| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Thoracic Imaging |
1 From the Department of Radiology, Weill Medical College of Cornell University, 525 E 68th St, New York, NY 10021 (W.J.K., D.F.Y., S.C.F., C.I.H.); and School of Electrical and Computer Engineering, Cornell University, Ithaca, NY (A.P.R.). Received April 8, 2003; revision requested June 20; revision received August 17; accepted September 29. Supported in part by National Institutes of Health grant R01-CA78905. Address correspondence to C.I.H. (e-mail: chensch@med.cornell.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: One hundred fifteen pulmonary nodules for which two thin-section small–field-of-view CT scans were obtained and which were stable during 2-year observation were evaluated. A standard group of 94 nodules (with no or minimal artifact) and an expanded group of 105 nodules (including those with moderate artifacts) were examined. Percentage volume change (PVC) and monthly volumetric growth index (MVGI) were computed for each nodule pair. By using estimates of the variation in PVC in stable nodules as a function of initial diameter, critical time to follow-up CT was estimated; this time is the earliest point at which growth in a nodule of a given size can be reliably identified with repeat CT.
RESULTS: The SD of PVC decreased with increasing nodule size from 18.5% in 2–5-mm nodules to 10.6% in 5–8-mm nodules and to 7.47% in 8–10-mm nodules. Inclusion of cases with moderate motion artifacts increased the SD of PVC to 27.4% in 2–5-mm nodules, to 17.1% in 5–8-mm nodules, and to 19.3% in 8–10-mm nodules. Critical time to follow-up CT for nodules detected at baseline screening was 12, 5, and 3 months and 1 month for those with initial sizes of 2, 5, 8, and 10 mm, respectively. For nodules detected at annual repeat screening, it was 4 and 3 months and 1 month for nodules that were 3, 4, and 5 mm or larger in size, respectively. Mean MVGI in 94 standard cases was 0.06%, and standard error was 0.21%.
CONCLUSION: Factors that affect reproducibility of nodule volume measurements and critical time to follow-up CT include nodule size at detection, type of scan (baseline or annual repeat) on which the nodule is detected, and presence of patient-induced artifacts.
© RSNA, 2004
Index terms: Cancer screening • Computed tomography (CT), artifact, 60.93 • Computed tomography (CT), thin-section, 60.12118 • Computed tomography (CT), three-dimensional, 60.12117 • Computed tomography (CT), volume rendering • Lung, nodule, 60.332 • Lung neoplasms, screening
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
One of the most compelling indicators of nodule malignancy is growth. The concept of estimation of nodule growth rate, expressed as doubling time, was introduced nearly 50 years ago (3,4). In these early studies, the doubling time was estimated on the basis of manual estimates of nodule diameter on chest radiographs. This method of growth rate estimation became the de facto standard for the evaluation of lesions that were found on chest radiographs and were suspected of being malignant. With the development of CT, a similar approach was taken; the CT section containing the largest cross section of the nodule was measured with physical or electronic calipers.
Early research on the application of computer vision techniques to the analysis of pulmonary nodule growth began with measurements similar to those of traditional evaluation. On a single CT section, automated techniques were used to make two-dimensional measurements, including cross-sectional area (5). This work was followed by a report of methods for the three-dimensional estimation of nodule volume and doubling time (6). In that study (6), researchers determined that the volume of synthetic nodule phantoms as small as 3 mm in diameter could be measured to within 3% accuracy. These techniques also had the potential to be used to distinguish malignant from nonmalignant nodules, and these determinations could possibly be made in relatively short intervals (6).
