocol is being adopted; such conditions could affect the accuracy of automated volume calculation. As an example, lung nodules may be initially detected on CT examinations performed without contrast medium, but follow-up CT studies might require contrast medium. Moreover, a nodule could be discovered on survey CT of the chest and abdomen using triple-phase scan delays, and the patient might come back a few months later for a dedicated chest CT scan using different scan delays.
In this study, we evaluated the variation of automated volume measurements of lung nodules during dynamic multiphase contrast-enhanced CT, using two different commercially available software packages. Thus, the aim of this study was to assess the effects of IV contrast material and different CT scan delays on automated volume measurement of pulmonary nodules and to determinate whether different software packages provide different contrast-enhanced measurements.
Materials and Methods
Data Source
All data used in this retrospective study were collected from a previous research protocol aimed at diagnosing the nature of undetermined pulmonary nodules and small mass lesions by dynamic contrast enhancement study. In the previous research protocol, which was conducted from December 2004 to October 2006, 78 patients (57 men and 21 women; mean age, 64 years; range, 52–76 years) with single undetermined solid pulmonary nodule ≥ 8 mm detected in a lung cancer screening program underwent dynamic contrast-enhanced chest CT. Our institutional review board approved the research protocol for this study, and written informed consent was provided by all patients.
Examinations were performed using a 16-MDCT system (LightSpeed 16, GE Healthcare). Before starting CT, patients were asked to repeat the same breath-hold to reduce significant respiratory excursions. To localize the nodule, a whole-chest low-dose CT scan was performed (140 kVp; 30 mA; tube rotation, 0.8 second; pitch, 1.75; slice thickness, 2.5 mm; standard reconstruction algorithm). The acquisition field of view (FOV) ranged from 320 to 370 mm, depending on the patient's body habitus.
Once the nodule was identified, we performed an unenhanced targeted FOV thin-section CT scan (120 kVp; automatic tube current modulation range, 100–440 mA; pitch, 0.938; slice thickness, 0.625 mm; standard reconstruction algorithm; FOV, 200 mm) through the nodule, for 50 mm along the z-axis. Nodules were excluded if they had calcification (four nodules), cavities (one nodule), or fat (three nodules) at thin-section CT. Thus, of the 78 enrolled patients, lung nodule enhancement was performed for 70 patients with a single nodule.
After the administration of contrast medium, five dynamic thin-section CT scans with the same unenhanced thin-section CT parameters were performed at 30, 60, 120, 180, and 300 seconds. Iodinated nonionic contrast medium (iobitridol, 350 mg I/mL [Xenetix, Guerbet]) was administered IV at a rate of 2 mL/s (dose, 2 mL/kg of body weight) using a power injector (Stellant, Medrad).
Present Study
Nodule selection and CT images analysis— From May 2008 through August 2009, all CT images obtained for the 70 patients for whom lung nodule enhancement was performed were retrospectively analyzed by two radiologists (with 8 and 3 years of experience in CT) using two different commercially available software packages dedicated to volume analysis: software A (ALA Single, GE Healthcare) and software B (LungCARE, Siemens Healthcare). Lung nodule volume and attenuation were calculated for each unenhanced and contrast-enhanced scan (Fig. 1).
In the first session, all anonymized CT data sets were transferred to a workstation (Advantage 4.2, GE Healthcare). Of 70 pulmonary nodules evaluated, 26 nodules with visible pleural or vessel attachment were excluded from the study.
The maximum transverse axial nodule diameter measured with an electronic caliper at unenhanced CT scan, the site (lobe), and the shape (round, irregular, or polygonal) of each nodule were reported by the two radiologists. A third radiologist (15 years of experience in CT) resolved by consensus differences in classification and discrepancy in measurements between the first two radiologists. In the second session, software A was used by a single radiologist, with 8 years of experience in CT, to calculate pulmonary nodule volume (cubic millimeters) of each unenhanced and contrast-enhanced acquisition.
