Cardiopulmonar y Imaging • Original Research

Kitazono et al.
Chest Radiography of Pleural Effusions

Cardiopulmonary Imaging
Original Research

Differentiation of Pleural Effusions

Downloaded from www.ajronline.org by 112.215.238.249 on 12/19/17 from IP address 112.215.238.249. Copyright ARRS. For personal use only; all rights reserved

From Parenchymal Opacities:

Accuracy of Bedside Chest
Radiography
Mary T. Kitazono1 OBJECTIVE. The purpose of this study was to determine, with CT as the reference stan-
Charles T. Lau dard, the ability of radiologists to detect pleural effusions on bedside chest radiographs.
Andrea N. Parada MATERIALS AND METHODS. Images of 200 hemithoraces in 100 ICU patients un-
Pooja Renjen dergoing chest radiography and CT within 24 hours were reviewed. Four readers with varying
Wallace T. Miller, Jr. levels of experience reviewed the chest radiographs and predicted the likelihood of the pres-
ence of an effusion or parenchymal opacity on independent 5-point scales. The results were
Kitazono MT, Lau CT, Parada AN, Renjen P, compared with the CT findings.
Miller WT Jr RESULTS. All readers regardless of experience had similar accuracy in detecting pleural
effusions. Among 117 pleural effusions, 66% were detected on chest radiographs (53%, 71%,
and 92% of small, moderate, and large effusions) with 89% specificity. Similarly, 65% of all
parenchymal opacities were detected on chest radiographs, also with 89% specificity. Most
(93%) of the misdiagnosed pulmonary opacities were simply not seen. Meniscus, apical cap,
lateral band, and subpulmonic opacity were highly specific findings but had low individual
sensitivity for effusions. The finding of homogeneous opacity, including both layering and
gradient opacities, was the most sensitive sign of effusion. Atelectasis can occasionally mimic
the pleural veil sign of effusion, accounting for most false-positive findings.
CONCLUSION. Radiologists interpreting bedside chest radiographs of ICU patients
detect large pleural effusions 92% of the time and can exclude large effusions with high con-
fidence. However, small and medium effusions often are misdiagnosed as parenchymal opac-
ities (45%) or are not seen (55%). Pulmonary opacities often are missed (34%) but are rarely
misdiagnosed as pleural effusions (7%).

Keywords: bedside, pleural effusion, semierect, supine
DOI:10.2214/AJR.09.2950
Received April 22, 2009; accepted after revision
August 8, 2009.
1
All authors: Department of Radiology, Hospital of the
University of Pennsylvania, 3400 Spruce St., Philadelphia,
PA 19104. Address correspondence to
M. T. Kitazono (Mary.Kitazono@uphs.upenn.edu).
AJR 2010; 194:407–412
0361–803X/10/1942–407
© American Roentgen Ray Society
P
leural effusions are exceedingly
common, occurring in an esti-
mated 1 million patients each
year in the United States, many
of whom are critically ill in an ICU [1]. Bed-
side supine and semierect chest radiography
is logistically easier to perform in the care of
critically ill patients and often is the primary
imaging technique used in the ICU. Howev-
er, bedside imaging often is suboptimal, and
numerous pitfalls of supine imaging of the
chest have been identified [2–4].
Basilar opacities are among the most com-
mon imaging findings in ICU patients and are
often a result of pleural effusion, parenchy-
mal lung disease, or a combination of these
conditions. Despite the frequent detection of
basilar lung opacities in our everyday prac-
tice, accuracy in detecting and differentiat-
ing pleural fluid from parenchymal opacities
on bedside supine and semierect chest ra-
diographs is not known. Previous studies in
the English language literature on the accu-
racy of supine radiography in the diagnosis
of pleural effusion have had highly variable
results, likely because of small sample siz-
es, the use of only known positive [5] or sus-
pected [6] findings of effusion, and the meth-
ods used to verify pleural fluid, which ranged
from posteroanterior and lateral upright radi-
ography, lateral decubitus chest radiography,
and ultrasound imaging [5–7]. Furthermore,
without the use of cross-sectional imaging as
the reference standard, the nature of false-
positive findings was not investigated.
We undertook this study to determine,
with CT as the reference standard, the accu-
racy of routine bedside radiography in the di-
agnosis of pleural effusions in critically ill
patients. We hypothesized that bedside chest
radiography in the ICU is moderately accu-
rate in the detection of pleural effusions and
in differentiating them from other causes of
basilar pulmonary opacity.