There is now widespread interest in the use of techniques for volumetric analysis of nodules, both in academic practice and in industry. Although the relative error in nodule volume measurement as a function of nodule size had been quantified by using nodule phantoms (6), the error for in vivo nodules must be greater, as it includes measurement error due to greater partial-volume effects, vascular geometry, and motion artifacts. Of primary concern are whether a nodule that appears to have grown in sequential volume measurements actually has grown and whether the difference in these measurements is not simply due to error. Thus, the purpose of our study was to determine the reproducibility of volume measurements of small pulmonary nodules on CT scans and to estimate the critical time to follow-up CT.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Imaging and Nodule Selection
We reviewed cases from our lung cancer screening program (1,2) in which thin-section CT scans were obtained at least twice. Informed consent for analysis of the CT scans was obtained from each of the subjects with a study protocol approved by our institutional review board. The CT scans were acquired with single–detector row (HiSpeed Advantage and CTi; GE Medical Systems, Milwaukee, Wis) and multi–detector row (LightSpeed; GE Medical Systems) CT scanners with 140 kVp, 200 mA, 1.0–1.25-mm section thickness, and a small field of view (9.6 cm), yielding maximum in-plane resolution of 0.1875 mm. The scans were reconstructed by using a high-spatial-frequency algorithm. As dictated by our original study protocol, these thin-section CT scans were obtained with standard-dose parameters after initial identification of the nodules on low-dose full-lung screening CT scans. At that time, standard-dose parameters were used to obtain the best possible thin-section scans. We are now using multi–detector row CT, and thus our protocol has changed to specify the use of low-dose parameters for scanning of all pulmonary nodules. Thin-section small–field-of-view images of individual nodules are now typically reconstructed from the raw data obtained during a low-dose full-lung screening examination.
We selected stable solid nodules that did not abut the pleura and thus excluded both juxtapleural and subsolid nodules because they present special challenges. Stability was defined as no increase or decrease in the size of the nodule for more than 2 years. Experienced thoracic radiologists (C.I.H., D.F.Y.) with 22 and 18 years experience, respectively, ascertained stability. To do so, they measured the average diameter (average of maximum length and maximum perpendicular width of the nodule on the two-dimensional image of maximum cross-sectional area) with electronic calipers on each of two thin-section CT scans.
A large number of nodules (n = 225) smaller than 5 mm in diameter met the inclusion criteria, and thus we randomly picked one-third of them (n = 75) by using a computer-based random-number generator. Because there were far fewer nodules larger than 5 mm (n = 45) that met the inclusion criteria, we included all of them. Initially, 120 stable nodules with 240 scans were available for the study. We excluded five (2%) of the scans because of technical artifacts; these scans included four in which the entire nodule volume was not acquired and one in which a system error resulted in missing data. Each of these corresponded to one of the two scans of five individual cases, and this process resulted in a reduction in the number of scans by 10 to a total of 230. Thus, there were 115 stable nodules that were evaluated for this study. These nodules were obtained from 115 subjects (59 men, 56 women) ranging in age from 43 to 83 years (median, 63 years).
We grouped the nodules in three categories according to initial diameter: 2–5 mm (range, 
2 to <5 mm), 5–8 mm (range, 5 to <8 mm), and 8–10 mm (range, 8 to <10 mm). With these categories, there were 72 2–5-mm nodules (range, 2.1–4.9 mm; median, 3.7 mm), 34 5–8-mm nodules (range, 5.1–7.8 mm; median, 5.8 mm), and nine 8–10-mm nodules (range, 8.1–9.5 mm; median, 8.3 mm).
Nodule Evaluation
The initial and follow-up scans of each nodule were visually assessed by one of the radiologists (D.F.Y.) to determine the presence of patient-induced artifacts that would prohibit accurate estimates of nodule volume. Patient-induced artifacts occurred when there was gross patient movement, respiratory motion, or cardiac motion such that the nodule volume was distorted (Fig 1). Each of these motion artifacts was graded by the radiologist on a five-point scale (none, minimal, moderate, pronounced, severe). With these scores, we grouped the cases for study into three overlapping sets: standard, expanded, and complete. The standard group included those cases for which artifacts on the scans were none or minimal, and this group represented the majority of cases (94 nodules) or those that typically would be seen at screening. The expanded group (105 nodules) included the cases in the standard group plus those for which the artifacts on scans were moderate. Although these artifacts were plainly visible to the radiologist, we saw value in performing an evaluation of an expanded group of cases that included scans with artifacts of greater severity that might require analysis when rescanning of the patient was not possible. The complete group (115 nodules) comprised all cases and included those with scans that had pronounced and severe artifacts that would certainly produce erroneous volume estimates.
