HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): Ichiro Shimada Stephen J. Rooney Robert S. Bonser Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Shimada, I. Articles by Bonser, R. S. Search for Related Content PubMed PubMed Citation Articles by Shimada, I. Articles by Bonser, R. S.

Eur J Cardiothorac Surg 1999;16:59-62
© 1999 Elsevier Science NL

Reproducibility of thoracic aortic diameter measurement using computed tomographic scans

^a Cardiothoracic Surgical Unit, Queen Elizabeth Hospital, Edgbaston, Birmingham B15 2TH, UK
^b Department of Radiology, Queen Elizabeth Hospital, Edgbaston, Birmingham B15 2TH, UK
^c Faculty of Mathematics, University of Birmingham, Birmingham, UK

Corresponding author. Tel.: +44-121-6272559; fax: +44-121-6272542
e-mail: r.s.bonser{at}bham.ac.uk

	Abstract

Top Abstract 1. Introduction 2. Materials and methods 3. Results 4. Discussion References

 
Objectives: Decisions to recommend elective surgical repair of thoracic aortic aneurysms (TAA) may be based on size or expansion rate, which are used as indices of the risk of rupture. Measurement error may thus affect clinical decision-making. In order to evaluate the reproducibility of aortic diameter measurements of TAA, we assessed departmental inter- and intra-observer variability of measurement of pre-selected computed tomographic scan images of aneurysmal segments of the thoracic aorta. Methods: We compared measurements of minimum aortic diameter made by four observers in 50 pre-selected scans and at different times by two observers using a calliper method and a measurement tool within the scan. Differences in measured dimension were analysed using Wilcoxon's signed ranks test and the repeatability assessed using the method of Bland and Altman. Results: There were no significant inter-observer differences among three observers but there were significant differences between another observer and two other observers (P<0.05). No significant intra-observer differences existed. The best intra-observer repeatability was 2.25 while the worst inter-observer repeatability was 4.37. The mean and maximum difference in measurement were ±0.88 mm and ±8.0 mm, respectively. Variability of measurement increased with aortic diameter. Conclusions: Calliper measurement of TAA is an acceptable measurement method for surveillance of TAA but appears most accurate with a single observer. Increasing error is seen with increasing diameter which may compound error in estimation of expansion rate. Standardisation of technique is advisable for multiple observers and aortic units should adopt quality assurance protocols to minimise error.

Key Words: Thoracic aortic aneurysm • Aortic diameter • Reproducibility

	1. Introduction

Top Abstract 1. Introduction 2. Materials and methods 3. Results 4. Discussion References

 
The cross-sectional diameter of aortic aneurysms, measured using computed tomographic (CT) scans, is the most important determinant of risk of rupture and is the key index determining when elective repair is justified [1,2]. Variation of measurement, which has been shown to be important in the ultrasound follow-up of small abdominal aortic aneurysms (AAAs) [3], may affect clinical decisions when these are based on actual diameter or demonstration of accelerated growth. The aorta is a 3-dimensional structure and aneurysms may expand in all dimensions [1,4]. Increasing aneurysmal length leads to aortic convolution or unfolding and may confound interpretation of measurements based on simple 2-dimensional images. In addition, there remains the possibility of observer error or bias. Therefore, we assessed departmental inter- and intra-observer variability of measurement of aortic diameter in CT images of thoracic aortic aneurysms (TAAs).

	2. Materials and methods

Top Abstract 1. Introduction 2. Materials and methods 3. Results 4. Discussion References

 
Four experienced observers, three cardiothoracic surgeons and one radiologist, all familiar with the interpretation of thoracic CT scans and involved in the management of patients with thoracic aortic disease measured aortic diameter in 50 pre-selected CT images of patients with TAA or thoracic aortic dissection. The observers (A, B, C, D) were unaware of each others analyses, and made their measurements using a calliper method (A1, B1, C1, D1). All the scans were appropriately enhanced and were available to the observer as a hard-copy film containing multiple 100x100 mm images. Each image represented an approximate 1/5 scaled reduction and contained a reference measurement scale in which approximately 10 mm represented a real dimension of 50 mm. Briefly, observers were instructed to measure the minor (minimum) outer diameter of the aorta in each designated scan directly using a pair of mathematical callipers and to then convert the calliper measurement on the CT reference scale to aortic diameter to the nearest 0.1 mm. Two observers (C, D) also measured the same 50 segments at remote times (2 and 4 weeks following the initial measurement (C2, D2)). In order to clarify inter-observer variations each set of the first measurements (A1, B1, C1, D1) was used and intra-observer variations were examined between the sets of measurement by the same person (C1:C2, D1:D2). Observers were blinded to other observers’ or their own previous measurements. Each combination of measurements were analysed by Wilcoxon's signed ranks test and P<0.05 was interpreted as representing a significant mean difference between two sets of measurement. Repeatability of diameter measurement was assessed using the method of Bland and Altman to determine inter- and intra-observer variations and coefficients of repeatability (CoR) [5,6].

