Home Comparison of the Accuracy of CT Volume Calculated by Circumscription to Prolate Ellipsoid Volume (Bidimensional Measurement Multiplied by Coronal Long Axis)
Post
Cancel

Comparison of the Accuracy of CT Volume Calculated by Circumscription to Prolate Ellipsoid Volume (Bidimensional Measurement Multiplied by Coronal Long Axis)

Posted Jan 31, 2009 Updated Apr 7, 2023
By Ali M. Rkein MSII , Chivonne Harrigal MD , Arnold C. Friedman MD , Daniel Persky MD , Elizabeth Krupinski PhD
4 min read

Rationale and Objectives

Tumor volume is one of the most important factors in evaluating the response to treatment of patients with cancer. The objective of this study was to compare computed tomographic (CT) volume calculation using a semiautomated circumscribing tracing tool (manual circumscription [MC]) to prolate ellipsoid volume calculation (PEVC; bidimensional measurement multiplied by coronal long axis) and determine which was more accurate and consistent.

Materials and Methods

The study included six patients with nine neoplasms, six phantoms, and two radiologists. The neoplasms and phantoms of varying sizes and shapes were imaged using multidetector CT scanners, with slice thicknesses ranging from 0.5 to 3 mm. Measurements were performed using a TeraRecon 3D workstation. Each lesion and phantom was manually circumscribed, and its three dimensions were measured. The measurements were repeated 2 weeks later.

Results

MC of the phantoms deviated from their true volumes by an average of 3.0 ± 1%, whereas PEVC deviated by 10.1 ± 3.99%. MC interobserver readings varied by 1.2 ± 0.6% and PEVC by 4.8 ± 3.3%. MC intraobserver readings varied by 1.95 ± 1.75% and PEVC by 2.5 ± 1.55%. Patient tumor volume predicted by MC and PEVC varied greatly; MC interobserver readings differed by 3.3 ± 2.1% and PEVC by 20.1 ± 10.6%. MC intraobserver readings varied by 2.5 ± 1.9% and PEVC by 5.5 ± 3.2%. Variability was greater for complex shapes than for simple shapes. Bidimensional analysis demonstrated an interobserver difference of 12.1 ± 8.7% and an intraobserver difference of 5.05 ± 3.3%. These results demonstrate large interobserver and intraobserver variability. Variability was greater for complex shapes than for simple shapes.

Conclusion

MC of neoplasms provided more accurate and consistent volume predictions than PEVC. More complicated shapes demonstrated the superiority of MC over PEVC.

Evaluating initial mass size and changes in mass size over time to differentiate malignancies from benign masses and evaluate the response to therapy of known neoplasms is often accomplished by performing bidimensional measurements. For example, lung nodules ≤ 4 mm are generally considered benign, whereas nodules > 8 mm bear a substantial risk for malignancy ( ).

Current clinical practice relies on diameter measurement as a surrogate for volume when evaluating the response to therapy of patients with cancer. Small inconsistencies in diameter measurement cause large changes in apparent volume (ie, a 25% increase in diameter results in a doubling of the volume). Growth rate can also be used to differentiate malignant masses from benign ones. The doubling time of malignant tumors generally ranges from 30 to 500 days, with a median of 100 days ( ).

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Materials and methods

Get Radiology Tree app to read full this article<

Figure 1, Bland-Altman plot comparing the accuracy of prolate ellipsoid volume calculation (PEVC) versus manual circumscription (MC) to the true phantom volumes for radiologist 1. SD, standard deviation.

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Results

Get Radiology Tree app to read full this article<

Figure 2, Bland-Altman plot comparing the accuracy of prolate ellipsoid volume calculation (PEVC) versus manual circumscription (MC) to the true phantom volumes for radiologist 2. SD, standard deviation.

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Figure 3, Percentage difference between the measurements completed by the radiologists and the true phantom volumes. Avg., average; MC, manual circumscription; PEVC, prolate ellipsoid volume calculation; Rad, radiologist.

Figure 4, Analysis of interobserver difference between the measurements conducted by the two radiologists of the phantom volumes. MC, manual circumscription; PEVC, prolate ellipsoid volume calculation.

Figure 5, Analysis of intraobserver measurements conducted by the two radiologists on the phantom volumes. MC, manual circumscription; PEVC, prolate ellipsoid volume calculation; Rad., radiologist.

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Figure 6, Depiction of a small clinical nodule. (Left) Cross-sectional slice with bidimensional measurements; (right) coronal image with height measurement.

Figure 7, Screenshot of the automated circumscribing tool workstation.

Figure 8, Automated circumscribing tool approximating the nodule.

Figure 9, Screenshot after manual circumscription. Only the nodule is circumscribed; the lung parenchyma and pleura and blood vessels are not circumscribed.

Figure 10, Clinical case demonstrating complex-shaped retroperitoneal adenopathy circumscribing the aorta. (Top left) Transverse scan demonstrating the bidimensional measurements; (bottom left) coronal image demonstrating the height.

Figure 11, The complex-shaped adenopathy shown in Figure 10 after manual circumscription. Only the adenopathy is circumscribed; the aorta is excluded.

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Figure 12, Bland-Altman plot comparing the interobserver volume measurements conducted on the patient neoplasms using prolate ellipsoid volume calculation (PEVC) versus manual circumscription (MC).

Figure 13, Analysis of interobserver difference of predicted volume by prolate ellipsoid volume calculation (PEVC) and manual circumscription (MC) by the two radiologists. Rad., radiologist.

Figure 14, Analysis of intraobserver difference of predicted volume by prolate ellipsoid volume calculation (PEVC) and manual circumscription (MC) by the two radiologists. Rad., radiologist.

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Figure 15, Intraobserver analysis of bidimensional (Bi-Dim) measurements of the neoplasms. Rad., radiologist.

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Discussion

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Conclusion

Get Radiology Tree app to read full this article<

References

  • 1. MacMahon H., Austin J.H.M., Gamsu G., et. al.: Guidelines for management of small pulmonary nodules detected on CT scans: a statement from the Fleischner Society. Radiology 2005; 237: pp. 395-400.

  • 2. Gurney J.W.: Doubling time. http://www.chestx-ray.com/SPN/DoublingTime.html Accessed June 2007

  • 3. Bolte H., Jahnke T., Schafer F.K., et. al.: Interobserver-variability of lung nodule volumetry considering different segmentation algorithms and observer training levels. Eur J Radiol 2007; 64: pp. 285-295.

  • 4. Cheson B.D., Pfistner B., Juweid M.E., et. al.: Revised response criteria for malignant lymphoma. J Clin Oncol 2007; 25: pp. 579-586.

Academic Radiology, Volume 16, Issue 2
Journals General Radiology
This post is licensed under CC BY 4.0 by the author.
Recently Updated
Contents