AJR:194, February 2010 407


Materials and Methods
After institutional review board approval and
in accordance with HIPAA regulations, the cas-
es of 100 patients in the surgical ICU who con-
secutively underwent CT of the chest less than 24
hours after undergoing chest radiography were
Downloaded from www.ajronline.org by 112.215.238.249 on 12/19/17 from IP address 112.215.238.249. Copyright ARRS. For personal use only; all rights reserved
retrospectively selected for review. Informed con-
sent was not required. The exclusion criteria in-
cluded complete omission of the lung bases, in-
terval thoracentesis, and chest tube placement or
a related procedure.
Chest Radiography
The chest radiographs were reviewed by four
readers: a thoracic imaging attending radiologist
with more than 18 years of experience, a thoracic
imaging fellow, a fourth-year radiology resident,
and a second-year radiology resident. No special
training was given before interpretation.
Each hemithorax was separately evaluated for
the presence of basilar opacities. If an opacity was
present, the characteristics were recorded (Table
1, Figs. 1–3). The imaging features of pleural ef-
fusions on supine and semierect radiographs can
be divided into two main categories. The first is
extrapulmonary signs, which identify an opacity
as being between the lung and chest wall or dia-
phragm; these signs include meniscus, apical cap,
lateral band, and subpulmonic opacity. The sec-
ond category is homogeneous densities overlying
the hemithorax; these signs include layering and
gradient opacities.
The readers independently determined the
likelihood of the presence of pleural effusion and
parenchymal opacity on a 5-point scale (5, defi-
nitely present; 4, probably present; 3, possibly
present; 2, probably not present; 1, definitely not
present). The finding of a meniscus, apical cap, or
lateral band was interpreted as certain evidence
of the presence of an effusion (typical score, 5).
The finding of a gradient opacity, layering opaci-
ty, or subpulmonic opacity was interpreted as high
likelihood of the presence of an effusion (typical
score, 4).
Air bronchograms were interpreted as certain
evidence of parenchymal opacities (typical score,
5), and heterogeneity was interpreted as high likeli-
hood of a parenchymal opacity (typical score, 4 or
5). If imaging findings were indicative of moderate
or large pleural effusion, it was assumed that as-
sociated passive atelectasis was present. Therefore,
the presence of a parenchymal opacity also was
predicted (typical score, 4 or 5) even in the absence
of any other signs of parenchymal opacification.
When an opaque hemithorax was encountered,
the location of the mediastinum was the principle
differentiating criterion: an ipsilateral mediastinal
shift indicated a parenchymal opacity, a contralat-
408 AJR:194, February 2010
Kitazono et al.
TABLE 1:  Definition of Terms Used to Characterize Opacities
Opacity Characteristic Definition
Apical capping Sharply marginated curvilinear opacity between aerated apical lung and the
apical ribs (Fig. 1)
Lateral band Sharply marginated linear opacity between aerated lateral border of lung and
medial border of the lateral aspects of the ribs (Fig. 1)
Meniscus Sharply marginated curvilinear opacity that obscures the lateral costophrenic
angle
Subpulmonic opacity Apparent elevation of the hemidiaphragm
Layering opacity Homogeneous opacity overlying the hemithorax with vascular markings
apparent through the opacity (Fig. 2); expected appearance of effusion on
supine image
Gradient opacity Homogeneous opacity overlying hemithorax that becomes progressively more
radiodense from cephalic to caudal aspect with vascular markings apparent
through the opacity (Fig. 3); expected appearance of effusion on a semierect
image
Air bronchograms Air-filled tubular or branching structures visible within an opacity
Uniformity of opacity Homogeneous, heterogeneous, or both homogeneous and heterogeneous
components
A B
Fig. 1—55-year-old woman with pleural effusions appearing as gradient opacities with apical caps and right-
sided lateral band.
A, Bedside anteroposterior semierect radiograph shows bilateral gradient opacities that become progressively
more radiopaque from lung apices to lung bases. Vascular markings are apparent through opacity. Prominent
right-sided lateral band and bilateral apical caps are evident. Bilateral pleural effusions with associated
passive atelectasis were predicted.
B, Unenhanced axial CT scan of chest at level of main pulmonary artery confirms diagnosis of bilateral pleural
effusions with effusion-related atelectasis.
eral mediastinal shift indicated probable pleural ed. Pleural effusion was diagnosed if low-attenu-
effusion, and an unshifted mediastinum indicat- ation (0–20 HU) material was visualized between
ed both possible effusion and possible parenchy- the lung parenchyma and the chest wall. The effu-
mal opacity. sion was considered loculated if it was clearly di-
All chest radiographs were obtained in the an- vided into more than one compartment. There were
teroposterior projection with a mobile radiogra- no cases of high-attenuation pleural effusion or he-
phy unit (AMX-4+, GE Healthcare). According mothorax in our patient sample. The quantity of
to the position indicators, no radiograph was ob- pleural fluid was semiquantitatively scored on CT
tained in the erect position, 28 were semierect, scans as small, moderate, or large. The quantity
and eight were supine. The position of 64 radio- depended on the maximum anteroposterior depth
graphs was not indicated. of pleural fluid (measured at the maximum effusion
depth, which varied by patient): less than 15% (≈ 75
CT mL), 15–30%, or greater than 30% (≈ 350 mL) of
All CT scans were reviewed by the same thorac- the diameter of the hemithorax [8].
ic imaging attending radiologist during a separate Although many patients had diffuse airspace
session after all radiographs had been interpret- disease, such as pulmonary edema, an opacity was
 Chest Radiography of Pleural Effusions
Downloaded from www.ajronline.org by 112.215.238.249 on 12/19/17 from IP address 112.215.238.249. Copyright ARRS. For personal use only; all rights reserved

