When you have The UPS Store pack and ship your items you get the benefit of The UPS Store Pack & Ship Guarantee. When we pack and ship your items using materials purchased from The UPS Store, we'll cover the cost of packing and shipping plus the value of your items, if lost or damaged*.
When it comes to cushioning and protecting your shipments, there's nothing The UPS Store Certified Packing Experts can't handle. We are trained in advanced packing techniques and specialize in properly packing fragile, high-value, large and odd-shaped items. If you're looking to ship electronics, artwork, antiques or luggage.
Volumetric measurements are usually based on segmentation of nodules on thin-section CT data sets and an algorithm that translates the segmented voxels into a nodule volume. For synthetic nodules, volumetry with various commercially available software packages proved to be very accurate . In vivo accuracy, however, is less due to less sharply defined nodule borders, motion effects and complex geometry of adjacent structures. Earlier studies showed that a large degree of the interexamination variability of nodule volumetry can be explained by segmentation errors [4, 5], which were found to be common in irregularly shaped nodules. Currently, several software packages have been developed that claim to be able to adequately segment nodules irrespectively of size, shape and location. Variation in the results of volumetry may result in false-positive or false-negative conclusions with potential serious consequences for the patient. To avoid overinterpreting random changes in volumetric measurements, the diagnosis of real growth or regression typically requires that the difference in measured nodule size exceeds the upper limit of agreement. This upper limit of agreement, however, may be dependent on the software package used.
In this study, we simulated the situation of a baseline and follow-up chest exam on which a suspicious nodule was detected and subsequently evaluated by commercially available software for the presence of change. In order to determine the minimum amount of change needed to be 95% sure that the change was due to a real growth, we determined the interexamination variability with patients scanned twice in the same session and moved on and off the table. This variability was influenced by both patient factors (e.g., inspiration depth) and the software algorithm . Since all software packages had to segment the same set of nodules, the approach we chose allows for determining a realistic number of the amount of change needed that incorporates all factors seen in clinical routine.
The objective of this study was to compare the interexamination variability of pulmonary nodule volumetry on repeat CTs using six currently available semi-automated software packages for CT. We used a dataset containing nodules of varying size, morphology and contact to pulmonary structures in order to define the upper limit of agreement for each software package and to determine whether these software packages could be used interchangeably.
All solid lesions with a minimum volume of 15 mm3 (corresponding to a diameter of about 3 mm) were included. Lung masses, defined as nodules exceeding 30 mm in diameter, were excluded from analysis. Nodules suspected of being metastases were included, as well as nodules that could potentially have a benign histology. Completely calcified nodules, however, were excluded. Only solid nodules were included since non-solid or partly solid nodules require a different segmentation approach, and not all of the evaluated segmentation software packages were developed for this task. A maximum of 50 nodules per patient were included.
The following segmentation algorithms were evaluated: Advantage ALA (GE, v7.4.63), Extended Brilliance Workspace (Philips, EBW v3.0), Lungcare I (Siemens, Somaris 5 VB 10A-W), Lungcare II (Siemens, Somaris 5 VE31H), OncoTreat (MEVIS, v1.6), and Vitrea (Vital images, v3.8.1, lung nodule evaluation add-on included). For the purpose of anonymization, the characters A to F were randomly assigned to the various packages.
In all algorithms, segmentation was initiated by clicking in the center of a nodule, starting a fully automated evaluation of the nodule. Next, all algorithms segment the nodule, calculate its volume and present the result. The segmented area was shown by the various software packages by a thin line surrounding the area of the nodule or by a colored overlay. This segmentation was visually judged for accuracy. In order to minimize observer influence, only these automated results were used for comparisons in this study, except for when explicitly written otherwise.
We did a separate analysis for results obtained using manual correction of incomplete segmentations. Manual correction by the user was allowed to correct the segmentation by four of six packages. In case of a mismatch between nodule and segmentation, this feature was used to obtain the most precise segmentation feasible. The type of manual correction varied between the packages (Table 1). Two packages also allowed for a complete manual segmentation in case of failure; this feature was not used. Next, the segmentation was again visually judged for accuracy.