After the nodule has been identified and manually marked by the operator with a mouse click, software A performs an automatic segmentation of the nodule by combining watershed segmentation and shape-analysis techniques [13] and elaborates a 3D template model, providing a nodular volume estimate in cubic millimeters. To guarantee a completely operator-independent process, nine nodules marked by the radiologist that could not be automatically and properly (partial segmentation or oversegmentation) segmented by dedicated software package were excluded; the completeness of each nodule segmentation was examined slice by slice on axial images. Thirty-five subjects (26 men and nine women; mean age, 60 years; range, 53–69 years) with a single pulmonary nodule or a small mass lesion up to 33 mm were eventually enrolled in this study.
In the third session, the anonymized data sets of the 35 pulmonary nodules evaluated with software A were transferred to a second workstation (Leonardo Workstation, Siemens Healthcare), and the same radiologist, who was blind to software A results, made the same evaluations described in the previous paragraph using software B. This software package is optimized for the evaluation of small soft-tissue density nodules of approximately spherical shape. After marking a nodule with a mouse click, a volume of interest is defined. An initial structure of interest is determined by applying a fixed density threshold and by extracting a 3D-connected structure consisting of the nodule and potential adjacent structures (e.g., vessels or parts of the chest wall). Subsequently, a small spherical 3D template originating from the click point is gradually expanded; its cross-correlation with the segmented nodule is computed for each step. The peak value of the cross-correlation curve is determined, and a cutoff value next to the peak value is set empirically. In this manner, an optimal spherical 3D template is generated that represents the nodule best. Finally, nodule segmentation is performed by fusing the optimal 3D template and the structure of interest, followed by spatial reasoning to remove adjacent structures [4].
Nodule attenuation—Pulmonary nodules attenuation (in Hounsfield units) was reported for each unenhanced and contrast-enhanced acquisition using both software packages. Software B provides nodule attenuation values calculated from the 3D template, so the Hounsfield unit value is an estimate of the whole nodule. Because our release of software A does not provide an automatic 3D attenuation of the nodule, 2D attenuation values (in Hounsfield units) were obtained placing a region of interest on the CT slice of maximal nodule diameter, covering 60% of the nodule area, as measured at mediastinal window settings (window width, 400 HU; window level, 40 HU) on transverse images.
We retrospectively reviewed the clinical histories of the 35 subjects to determine the nodule diagnosis. Nine nodules were histologically proven lung cancer (six adenocarcinomas and three squamous cell carcinomas). The 26 remaining nodules were considered benign according to dimensional stability or regression at further follow-up CT examinations (mean follow-up, 37 months; range, 25–47 months).
Statistical Analysis
Because the measured volumes and densities were not normally distributed, we used the nonparametric Wilcoxon's signed rank test to compare the measures of pulmonary nodule volume and density before and after contrast medium administration, at each contrast-enhanced scan delay. The same test was used to further evaluate whether there were significant variations in volume measurements among different contrast-enhanced delays. We used the Wilcoxon's two-sample test to evaluate whether increased volume and density after administration of contrast medium differed by nodule diagnosis (benign or malignant nodule).
Contrast-enhanced volume ratio was calculated for each CT scan delay as contrast-enhanced nodule volume divided by unenhanced nodule volume. We evaluated the correlation between contrast-enhanced volume ratio with baseline characteristics using the Kruskal-Wallis test for categorical variables (sex, site, and shape) and the Spearman's rank correlation coefficient for continuous variables (age and unenhanced density). To evaluate whether variation between unenhanced and contrast-enhanced volume measures correlated with variation between unenhanced and contrast-enhanced density measures, we calculated the contrast-enhanced density ratio, as described for volume, and correlated it with contrast-enhanced volume ratio by Spearman's rank correlation coefficient.
A p value of < 0.05 was considered statistically significant. All the statistical analysis were performed using SAS software, version 8.2 (SAS Institute).
