A New Formula for Rapid Assessmentof Pericardial Effusion Volume

by Computed Tomography

Michael Groth, MD, Marc Regier, MD, Kai Muellerleile, MD,

Peter Bannas, MD, Gerhard Adam, MD, Frank Oliver Henes, MD

Ac

FranCGHNto

ªdo

71

Rationale and Objectives: The aim of this study was to evaluate a new formula for the rapid assessment of pericardial effusion (PE)

volume by computed tomography.

Materials and Methods: Twenty computed tomographic scans positive for PE were reviewed by two observers. Diameters of PE were

measured at four locations. Additionally, PE volume was assessed by volumetry. The correlation between PE diameters and volume

was evaluated, and a linear equation was derived for each diameter location. To test validity and reliability of the measurements, intraclass

correlation and Bland-Altman analysis were performed.

Results: Good validity was expressed by strong correlations between diameter measurements at all four locations and PE volume (all

R values >0.80 and P values <.0001). Intraclass correlation (all coefficients >0.75) and Bland-Altman analysis revealed good interobserver

and intraobserver reliability of diameter measurements. The best values were observed for apical diameter measurements. The followinglinear equation was derived for apical diameter measurements: PE volume = 296 (mL/cm) � apical diameter (cm) � 32 mL.

Conclusions: PE volume can rapidly be assessed by apical PE diameter measurement using the simplified formula PE volume =

0.3 (L/cm) � apical diameter (cm).

Key Words: Pericardial effusion; computed tomography; volume; formula.

ªAUR, 2012

The heart is surrounded by the pericardium, which

consists of two layers. The outer layer is called the

parietal pericardium and the inner layer the visceral

pericardium (1,2). The space between the two layers, the

pericardial cavity, normally contains a small amount of

pericardial fluid, typically between 15 and 50 mL (1,3,4),

which reduces friction within the pericardium by allowing

the pericardial layers to glide over each other with every

heartbeat. Abnormal accumulation of fluid in the pericardial

cavity is called pericardial effusion (PE). The etiology of

PE can be idiopathic, iatrogenic, metabolic, or caused by

neoplasia, infection, or connective tissue disease (5–8). If

a critical rate of pericardial fluid accumulation relative to

pericardial stretch is reached, increasing intrapericardial

pressure can lead to a severe decrease in cardiac function (9).

This life-threatening condition is called cardiac tamponade.

The critical pericardial fluid volume that leads to cardiac

ad Radiol 2012; 19:718–722

om the Center for Radiology and Endoscopy, Department of Diagnosticd Interventional Radiology (M.G., M.R., P.B., G.A., F.O.H.), and theenter for Cardiology and Cardiovascular Surgery, Department ofeneral and Interventional Cardiology (K.M.), University Medical Centeramburg-Eppendorf, Martinistraße 52, 20246 Hamburg, Germany. Receivedovember 24, 2011; accepted January 25, 2012. Address correspondence: M.G. e-mail: groth.michael@googlemail.com

AUR, 2012i:10.1016/j.acra.2012.01.008

8

tamponade depends on whether the volume increases slowly

or rapidly over time. Nevertheless, volumes of nonhemorrha-

gic effusions ranging from 300 to 600 mL have been reported

to cause cardiac tamponade (9).

In certain patient cohorts, PE can be frequently detected

as an incidental finding on computed tomographic (CT)

examinations. For example, the incidence rate of PE in

patients undergoing CT imaging of the pulmonary arteries

for suspected pulmonary embolism ranges between 2% and

5% (10–12).

PE volume can be calculated using CT volumetry (13).

Because of the need to measure PE area on various CT slices,

CT volumetry is very time consuming. Nevertheless, estima-

tion of PE volume is important for clinicians to evaluate

hemodynamic relevance and to consider relief by pericardial

paracentesis.

For this reason, the aim of this study was to introduce a new

formula for the rapid assessment of PE volume by CT

imaging.

MATERIALS AND METHODS

Study Population

With the use of our radiology information system, 20 patients

(eight women, 12 men; mean age, 61.6� 14.0 years) with PE

Academic Radiology, Vol 19, No 6, June 2012 PERICARDIAL EFFUSION VOLUME ASSESSMENT BY CT

were identified among 915 contrast-enhanced thorax-only

CT examinations and included in this study. Institutional

review board approval was obtained for this retrospective

study.