The histogram of relative differences showed a normal distribution for all packages (tested with the Kolmogorov-Smirnov test). Because the same nodule was measured twice on successive chest CTs, a mean relative difference close to 0 can be expected. In fact, none of the packages had a mean relative difference higher than 1.1%. We therefore decided to use only the upper limit of agreement of the 95% CI of the relative differences as assessed according to the method proposed by Bland and Altman  as the measure of interexamination variability. An increase in nodule volume above this upper limit of agreement can, with 95% confidence, be attributed to real growth.
We also tested whether there was a significant difference in interexamination variability between excellently and satisfactorily segmented nodules. For each software package separately, an F-test was used to compare interexamination variability for all those nodules that were classified as excellently or satisfactorily segmented with this specific software.
In order to detect systematic differences in measured volumes between packages, we performed a mixed model variance analysis of nodule volumes on a subset of nodules that were adequately segmented by all programs.
Eighty-nine (42%) of all nodules were adequately segmented by all packages. In this dataset, interexamination variability of software packages B and D was significantly lower than that of package C, E and F. Extent of variability of package C also differed significantly from package A (Table 2).
The overall measurement variability of the software packages before and after manual correction is given in Fig. 4. For each individual software package, Fig. 4 shows the variability for all nodules adequately segmented by this particular software package. Note that the numbers differ from those in Table 2 because for each package all adequately segmented nodules were included and not only the subset of 89 nodules that were adequately segmented by all software packages. Compared to the automated results, the upper limit of agreement did not change significantly after manual correction for any of the software packages.
Comparison of overall variability in repeated CT examinations for the various software packages, without (a) and after (b) manual adjustments. For each package only those nodules were considered for which segmentation was visually rated as adequate (excellent or satisfactory on both scans). The percentage of nodules included in these calculations therefore varied between (a) and (b), and per software package (see Fig. 3). Overall interexamination variability is expressed as the upper limit of the 95% CI for the relative differences between scan 1 and 2. a Variability in repeated CT examinations (without manual adjustment). b Variability in repeated CT examinations (after manual adjustments)
In the dataset of nodules that were adequately segmented by all packages, mean nodule volume per software packages is given in Table 2. The mixed model variance analysis showed significant systematic differences in mean volumes in 11 of the 15 (6*5/2) possible pairs of software packages.
In this study we show substantial differences in segmentation performance among six currently available pulmonary nodule segmentation software packages in a dataset of nodules with a variety in size, morphology and contact to pulmonary structures. The best package segmented 86% of all nodules in both the first and the second examination with excellent or satisfactory accuracy. All software packages showed similar interexamination variability, but there were significant differences in absolute nodule volumes between software packages. Manual correction substantially improved the number of accurate segmentations without significantly affecting reproducibility.
High segmentation accuracy is a prerequisite for adequate performance of nodule volumetry software. It is obvious that segmentations that include surrounding structures or do not include part of a nodule may lead to wrong management decisions. We found substantial variations in segmentation accuracy between software packages.
We also found that the differences in variability were comparatively small when excellent and satisfactory (small segmentation errors included) segmentations were compared. While there were significant differences between excellent and satisfactory segmentations in some packages, these differences were comparatively small and in the order of the differences between software packages.
There was, however, a significant difference in absolute nodule volumes among software packages. This can lead to variations in management decisions: Software X and Y will measure higher volume (and thus, effective diameter) than software Z, for example, and will therefore induce more aggressive management decisions if recommendations by the Fleischner society are followed. . Consequently, the nodule shown in Fig. 5 would be treated differently depending on the segmentation algorithm used. As shown in Table 2, the size changes required to detect significant growth may be substantially greater when different software packages are used for baseline and follow-up evaluation instead of the same package for both examinations. In addition, there is a substantial bias, which means that the systematic differences in mean volumes can lead to a situation in which a growing nodule will appear to have shrunk (e.g., first measurement with software F, second measurement with software E) or a stable nodule will appear to have grown (e.g., first measurement with software E, second measurement with software F). 041b061a72