Results
The 35 nodules were distributed as follows: 31% (n = 11) right upper lobe, 6% (n = 2) middle lobe, 23% (n = 8) right lower lobe, 20% (n = 7) left upper lobe, and 20% (n = 7) left lower lobe. The mean diameter of the 35 nodules, measured at unenhanced thin-section CT, was 13.8 mm (range, 8–33 mm; SD, 5.37 mm); the mean (SD) diameter of benign nodules was 12.7 mm (3.98 mm) versus 16.5 mm (7.42 mm) for malignant nodules. Fifty-two percent were round, 34% were irregular, and 14% were polygonal.
Mean volume and mean density before and after administration of contrast medium are shown in Table 1 and Figure 2A, 2B. The median increase in nodule volume and density after contrast medium for each CT scan delay is shown in Table 2. The measured volumes and densities at contrast-enhanced imaging were significantly larger than those at unenhanced imaging for all the timing delays for both software packages, except at 30 seconds for software A (Table 2). We did not observe any significant difference among measurements of volume calculated at different contrast-enhanced delays for both software packages (p > 0.05).
| Nodule Volume (mm3), Mean (SD) | Nodule Density (HU), Mean (SD) | |||
|---|---|---|---|---|
| CT Scan Delay (s) | Software A | Software B | Software A | Software B |
| 0 | 1,295 (1,764) | 1,226 (1,767) | 35.49 (19.89) | 20.74 (18.90) |
| 30 | 1,369 (1,898) | 1,268 (1,828) | 38.44 (19.05) | 29.38 (19.32) |
| 60 | 1,395 (1,947) | 1,278 (1,855) | 47.05 (34.10) | 40.56 (26.62) |
| 120 | 1,406 (1,926) | 1,299 (1,780) | 57.60 (28.31) | 47.21 (24.84) |
| 180 | 1,370 (1,903) | 1,290 (1,803) | 55.18 (24.58) | 46.68 (21.87) |
| 300 | 1,394 (1,967) | 1,299 (1,817) | 56.19 (28.16) | 46.47 (20.66) |
Note—Software A is ALA Single (GE Healthcare), and software B is LungCARE (Siemens Healthcare).
| Software A | Software B | |||
|---|---|---|---|---|
| Nodule Measurement, CT Scan Delay | Difference Between Contrast-Enhanced and Unenhanced Scan, Median (Range) | pa | Difference Between Contrast-Enhanced and Unenhanced Scan, Median (Range) | pa |
| Volume (mm3) | ||||
| 30 s | 37 (-277 to 671) | 0.06 | 33.5 (-114 to 380) | 0.02 |
| 60 s | 35 (-487 to 1,082) | 0.04 | 52 (-393 to 545) | 0.01 |
| 120 s | 31 (-1,057 to 800) | 0.004 | 72 (-69 to 526) | < 0.0001 |
| 180 s | 36 (-1,109 to 689) | 0.004 | 64 (-214 to 579) | 0.003 |
| 300 s | 14 (-481 to 1,123) | 0.007 | 73 (-358 to 539) | 0.007 |
| Density (HU) | ||||
| 30 seconds | 3.15 (-19.99 to 44.08) | 0.22 | 3.00 (-21.00 to 55.00) | 0.01 |
| 60 seconds | 8.66 (-82.87 to 69.88) | 0.01 | 10.50 (-24.00 to 120.00) | < 0.0001 |
| 120 seconds | 19.19 (-13.73 to 70.32) | < 0.0001 | 19.00 (-5.0 to 129.00) | < 0.0001 |
| 180 seconds | 18.89 (-17.84 to 69.98) | < 0.0001 | 24.00 (-3.0 to 112.00) | < 0.0001 |
| 300 seconds | 16.24 (-18.09 to 65.45) | < 0.0001 | 23.00 (-2.00 to 113.00) | < 0.0001 |
Note—Software A is ALA Single (GE Healthcare), and software B is LungCARE (Siemens Healthcare).
a
Wilcoxon's signed rank test.