|
|
For each case, we computed the percentage volume change (PVC) with
|
To assess the variability of growth rate estimation of stable nodules induced by the variability in volume estimation, we calculated the monthly volumetric growth index (MVGI), which is used to quantify the percentage change in nodule volume per month (10). It is defined as follows:
|
|
t days apart (11). Alternatively, it can be expressed directly in terms of the ratio of V2 to V1 and the time between scans thus:
|
We wanted to determine the critical time to perform follow-up CT as a function of the initial nodule size. We defined the critical time to be the earliest point at which we can reliably identify growth in a nodule of a given size with repeat CT. This time interval is a function of 
(d); the reliably detectable PVC; and DTT, a doubling time threshold between growing and stable nodules. The parameter (d) is a function of nodule diameter, d, and exhibits an inverse relationship to it, as observed in phantom experiments (6). Since (d) is larger for smaller nodule diameters, the time to follow-up CT for smaller nodules must be longer. Rewriting Equation (3) and solving for the critical time to follow-up CT, ti, we found that
|
We used 2 SDs of PVC as our estimate of (d), as it would include any change in the volume of the majority of stable nodules.
The value of DTT in Equation (5) is influenced by the conditions under which the nodule was detected. For nodules detected at baseline screening or incidentally at CT, we used a DTT value of 400 days. This value was chosen because it corresponded approximately to the doubling time of the slowest-growing lung cancers in several published series (12–14). A lower DTT was selected for nodules that are found at annual repeat screening after a negative prior screening (baseline or otherwise) since these nodules are more likely to be malignant, as they have already demonstrated growth during the intervening year and as a group are growing faster. An estimate of the growth rate of a nodule first detected at annual repeat screening (ie, a nodule not identifiable on the CT scan 1 year prior to detection) may be determined with the assumption that the smallest nodules that can be reliably seen on modern multi–detector row CT scans are approximately 2 mm in diameter. By using this visibility threshold, we can determine the upper bound on their doubling time (the lower bound on their growth rate). For example, if a nodule is 3 mm in diameter but was not visible on a scan obtained 1 year prior to its detection, we can determine that the upper bound on the doubling time would be 208 days, with the two-dimensional analogue of Equation (3) as follows:
|
t is 365.25, D2 is 3, and D1 is at most 2, DT could be at most 208 days. For a nodule larger than 3 mm, the doubling time would be shorter, as it would have to grow faster to progress from being undetectable to its larger size at detection. Thus, at annual repeat screening, as distinct from baseline screening, the value of DTT would be approximately 200 days for 3-mm nodules, 90 days for 5-mm nodules, and 60 days for 8-mm nodules.
Data Analysis
We calculated the mean and SD of PVC values for each category with initial nodule size and time to follow-up CT. We also calculated the mean and SD of MVGI for each category with initial size and time to follow-up CT. We determined the critical time to follow-up CT for nodules of varying size by using empirical estimates of (d) derived from the distribution of PVC values observed in our experiments.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
|
|
|
Time to Follow-up
With the methods described, we found that the critical time to follow-up CT (when change in nodule volume could be determined) for nodules detected at baseline screening or those detected at CT performed for other reasons, was 12 months for those with an initial diameter of 2 mm, 5 months for those with an initial diameter of 5 mm, 3 months for those with an initial diameter of 8 mm, and 1 month for those with an initial diameter of 10 mm. For nodules detected at annual repeat screening, we determined the critical times to be 4 months for 3-mm nodules, 3 months for 4-mm nodules, and 1 month for nodules 5 mm or larger in diameter.
In comparison, when we considered the expanded group of 115 nodules, the estimate of the critical time to follow-up CT increased. For nodules detected at baseline screening, the estimate of critical time to follow-up was 8 months for those with an initial diameter smaller than 5 mm and 6 months for those larger than 5 mm in diameter. For nodules detected at annual repeat screening, the critical times to follow-up increased to 5 months for nodules smaller than 5 mm in diameter and to 2 months for those 5 mm in diameter or larger.