Imaging Technique

All CT studies were performed by one of three multidetector

CT (MDCT) scanners available at our institution, which

include a 16-slice MDCT scanner (MX 8000IDT; Philips

Healthcare, Best, The Netherlands), a 64-slice MDCT

scanner (Brilliance 64; Philips Healthcare), and a 256-slice

MDCT scanner (Brilliance iCT; Philips Healthcare). All

examinations were performed with either the contrast-

enhanced thorax protocol (protocol A) or the pulmonary

angiographic protocol (protocol B).

The scan parameters for protocol A using the 16-slice,

64-slice, and 256-slice MDCT scanners, respectively, were

as follows: rotation time, 0.75, 0.75, and 0.5 seconds; collima-

tion, 16 � 1.5, 64 � 0.625, and 128 � 0.625; and tube

voltage, 120, 120, and 120 kV. Effective tube current was

automated.

The following scan parameters were used for protocol B

using the 16-slice, 64-slice, and 256-slice MDCT scanners,

respectively: rotation time, 0.75, 0.75, and 0.33 seconds;

collimation, 16 � 1.5, 64 � 0.625, and 128 � 0.625; and

tube voltage, 120, 120, and 120 kV. Effective tube current

was automated.

Scan volume included the entire chest from the lung apex

to the lung base. CT images were obtained in a single

breath-hold at end-inspiration in the supine position.

Protocol A consisted of the administration of 70 to 80 mL

of nonionic contrast material with an iodine concentration of

300 mg/mL (Imeron 300; Bracco Altana Pharma, Milan,

Italy) injected at a rate of 2 mL/s through a 20-gauge or

18-gauge peripheral intravenous catheter. Imaging was

performed 40 seconds after injection.

Protocol B included the administration of 80 to 100 mL of

nonionic contrast material with an iodine concentration of

400 mg/mL (Imeron 400; Bracco Altana Pharma) injected

at a rate of 4 mL/s through an 18-gauge peripheral intrave-

nous catheter placed in the antecubital vein. The threshold

level for triggering the scan was achieved when the attenua-

tion of the main pulmonary artery measured $130 Houns-

field units (HU).

Image Analysis

All studies were evaluated using standard window and level

settings (protocol A: width, 360 HU; level, 60 HU; protocol

B: width, 560 HU; level, 60 HU). The reviewers were

allowed to change these settings and to zoom into areas of

interest. The CT data set was reconstructed at 5.0-mm and

3.0-mm slice thicknesses for protocols A and B, respectively.

Additional coronal reformations with section thicknesses of

5.0 and 3.0 mm, respectively, were reconstructed as well.

CT scans were reviewed independently by two observers

(M.G. and F.O.H., each with 4 years of experience in CT

imaging). Measurements of pericardial fluid volumes were

measured once by each of the two observers. Diameter

measurements were measured twice by the first (M.G.) and

once by the second (F.O.H.) investigator. There was a 4-

week interval between the readings of the first investigator.

For the calculation of PE volume, a region of interest was

drawn around the PE area on every CT slice (13). The volume

was calculated by multiplying the sum of areas from each slice

by the reconstruction interval.

PE diameters were measured at four predetermined

locations: superior aortic recess (SAR), anterior, posterior,

and apical (Fig 1).

The SAR is the superior extent of the pericardial transverse

sinus. It is located anterior to the aorta and pulmonary artery,

forming a cleft as it indents between the great vessels (1). The

transverse extension of the SAR was measured from the

indention between the great vessels at the level of right pulmo-

nary artery and the most ventral portion of the SAR (Fig 1a).

Anterior PE was measured as the largest diameter in front of

the right ventricle and posterior pericardial fluid as the largest

diameter behind the mid posterior free wall of the left

ventricle (Fig 1b). Both measurements were done in the

transverse plane.

The diameter of apical PE was assessed as the widest

distance between the left ventricular apex and the most apical

part of the parietal pericardium in the coronal plane (Fig 1c).

Statistical Analysis

Statistical analysis was performed using Microsoft Excel 2003

(Microsoft Corporation, Redmond, WA) and MedCalc for

Windows (MedCalc Software, Mariakerke, Belgium).

Reliability is the capability of a method to give the same

results in the same observer on different occasions (intraob-

server reliability) or with different observers on the same occa-

sion (interobserver reliability). It expresses the reproducibility

of a method (14). Validity assessesses whether a test is actually

measuring what it is indicating to measure and therefore

expresses the measurement accuracy (14).