Median volume ratios between unenhanced and contrast-enhanced imaging ranged from 1.04 to 1.07 for software A and from 1.04 to 1.06 for software B. Median volume ratios within different contrast-enhanced delays ranged from 0.99 to 1.03 for software A and from 0.99 to 1.04 for software B (Table 3).
| Volume Ratio, Median (Range) | ||
|---|---|---|
| Time Comparison (s) | Software A | Software B |
| 30 vs 0 | 1.06 (0.67-1.41) | 1.04 (0.80-1.66) |
| 60 vs 0 | 1.06 (0.60-1.61) | 1.04 (0.62-1.35) |
| 120 vs 0 | 1.06 (0.57-1.39) | 1.06 (0.78-1.46) |
| 180 vs 0 | 1.07 (0.55-1.41) | 1.06 (0.50-1.42) |
| 300 vs 0 | 1.04 (0.67-1.57) | 1.04 (0.65-1.57) |
| 60 vs 30 | 1.01 (0.66-1.56) | 1.00 (0.67-1.32) |
| 120 vs 30 | 1.02 (0.64-1.61) | 1.02 (0.88-1.31) |
| 180 vs 30 | 1.03 (0.56-1.37) | 1.04 (0.51-1.25) |
| 300 vs 30 | 1.03 (0.74-1.50) | 1.03 (0.70-1.26) |
| 120 vs 60 | 1.00 (0.71-1.76) | 1.03 (0.87-1.62) |
| 180 vs 60 | 1.00 (0.51-1.65) | 1.02 (0.47-1.53) |
| 300 vs 60 | 1.00 (0.78-1.52) | 1.02 (0.87-1.64) |
| 180 vs 120 | 0.99 (0.51-1.15) | 0.99 (0.55-1.16) |
| 300 vs 120 | 1.00 (0.62-1.41) | 1.00 (0.65-1.21) |
| 300 vs 180 | 1.00 (0.68-1.81) | 1.00 (0.69-1.94) |
Note—Software A is ALA Single (GE Healthcare), and software B is LungCARE (Siemens Healthcare).
According to nodule diagnosis, the median increase in volume after administration of contrast medium was larger in malignant nodules than benign nodules at 120 and 180 seconds (p = 0.04) for software A and at all time delays but 30 seconds for software B (Table 4). We did not find any significant association between contrast-enhanced volume ratio and sex, site, shape, age, unenhanced diameter, unenhanced density, or contrast-enhanced density ratio (p > 0.05).
| Software A | Software B | |||||
|---|---|---|---|---|---|---|
| Nodule Measurement, CT Scan Delay | Benign Nodule | Malignant Nodule | pa | Benign Nodule | Malignant Nodule | pa |
| Volume (mm3) | ||||||
| 30 s | 15.5 | 55.0 | 0.29 | 18.0 | 59.0 | 0.08 |
| 60 s | 28.0 | 59.0 | 0.38 | 22.0 | 48.0 | 0.37 |
| 120 s | 16.0 | 313.0 | 0.04 | 22.0 | 97.0 | 0.08 |
| 180 s | 25.0 | 110.0 | 0.04 | 22.0 | 71.0 | 0.25 |
| 300 s | 12.5 | 131.0 | 0.23 | 22.0 | 46.0 | 0.37 |
| Density (HU) | ||||||
| 30 s | 2.14 | 5.00 | 0.69 | 0.00 | 12.00 | 0.08 |
| 60 s | 1.81 | 36.13 | 0.05 | 4.00 | 27.00 | 0.01 |
| 120 s | 12.51 | 40.83 | 0.05 | 12.00 | 46.00 | 0.01 |
| 180 s | 9.09 | 38.37 | 0.01 | 12.00 | 35.00 | 0.01 |
| 300 s | 12.12 | 34.49 | 0.005 | 17.00 | 37.00 | 0.01 |
Note—Except for p values, data are median differences between contrast-enhanced and unenhanced scans. Software A is ALA Single (GE Healthcare), and software B is LungCARE (Siemens Healthcare).
a
Wilcoxon's two-sample test.