MVGI Data
As expected for these stable nodules, the overall mean growth rate in the standard group was essentially zero (0.06%), and the standard error was 0.21%. The SD of MVGI was inversely related to nodule size, from 2.21 in 2–5-mm nodules to 1.68 in 5–8-mm nodules to 1.23 in 8–10-mm nodules. Unlike that of PVC, the SD of MVGI decreased with increasing time to follow-up CT (2.22 for 0–6 months, 2.04 for 6–12 months, and 1.41 for 12–30 months) (Table 4). All of the stable nodules had a rate of growth that was less than 2 SDs above the corresponding mean value for their size category: 4.4% for 2–5-mm nodules, 3.6% for 5–8-mm nodules, and 2.3% for 8–10-mm nodules.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
The time to follow-up CT is one of the factors that affects the accuracy of growth rate estimation. For nodules detected at baseline screening, follow-up CT could be performed in as short an interval as 3 months. The critical time to follow-up CT for a nodule detected at annual repeat screening should be shorter than that for a nodule detected at baseline screening (given that a higher proportion of these nodules have a faster growth rate) and was found to be as short as 1 month for nodules larger than 5 mm in diameter. Since the critical time to follow-up CT is a function of the nodule size on the initial scan, follow-up of smaller nodules should necessarily require a longer delay than that for larger nodules, unless there is other clinical information that indicates a high likelihood of malignancy. These results are in accord with those used in current CT screening protocols (15).
When we examined the effect of increased artifact on the reproducibility of volume measurements (by using the expanded group), we found that the SD of PVC increased by 50%–60% in nodules between 2 and 8 mm in diameter. In addition, although we found a greater increase for nodules between 8 and 10 mm in diameter, the sample size was too small for the effect to be assessed reliably. Given these increases in variability of volume estimation and the corresponding increases in the value of (d), we determined that the time to follow-up CT would be 2–3 months longer when moderate motion artifacts are present, and these would be easily identifiable by the radiologist. This finding may be of value when the time to follow-up CT is determined for a nodule that was initially scanned with a higher degree of motion artifact caused by patient movement or one in which the effect of cardiac motion would typically be great.
We also evaluated the apparent growth rates in these stable nodules induced by both variability of our measurement techniques and any small but true change in the nodule size. We found that MVGI improves in accuracy with increasing time between acquisition of scans. As a measure of growth rate, MVGI can be used to standardize the volume change according to the time between scans and should be, in principle, independent of this parameter (t) if the nodule volumes are determined perfectly without error. In practice, however, an MVGI estimate actually improves with increasing time between acquisition of scans because its variance is reduced as this time interval between acquisition of scans increases.
In this study, we examined the use of PVC to determine the measured percentage change in the volume of stable nodules. The measurement error was largely quantified by using phantom experiments in which it was determined that nodules of 3–10 mm in diameter could be reproducibly measured with an SD of less than 3% and that there was a reduction in this error with increasing nodule size (6). The larger SD of PVC values in stable nodules seen in the current study was caused by the addition of a number of other sources of error, including but not limited to motion artifacts, segmentation issues, and small but potentially real increases or decreases in nodule volume. We observed that the SD of PVC in stable nodules increased with increasing time to follow-up CT, which might partially be explained by the latter source of error.
The decrease in the SD of MVGI and PVC with increasing nodule size was in part due to the increasing accuracy of nodule segmentation for larger nodules. For smaller nodules, the accuracy is compromised by the larger proportion of voxels that are on the surface of the nodule and are thus affected by partial-volume effect (6,9). A less important but competing factor for larger nodules is that there may be some difficulty with the segmentation of the nodule from its surrounding larger local vasculature that contributes to measurement error (9).
Pitfalls in growth rate determination, however, should be recognized. We found that with decreasing nodule size, there was an increase in motion artifacts that prohibited accurate size assessment. This relationship appears to be caused by the fact that for an artifact of a given magnitude, the apparent proportional effect is smaller with increasing nodule size. Thus, while a small degree of motion might not appreciably affect the scan quality for a larger nodule, it might have a dramatic effect on that for a smaller nodule. Because such artifacts, whether on the initial or follow-up CT scans, affect growth analysis, the accuracy of the growth rate estimate is a function of the quality of both scans. Although patient-induced artifacts will continue to decrease in importance with increasing scanner acquisition speed, training of CT technicians and careful data management policies should be practiced to keep technical artifacts (eg, incomplete acquisition of the entire nodule volume, data loss due to hardware or software error) to a minimum. In the future, automated computer methods should be used to analyze the scans as they are obtained and to detect the presence of such artifacts. In the event that an artifact is detected, the operator could be alerted and given the option of rescanning the nodule. Gating, however, ultimately may be necessary to compensate for cardiac motion.