Reliability was examined using the intraclass correlation

coefficient (ICC) to assess the agreement of measurements

and Bland-Altman analysis to present the bias and limits of

agreement (15–17). For interobserver reliability of diameter

measurements, the first or second measurement from the

first observer was compared to the measurement from the

second observer. For even consecutive inclusion numbers,

the first measurement of the first observer was used, and for

uneven numbers, the second was used. An ICC >0.75 was

considered to signify good agreement.

Validity was assessed by comparing the volumetric and diam-

eter measurements of both observers using Pearson’s correlation

coefficient. Moreover, a linear equation was derived for each

diameter location. A correlation coefficient >0.80 has been

considered strong, between 0.80 and 0.50 moderate, between

719

Figure 1. Diameter measurements ofpericardial effusion (PE) at four predefined

landmarks. The largest diameter of fluid

within the superior pericardial recess was

measured in transverse orientation betweenthe ascending aorta (AA) and the main

pulmonary artery (MPA) at the level of right

pulmonary artery (RPA) (white double arrow

in a). Anterior PE was measured in front ofthe right ventricle (RV) (white double arrow

in b) and posterior pericardial fluid behind

themid posterior freewall of the left ventricle(LV) (black double arrow in b). The diameter

of apical PE was assessed as the largest

PE extension apical to the left ventricle in

coronal orientation (white double arrow inc). LA, left atrium; RA, right atrium.

TABLE 1. Reliability of Pericardial Diameter Measurements at Distinct Locations

Anterior Posterior Apical Superior Aortic Recess

Intraobserver reliability

Intraclass correlation coefficient 0.80 0.87 0.97 0.92

Bland-Altman (cm) 0.04 � 0.82 �0.08 � 0.79 0.01 � 0.42 0.06 � 0.72

Interobserver reliability

Intraclass correlation coefficient 0.81 0.82 0.95 0.89

Bland-Altman (cm) 0.03 � 0.88 0.12 � 0.99 0.07 � 0.55 0.06 � 0.85

GROTH ET AL Academic Radiology, Vol 19, No 6, June 2012

0.50 and 0.30 weak, and between 0.30 and 0 almost nonexis-

tent. P values <.05 indicated statistical significance.

Diameter and volume measurements were tested for

normality with using the Kolmogorov-Smirnov test for

normal distribution.

For validity testing, a sample size $16 provided >80%

power to detect a correlation >0.60 with an a value of 0.05.

RESULTS

Data Distributions

The Kolmogorov-Smirnov test for normal distribution

accepted normality for all measurements: anterior (P = .21),

posterior (P = .40), apical (P = .11), SAR (P = .94), and

volume measurements (P = .14).

Reliability of Volume Measurements

PE volume measurements showed excellent agreement

between the two observers, with an ICC of 0.99. Bland-

Altman analysis displayed a mean bias of �2.5 mL (95% limits

of agreement, �67.2 mL).

Reliability of Diameter Measurements

ICCs revealed good interobserver agreement for all measure-

ment locations, with the strongest ICC for apical measure-

ments (Table 1). Moreover, Bland-Altman analysis displayed

the smallest limits of interobserver agreement for apical

measurements (Table 1, Fig 2a).

720

Good intraobserver agreement was found for all measure-

ment locations as well (Table 1). However, apical measure-

ments showed the strongest ICC and the smallest limits of

agreement (Table 1, Fig 2b).

Validity of Diameter Measurements

Significant and strong Pearson’s correlation coefficients

between diameter measurements and PE volume were

demonstrated for all measurement locations. Pearson’s corre-

lation coefficients, P values,R2 values, and equations are listed

in Table 2. The strongest Pearson’s correlation coefficient was

found for apical measurements (Fig 3).

Derivation of Equation for PE Volume Estimation

Of the four measured PE diameters, the one with best validity

and reliability was chosen to derive a simplified formula for PE

volume estimation. For this purpose, a linear regression equa-

tion for apical diameter measurements against PE volume

(Table 2) was simplified into the following formula: PE

volume = 0.3 (L/cm) � apical diameter (cm).

DISCUSSION

PE volume can be estimated easily and rapid with the intro-

duced formula. For clinical applications, the estimation of

PE volume can be useful for planning surgical drainage and

pericardiocentesis (18).

In this study, PE diameters were measured at four distinct

locations: SAR, anterior, posterior, and apical.