	3. Results

Top Abstract 1. Introduction 2. Materials and methods 3. Results 4. Discussion References

 
Table 1 shows inter-observer P values in Wilcoxon signed ranks test and the CoR. Fig. 1 shows inter-observer variation of aortic diameter using a Bland Altman plot. There were no significant differences between measurements between three different observers (A, B, D) but measurement of one observer (C) was significantly different from that of A or D. The intra-observer CoRs (Table 2) demonstrated improved repeatability than that seen between different observers. The mean and maximum difference in measurements between observers were ±0.88 mm and ±8.0 mm, respectively. Variability of measurement increased with increasing aortic diameter in both inter- and intra-observer measurements as assessed by Bland Altman plot. The best intra-observer coefficient of repeatability was 2.25 while the worst inter-observer repeatability was 4.37.

View larger version (13K): [in this window] [in a new window]   Fig. 1. Bland-Altman's plot of inter-observer differences. Variability of measurements increased with mean aortic diameter. There was a weak positive correlation between absolute difference and mean aortic diameter (AD) in each group. Comparisons B-D and C-D were significantly different (P<0.05, Wilcoxon signed rank test).

	4. Discussion

Top Abstract 1. Introduction 2. Materials and methods 3. Results 4. Discussion References

Compared to ultrasound assessment or echocardiographic examination of the ascending aorta, CT scanning would appear to provide a more objective assessment of aneurysm diameter. However, there remain several factors that may impact upon measurement consistency. These factors predominantly relate to the problems of assessment of a three-dimensional structure using a two-dimensional image. As imaging protocols may vary in the distance between each level of tomography, it is possible that different cross-sections of aneurysms are measured at different times. Even if measurements are performed by alignment with pre-determined anatomical landmarks, as aneurysms expand in three dimensions, the anatomical relationship of an aortic segment to the control landmark may change with time. Error produced by scaled reduction of CT images could be reduced by reconstruction of images targeted to the aorta with minimal scaling. Aneurysmal aortas are commonly elongated and demonstrate folding or convolution effect. Tomographic scans in situations where the aorta does not lie perpendicular to the plane of the scan produces an elliptical image with a major (maximum) and minor (minimum) diameter. In most natural history studies of aneurysm expansion, the minimum diameter has been reported to avoid the effect of convolution [1,2,4]. This approach appears prudent but may lead to under-estimation of effective diameter if the aneurysm is truly elliptical or saccular in cross-section along its long axis at the level of measurement. It is not known if the risk of rupture is more closely related to the major or minor diameter of truly elliptical or non-circular, e.g. saccular aneurysms, but wall tension is directly related to diameter. This suggests that a method to image the aorta, perpendicular to its long axis, throughout its length would be an important advance to monitor the natural history of thoracic aortic aneurysms, providing a more consistent dimension that would accommodate tortuosity and convolution. A further source of potential error relates to the identification of the outer margin of the aortic wall especially as adjacent sub-pleural pulmonary atelectasis may be mistakenly included in measurement. Computed tomography has inadequate resolution to accurately identify aortic wall thickness. This drawback of current imaging can compound accurate measurement and frustrates our ability to describe aneurysm growth in terms of LaPlace's law, in which wall thickness is an integral part of the wall stress equation. One of the limitations of this study is that a calliper technique was used to assess diameter in hard copy films. We do not have data comparing this technique with computerised measurement in which two sets of cross-hairs are targeted by the investigator and the dimension calculated by integrated software. The accuracy of this technique would be dependent upon the calibration of the scan measurement reference and the precise siting of the cross-hairs and would therefore be susceptible to similar errors as the method studied. This study, identifies the potential for measurement error even when pre-selected scan images are used and clear instructions are given to each observer. This suggests that the potential for error in the clinical environment, when different scans, at different levels are assessed by different observers with different measurement criteria could be substantial. Importantly, both inter- and intra-observer variability increase with aortic diameter and any inaccuracy in calculating diameter in larger aneurysms will exponentially increase error in expansion rate calculations. Thus, if at time zero an aneurysm of 60 mm is erroneously measured as 57 mm and three months later, the same unchanged aneurysm is reported as measuring 63 mm, the calculated annual linear expansion rate would be 24 mm per annum. Such a figure might well precipitate a recommendation for surgery. It is therefore recommended that aortic surgical units clearly define the measurement criteria to be adopted and set up internal quality assurance checks of diameter reporting. Decisions to operate based upon presumed accelerated expansion should be carefully validated.