Discussion
This study shows that automated volume calculation of lung nodules with contrast-enhanced imaging yields results larger than those for unenhanced imaging, with a median increase in contrast-enhanced volume of 4–7% for software A and 4–6% for software B, according to different scan delays. Our results were similar to those from a previous study [12], in which volume ranged by a mean of 5.4–6.5%, depending on whether a high- or low-spatial-frequency algorithm was used; however, in that study, various contrast medium concentrations, doses, injection rates, and scan delays (range, 25–64 seconds) were used, which partially influenced the results. Although we used the same CT protocol, contrast material administration, and two different commercially available software packages, the reason for the increased contrast-enhanced volume is still not clear.
Because nodules with visible vessel attachment were excluded in this study, surrounding microvessel enhancement could not significantly increase nodule volume. Supposing that the enhancement of the peripheral part of the nodule could enlarge the shape of the nodule, we presume that contrast-enhanced volume is affected by overestimated automated segmentation as a result of increased contrast-enhanced attenuation of the nodule's edge. This could explain why contrast material injection increased nodule volume. However, further investigations would be necessary to clarify this important issue, particularly regarding the correlation between degree of lung nodule attenuation values and volume variation.
To our knowledge, no previous studies have investigated the effects of different contrast-enhanced CT scan delays on nodule volumetric measurements. Our results show that automated volume variation was not significantly different within contrast-enhanced CT scan delays, ranging from –1% to 3% for software A and from –1% to 4% for software B; thus, in clinical practice, volume measurements obtained by enhanced CT performed with different scan delays are comparable.
In addition, we have evaluated the association between increased contrast-enhanced volume and nodule diagnosis, obtaining a median increase in volume larger in malignant than in benign nodules at 120 and 180 seconds only for software A and at all time delays except 30 seconds for software B. This finding could be explained by the higher degree of contrast enhancement of malignant nodules, but in view of the small number (n = 9) of malignant nodules, we cannot draw any significant conclusions.
The standard algorithm (relatively low-spatial-frequency algorithm) was used for image reconstruction in this study. Even though a high-spatial-frequency algorithm has been recommended for volumetric evaluation, the standard algorithm used can be considered accurate because no substantial differences in the accuracy of volume estimations across seven different reconstruction kernels were reported in a previous in vitro study [14].
This study has some limitations. First, a small number of nodules were examined. Although our results achieved statistical significance, a larger number of nodules would improve information and clarify some aspects of this topic. A second limitation concerns the different methods used to calculate the lung nodule attenuation, which was evaluated in 2D with software A and in 3D with software B; however, with both methods, we did not find any significant correlation between volume variation and degree of contrast-enhanced nodule attenuation. A third limitation concerns the characteristics of the nodules chosen as targets of the study (i.e., solid nodules with no visible pleural or vascular contact). We decided to use this target to allow evaluation of volume variation in an optimal condition. In fact, evaluation of repeatability could have been further influenced by juxtapleural or juxtavascular or nonsolid nodules [15]. Finally, the results reported in this study are strictly tied to the performance of the two software packages used, and the inherent variability of the software could affect the volumetric analysis, particularly for small nodules [16]; however, in this study, we did not find any significant association between contrast-enhanced volume variation and nodule diameter.
In conclusion, regardless of the two different software packages used, automated volume calculation of pulmonary nodules is significantly larger after contrast material administration, and different contrast-enhanced CT scan delays provided similar volumetric measurements. The meaning of contrast-enhanced variation is even more important if it is added to other variables that could affect nodule volume calculation in follow-up CT examinations [5–11]. Thus, in clinical practice, we recommend comparing automated volume of lung nodules obtained from CT examinations only if they are all performed with or without contrast material, whereas nodule volume obtained from enhanced CT examinations performed with different scan delays are comparable.