There are further challenges that need to be addressed before these methods become applicable to all types of nodules. Subsolid nodules (both nonsolid and part solid) pose special challenges to analysis of growth (16). Delineation of the boundaries between nonsolid components and either the lung parenchyma or any solid components present (due to reduced image contrast) is more difficult than segmentation of solid nodules, and therefore, work on robust segmentation methods for these nodules is still ongoing. Research is now also being performed in regard to methods to allow the volume change of the solid and nonsolid components of part-solid nodules to be quantified individually (17). This quantification involves the development of new segmentation methods that allow delineation of the boundaries of these components in an automated way. Juxtapleural nodules also pose a challenge to the assessment of nodule growth. Definition of the interface between the nodule and the pleural surface is more difficult than definition of that between the nodule and vascular structures, because the larger proportion of the nodule surface typically abuts the pleura. Methods are being developed to model the pleural surface and thereby better determine the boundary between the nodule and the pleura (9).
As noted in Materials and Methods, the key limitation of this study was that the times at which the two scans of each nodule were acquired were separated in time. Perhaps a better measure of reproducibility of nodule volume measurements would be to scan the same nodule twice in short succession, but this was not possible because it was not part of our protocol and because of concerns about radiation exposure to patients. However, we followed cases of each nodule for a minimum of 2 years, established size stability during that interval with measurement by experienced thoracic radiologists, and then compared two scans that were obtained typically much closer chronologically (to minimize the effect of any true volume change in the nodule over time). Another limitation was that the critical time to follow-up CT was determined in this study by using standard-dose and thin-section CT parameters. Newer screening protocols now typically rely on low-dose parameters, and this would likely increase the critical time for nodules of a given size due to increased image noise. In general, screening protocol development is influenced by factors beyond just the critical time to follow-up CT (eg, prevalence of a given finding in the population under study).
In conclusion, three-dimensional computer methods can be used to assess stability and characterize growth in small solid pulmonary nodules. Factors that affect the reproducibility of nodule volume measurements and the critical time to follow-up CT include the nodule size at detection, the type of scan (baseline or annual repeat) on which the nodule is detected, and the presence of patient-induced artifacts.
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, W.J.K., D.F.Y., C.I.H.; study concepts and design, W.J.K., D.F.Y., A.P.R., C.I.H.; literature research, W.J.K., S.C.F.; clinical studies, D.F.Y., C.I.H.; experimental studies, W.J.K., D.F.Y., S.C.F., C.I.H.; data acquisition, W.J.K., D.F.Y., S.C.F., C.I.H.; data analysis/interpretation, W.J.K., D.F.Y., C.I.H.; statistical analysis, W.J.K., C.I.H.; manuscript preparation, W.J.K., D.F.Y., C.I.H.; manuscript definition of intellectual content, W.J.K., D.F.Y., A.P.R., C.I.H.; manuscript editing, W.J.K., D.F.Y., C.I.H.; manuscript revision/review and final version approval, all authors
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  Y. Wang, R. J. van Klaveren, H. J. van der Zaag-Loonen, G. H. de Bock, H. A. Gietema, D. M. Xu, A. L. M. Leusveld, H. J. de Koning, E. T. Scholten, J. Verschakelen, et al. Effect of Nodule Characteristics on Variability of Semiautomated Volume Measurements in Pulmonary Nodules Detected in a Lung Cancer Screening Program Radiology, August 1, 2008; 248(2): 625 - 631. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Jeudy, C. S. White, R. F. Munden, and P. M. Boiselle Management of Small (3-5-mm) Pulmonary Nodules at Chest CT: Global Survey of Thoracic Radiologists Radiology, June 1, 2008; 247(3): 847 - 853. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. G. Ravenel, W. M. Leue, P. J. Nietert, J. V. Miller, K. K. Taylor, and G. A. Silvestri Pulmonary Nodule Volume: Effects of Reconstruction Parameters on Automated Measurements--A Phantom Study Radiology, May 1, 2008; 247(2): 400 - 408. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. I. Henschke and D. F. Yankelevitz CT Screening for Lung Cancer: Update 2007 Oncologist, January 1, 2008; 13(1): 65 - 78. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. A. Gietema, C. M. Schaefer-Prokop, W. P. T. M. Mali, G. Groenewegen, and M. Prokop Pulmonary Nodules: Interscan Variability of Semiautomated Volume Measurements with Multisection CT Influence of Inspiration Level, Nodule Size, and Segmentation Performance Radiology, December 1, 2007; 245(3): 888 - 894. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. S. Gierada, T. K. Pilgram, M. Ford, R. M. Fagerstrom, T. R. Church, H. Nath, K. Garg, and D. C. Strollo Lung Cancer: Interobserver Agreement on Interpretation of Pulmonary Findings at Low-Dose CT Screening Radiology, December 1, 2007; 246(1): 265 - 272. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Aisner CT Screening for Lung Cancer: Are We Ready for Wide-Scale Application? Clin. Cancer Res., September 1, 2007; 13(17): 4951 - 4953. [Full Text] [PDF]
|
|
|
|
|
|
New York Early Lung Cancer Action Project Investig CT Screening for Lung Cancer: Diagnoses Resulting from the New York Early Lung Cancer Action Project Radiology, April 1, 2007; 243(1): 239 - 249. [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]
|
|
|
|
|
|
S. G. Jennings, H. T. Winer-Muram, M. Tann, J. Ying, and I. Dowdeswell Distribution of Stage I Lung Cancer Growth Rates Determined with Serial Volumetric CT Measurements Radiology, November 1, 2006; 241(2): 554 - 563. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 The International Early Lung Cancer Action Program Survival of Patients with Stage I Lung Cancer Detected on CT Screening N. Engl. J. Med., October 26, 2006; 355(17): 1763 - 1771. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. A. Gietema, Y. Wang, D. Xu, R. J. van Klaveren, H. de Koning, E. Scholten, J. Verschakelen, G. Kohl, M. Oudkerk, and M. Prokop Pulmonary Nodules Detected at Lung Cancer Screening: Interobserver Variability of Semiautomated Volume Measurements Radiology, October 1, 2006; 241(1): 251 - 257. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Plathow, M. Schoebinger, C. Fink, H. Hof, J. Debus, H.-P. Meinzer, and H.-U. Kauczor Quantification of Lung Tumor Volume and Rotation at 3D Dynamic Parallel MR Imaging with View Sharing: Preliminary Results Radiology, August 1, 2006; 240(2): 537 - 545. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. J. Pickhardt, V. T. Lehman, T. C. Winter, and A. J. Taylor Polyp volume versus linear size measurements at CT colonography: implications for noninvasive surveillance of unresected colorectal lesions. Am. J. Roentgenol., June 1, 2006; 186(6): 1605 - 1610. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. P. Ko, R. Marcus, E. Bomsztyk, J. S. Babb, C. Stefanescu, M. Kaur, D. P. Naidich, and H. Rusinek Effect of blood vessels on measurement of nodule volume in a chest phantom. Radiology, April 1, 2006; 239(1): 79 - 85. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Awai, K. Murao, A. Ozawa, Y. Nakayama, T. Nakaura, D. Liu, K. Kawanaka, Y. Funama, S. Morishita, and Y. Yamashita Pulmonary Nodules: Estimation of Malignancy at Thin-Section Helical CT--Effect of Computer-aided Diagnosis on Performance of Radiologists Radiology, April 1, 2006; 239(1): 276 - 284. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. MacMahon, J. H. M. Austin, G. Gamsu, C. J. Herold, J. R. Jett, D. P. Naidich, E. F. Patz Jr, and S. J. Swensen Guidelines for Management of Small Pulmonary Nodules Detected on CT Scans: A Statement from the Fleischner Society Radiology, November 1, 2005; 237(2): 395 - 400. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. P. Ko, J. Chang, E. Bomsztyk, J. S. Babb, D. P. Naidich, and H. Rusinek Effect of CT Image Compression on Computer-assisted Lung Nodule Volume Measurement Radiology, October 1, 2005; 237(1): 83 - 88. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. Mulshine Clinical Issues in the Management of Early Lung Cancer Clin. Cancer Res., July 1, 2005; 11(13): 4993s - 4998s. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