	Footnotes

 
Presented at the Aortic Surgery Symposium VI, New York, NY, USA, April 30 – May 1, 1998.

	References

Top Abstract 1. Introduction 2. Materials and methods 3. Results 4. Discussion References

 

1. Dapunt O.E., Galla J.D., Sadeghi A.M., Lansman S.L., Mezrow C.K., de Alsa R.A., Quintana C., Wallenstein S., Ergin A.M., Griepp R.B. The natural history of thoracic aortic aneurysms. J Thorac Cardiovasc Surg 1994;107:1323-1333.[Abstract/Free Full Text]
2. Coady M.A., Rizzo J.A., Hammond G.L., Mndapati D., Darr U., Kopf G.S., Elefteriades J.A. What is the appropriate size criterion for resection of thoracic aortic aneurysms?. J Thorac Cardiovasc Surg 1997;113:476-491.[Abstract/Free Full Text]
3. Ellis M., Powell J.T., Greenhalgh R.M. Limitation of ultrasonography in surveillance of small abdominal aortic aneurysms. Br J Surg 1991;78:614-616.[Medline]
4. Galla J.D., Ergin M.A., Lansman S.L., DeAsla R.A., Nguyen K.H., McCullough J.N., Griepp R.B. Identification of risk factors in patients undergoing thoracoabdominal aneurysm repair. J Card Surg 1997;12:292-299.[Medline]
5. Bland J.M., Altman D.G. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;i:307-310.
6. British Standards Institution. Precision of Test Methods. I. Guide for the Determination and Reproducibility for a Standard Test Method (BS 5497, part I). London, BSI, 1979.
7. Pitt M.P.I., Bonser R.S. The natural history of thoracic aortic aneurysms: an overview. J Card Surg 1997;12:270-278.[Medline]
8. Juvonen T., Ergin M.A., Galla J.D., Lansman S.L., Nguyen K.H., McCullough J.N., Levy D., de Alsa R.A., Bodian C.A., Griepp R.B. Prospective study of the natural history of thoracic aortic aneurysms. Ann Thorac Surg 1997;63:1533-1545.[Abstract/Free Full Text]
9. Cambria R.A., Gloviczki P., Stanson A.W., Cherry J.K., Jr, Bower T.C., Hallett W.J., Jr, Pairolero R.C. Outcome and expansion rate of 57 thoracoabdominal aortic aneurysms managed nonoperatively. Am J Surg 1995;170:213-217.[Medline]
10. Hirose Y., Hamada S., Takamiya M., Imakita S., Naito H., Nishimura T. Aortic aneurysms: growth rates measured with CT. Radiology 1992;185:249-252.[Abstract/Free Full Text]
11. Hirose Y., Hamada S., Takamiya M., Imakita S., Naito H. Growth rates of aortic aneurysms as a risk factor in rupture: an evaluation with CT. Nippon Acta Radiol 1993;53:635-640.
12. Hirose Y., Hamada S., Takamiya M. Predicting the growth of aortic aneurysms: a comparison of linear vs exponential models. Angiology 1995;46:413-419.
13. Masuda Y., Takanashi K., Takasu J., Morooka N., Inagaki Y. Expansion rate of thoracic aortic aneurysms and influencing factors. Chest 1992;102:461-466.[Abstract/Free Full Text]

This article has been cited by other articles:

				A. Zierer, R. K. Voeller, K. E. Hill, N. T. Kouchoukos, R. J. Damiano Jr, and M. R. Moon Aortic Enlargement and Late Reoperation After Repair of Acute Type A Aortic Dissection Ann. Thorac. Surg., August 1, 2007; 84(2): 479 - 487. [Abstract] [Full Text] [PDF]

				M. S. Kalkat, I. Rahman, K. Kotidis, B. Davies, and R. S. Bonser Presentation and outcome of Marfan's syndrome patients with dissection and thoraco-abdominal aortic aneurysm Eur. J. Cardiothorac. Surg., August 1, 2007; 32(2): 250 - 254. [Abstract] [Full Text] [PDF]

				R S Bonser, D Pagano, M E Lewis, S J Rooney, P Guest, P Davies, and I Shimada Clinical and patho-anatomical factors affecting expansion of thoracic aortic aneurysms Heart, September 1, 2000; 84(3): 277 - 283. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): Ichiro Shimada Stephen J. Rooney Robert S. Bonser Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Shimada, I. Articles by Bonser, R. S. Search for Related Content PubMed PubMed Citation Articles by Shimada, I. Articles by Bonser, R. S.