Footnote
Address correspondence to C. Rampinelli ([email protected]).
References
1.
Revel MP, Bissery A, Bienvenu M, Aycard L, Lefort C, Frija G. Are two-dimensional CT measurements of small noncalcified pulmonary nodules reliable? Radiology 2004; 231:453–458
2.
Marten K, Auer F, Schmidt S, Kohl G, Rummeny EJ, Engelke C. Inadequacy of manual measurements compared to automated CT volumetry in assessment of treatment response of pulmonary metastases using RECIST criteria. Eur Radiol 2006; 16:781–790
3.
Goodman LR, Gulsun M, Washington L, Nagy PG, Piacsek KL. Inherent variability of CT lung nodule measurements in vivo using semiautomated volumetric measurements. AJR 2006; 186:989–994
4.
Wormanns D, Kohl G, Klotz E, et al. Volumetric measurements of pulmonary nodules at multi-row detector CT: in vivo reproducibility. Eur Radiol 2004; 14:86–92
5.
Petkovska I, Brown MS, Goldin JG, et al. The effect of lung volume on nodule size on CT. Acad Radiol 2007; 14:476–485
6.
Gietema HA, Schaefer-Prokop CM, Mali WP, Groenewegen G, Prokop M. Pulmonary nodules: interscan variability of semiautomated volume measurements with multisection CT influence of inspiration level, nodule size, and segmentation performance. Radiology 2007; 245:888–894
7.
Boll DT, Gilkeson RC, Fleiter TR, Blackham KA, Duerk JL, Lewin JS. Volumetric assessment of pulmonary nodules with ECG-gated MDCT. AJR 2004; 183:1217 –1223
8.
Goo JM, Kim KG, Gierada DS, Castro M, Bae KT. Volumetric measurements of lung nodules with multi-detector row CT: effect of changes in lung volume. Korean J Radiol 2006; 7:243–248
9.
Wang Y, van Klaveren RJ, van der Zaag-Loonen HJ, et al. Effect of nodule characteristics on variability of semiautomated volume measurements in pulmonary nodules detected in a lung cancer screening program. Radiology 2008; 248:625–631
10.
Honda O, Sumikawa H, Johkoh T, et al. Computer-assisted lung nodule volumetry from multi-detector row CT: influence of image reconstruction parameters. Eur J Radiol 2007; 62:106–113
11.
Goo JM, Tongdee T, Tongdee R, Yeo K, Hildebolt CF, Bae KT. Volumetric measurement of synthetic lung nodules with multi-detector row CT: effect of various image reconstruction parameters and segmentation thresholds on measurement accuracy. Radiology 2005; 235:850–856
12.
Honda O, Johkoh T, Sumikawa H, et al. Pulmonary nodules: 3D volumetric measurement with multidetector CT—effect of intravenous contrast medium. Radiology 2007; 245:881–887
13.
Revel MP, Lefort C, Bissery A, et al. Pulmonary nodules: preliminary experience with three-dimensional evaluation. Radiology 2004; 231:459–466
14.
Ravenel JG, Leue WM, Nietert PJ, Miller JV, Taylor KK, Silvestri GA. Pulmonary nodule volume: effects of reconstruction parameters on automated measurements—a phantom study. Radiology 2008; 247:400 –408
15.
Ko JP, Marcus R, Bomsztyk E, et al. Effect of blood vessels on measurement of nodule volume in a chest phantom. Radiology 2006; 239:79–85
16.
Rampinelli C, De Fiori E, Raimondi S, Veronesi G, Bellomi M. In vivo repeatability of automated volume calculations of small pulmonary nodules with CT. AJR 2009; 192:1657 –1661
Information & Authors
Information
Published In
Copyright
© American Roentgen Ray Society.
History
Submitted: June 19, 2009
Accepted: January 5, 2010
First published: November 23, 2012
Keywords
Authors
Metrics & Citations
Metrics
Citations
Export Citations
To download the citation to this article, select your reference manager software.