Figure 2. Bland-Altman plots showing inter-

observer (a) and intraobserver (b) reliability ofapical diameter measurements.

TABLE 2. Correlation (Validity) between Diameter Measurements and Pericardial Effusion Volume Assessed by Volumetry andDerived Linear Equation

Anterior Posterior Apical Superior Aortic Recess

Correlation coefficient 0.83 0.88 0.89 0.83

P <.0001 <.0001 <.0001 <.0001

Coefficient of determination 0.69 0.78 0.80 0.70

Linear equation y = 332x � 132 y = 308x � 35 y = 296x � 32 y = 263x � 287

Figure 3. Plot of pericardial effusion volume against apical diameter

measurements (n = 40 measurements performed by the two

observers). Ninety-five percent confidence intervals (dotted lines)

were fitted to the regression line (solid line).

Academic Radiology, Vol 19, No 6, June 2012 PERICARDIAL EFFUSION VOLUME ASSESSMENT BY CT

Apical PE measurements for the estimation of PE volume

revealed the best validity and reliability. This fact might be

especially useful in daily clinical routine, because the heart

apex is normally displayed in thorax-only as well as

abdomen-only CT examinations. Therefore, the evaluated

formula allows the assessment of PE volume not only if the

whole heart is imaged but even on abdomen-only CTexam-

inations without coverage of the whole heart.

Nevertheless, our study showed that anterior and posterior

as well as diameter measurements obtained at the level of the

SAR showed strong correlation with PE volume as well.

In 1991, D’Cruz and Hoffman (19) evaluated a formula for

estimating the volume of PE by echocardiography. Leibowitz

et al (20) investigated the accuracy of this formula to estimate

PE volume by echocardiography and CT imaging. Although

their study showed a significant correlation between PE

quantification by CT imaging and actual PE volume

(r = 0.40, P = .004), echocardiographic quantification was

more accurate. Because the formula of D’Cruz and Hoffmann

was validated only by echocardiography, Leibowitz et al

considered a CT imaging–derived anatomic formula to be

more accurate. However, the CT imaging–based formula

evaluated in our study shows a stronger correlation with actual

PE volume than the formula of D’Cruz and Hoffmann.

Therefore, our formula seems to be more appropriate for

the estimation of PE volume by CT imaging.

Some limitations of our study must be addressed. First,

PE volume was calculated only by tomographic volumetry,

and no comparison with surgically drained volume was per-

formed, because none of the investigated subjects underwent

surgical PE drainage. Nevertheless, a previous study demon-

strated PE volume calculated by tomographic volumetry to

be almost consistent with volume drained surgically (10).

Moreover, this study showed that aspiration of PE by inserting

a small-gauge tube showed smaller PE volumes, most probably

because complete aspiration of the fluid was not possible (13).

Second, because of the retrospective character of our study,

CTexaminations positive for PE were selected from the daily

clinical program at our institution. Hence, in most cases, PE

was an incidental finding in patients without clinical signs or

symptoms of PE. This might have led to the inclusion of

a larger number of small PEs.

Third, the need for radiation exposure limits the use of CT

imaging for the evaluation of PE. Especially because of the

wide availability of other diagnostic modalities, such as echo-

cardiography, CT imaging should not be used for the first-line

evaluation of PE. Therefore, at our institution, most patients

with symptomatic PE undergo echocardiography rather than

721

GROTH ET AL Academic Radiology, Vol 19, No 6, June 2012

CT imaging for diagnostic evaluation, which might have

influenced the study population as well.

Fourth, the study collective was small and contained just 20

patients with PE. These patients were identified among 915

CT examinations of the thorax, which were retrospectively

reviewed for the presence or absence of PE. To include larger

number of patients, a study carried out prospectively would be

preferable. However, because of the need for radiation expo-

sure, conducting such a prospective study would be

questionable.

However, although echocardiography enables the differen-

tiation of large PE from moderate or small PE, it lacks accu-

racy in quantitating PE volume (21). Another limitation of

echocardiography may be hampered performance due to

the lack of an optimal ultrasound window after chest surgery

(18). Thus, CT imaging for the evaluation of PE might be

indicated in special cases. However, because PE is a common

additional finding in patients undergoing CT imaging

(10,11), a formula for easy and fast assessment of PE is very

helpful in daily clinical routine.

CONCLUSIONS

We recommend a simplified equation for the rapid assessment

of PE volume detected by CT imaging:

PE volume ¼ 0:3ðL=cmÞ � apical diameterðcmÞ:

This formula might be very helpful for clinicians in the

evaluation of hemodynamic relevance and in considering

relief by pericardial paracentesis.

REFERENCES

1. Truong MT, Erasmus JJ, Gladish GW, et al. Anatomy of pericardial

recesses on multidetector CT: implications for oncologic imaging. AJR

Am J Roentgenol 2003; 181:1109–1113.

2. Scherer A, Choy G, Kropil P, et al. Cardiac pathologies incidentally

detected with non-gated chest CT [article in German]. Rofo 2009; 181:

1127–1134.

3. Restrepo CS, Lemos DF, Lemos JA, et al. Imaging findings in cardiac

tamponade with emphasis on CT. Radiographics 2007; 27:1595–1610.

722

4. YaredK, Baggish AL, PicardMH, et al. Multimodality imaging of pericardial

diseases. JACC Cardiovasc Imaging 2010; 3:650–660.

5. Tian L, Liu LZ, Cui CY, et al. CT findings of primary non-teratomatous

germ cell tumors of the mediastinum—a report of 15 cases. Eur J Radiol.

In press.

6. Imazio M, Spodick DH, Brucato A, et al. Controversial issues in the

management of pericardial diseases. Circulation 2010; 121:916–928.

7. Eichler K, Zangos S, Thalhammer A, et al. CT-guided pericardiocenteses:

clinical profile, practice patterns and clinical outcome. Eur J Radiol 2010;

75:28–31.

8. Da Ines D, Chabrot P, Motreff P, et al. Cardiac tamponade after malignant

superior vena cava stenting: two case reports and brief review of the liter-

ature. Acta Radiol 2010; 51:256–259.

9. Spodick DH. Acute cardiac tamponade. N Engl J Med 2003; 349:

684–690.

10. Hall WB, Truitt SG, Scheunemann LP, et al. The prevalence of clinically

relevant incidental findings on chest computed tomographic angiograms

ordered to diagnose pulmonary embolism. Arch Intern Med 2009; 169:

1961–1965.

11. Lee EY, Kritsaneepaiboon S, Zurakowski D, et al. Beyond the pulmonary

arteries: alternative diagnoses in children with MDCT pulmonary

angiography negative for pulmonary embolism. AJR Am J Roentgenol

2009; 193:888–894.

12. Groth M, Henes FO, Mayer U, et al. Age-related incidence of pulmonary

embolism and additional pathologic findings detected by computed

tomography pulmonary angiography. Eur J Radiol. In press.

13. Tomoda H, Hoshiai M, Furuya H, et al. Evaluation of pericardial effusion

with computed tomography. Am Heart J 1980; 99:701–706.

14. Busing KA, Kilian AK, Schaible T, et al. Reliability and validity of MR image

lung volumemeasurement in fetuses with congenital diaphragmatic hernia

and in vitro lung models. Radiology 2008; 246:553–561.

15. Bland JM, Altman DG. Statistical methods for assessing agreement

between two methods of clinical measurement. Lancet 1986; 1:307–310.

16. Costa-Santos C, Bernardes J, Ayres-de-Campos D, et al. The limits

of agreement and the intraclass correlation coefficient may be

inconsistent in the interpretation of agreement. J Clin Epidemiol 2011;

64:264–269.

17. Groth M, Henes FO, Bannas P, et al. Intraindividual comparison

of contrast-enhanced MRI and unenhanced SSFP sequences of

stenotic and non-stenotic pulmonary artery diameters. Rofo 2011; 183:

47–53.

18. Tsang TS, Barnes ME, Hayes SN, et al. Clinical and echocardiographic

characteristics of significant pericardial effusions following cardiothoracic

surgery and outcomes of echo-guided pericardiocentesis for manage-

ment: Mayo Clinic experience 1979-1998. Chest 1999; 116:322–331.

19. D’Cruz IA, Hoffman PK. A new cross sectional echocardiographic method

for estimating the volume of large pericardial effusions. Br Heart J 1991;

66:448–451.

20. Leibowitz D, Perlman G, Planer D, et al. Quantification of pericardial

effusions by echocardiography and computed tomography. Am J Cardiol

2011; 107:331–335.

21. Prakash R, Moorthy K, Del Vicario M, et al. Reliability of echocardiography

in quantitating pericardial effusion: a prospective study. J Clin Ultrasound

1977; 5:398–402.