A B
Fig. 2—57-year-old man with pleural effusions appearing as layering opacities. Fig. 3—52-year-old man with pleural effusions
A, Supine chest radiograph shows homogeneously increased opacification of right hemithorax typical of appearing as gradient opacities. Semierect chest
layering opacity, typically scored 4 for probable pleural effusion. Right hemithorax also was typically scored 4 radiograph shows bilateral homogeneous opacities
for pulmonary opacity because readers were predicting presence of atelectasis independent of effusion. Most at lung bases interpreted as gradient opacities and
readers scored left lung base 4 for opacity and 2 for effusion. scored 5 for pleural effusions by all readers. Probable
B, CT scan confirms presence of moderate right pleural effusion and associated passive atelectasis in adjacent associated bibasilar passive atelectasis was inferred
right lower lobe. from moderate size of pleural effusions. Pleural
effusions and passive atelectasis were confirmed at
CT (not shown).

diagnosed only when a discrete region of alveolar
opacification was found, either alone or superim-
posed on a diffuse alveolar process. For the pur-
poses of this study, we made no distinction between
opacification due to volume loss (atelectasis) and
opacification due to alveolar filling (e.g., pneumo-
nia, contusion, infarction).
Various CT scanners were used, including a
single-detector scanner (CTi, GE Healthcare),
4-MDCT scanners (LightSpeed QXI, GE Health-
care; Volume Zoom, Siemens Healthcare), a 16-
MDCT scanner (Sensation 16, Siemens Health-
care), a 64-MDCT scanner (Sensation 64, Siemens
Healthcare), and a dual-source 64-MDCT scanner
(Somatom Definition, Siemens Healthcare). Sev-
eral imaging techniques were used. Sixty-one ex-
aminations were performed with IV contrast en-
hancement: 47 with 100 mL of iohexol 300 mg I/
mL (Omnipaque 300, GE Healthcare) and 14 with
100 mL of iohexol 350 mg I/mL (Omnipaque 350,
GE Healthcare) at injection rates of 1.5–4.0 mL/s.
Various collimations and pitches were used de-
pending on the type of scanner and imaging proto-
col. Images were reviewed on a PACS workstation
(RadWorks, GE Healthcare) at a slice thickness of
1, 3, or 5 mm.
Statistics
The sensitivity, specificity, and receiver operat-
ing characteristic (ROC) curves in the detection
of pleural effusion on bedside chest radiographs
were calculated with CT as the reference standard.
Area under the ROC curves was calculated with a
Web-based calculator [9]. Separate analyses were
conducted for each reader and for small, moderate,
and large pleural effusions. The sensitivity, speci-
ficity, and areas under the ROC curve in the de-
tection of parenchymal opacities on bedside chest
radiographs also were calculated for each reader.
We considered scores of 3 and higher positive pre-
dictions of effusion or opacity.
Results
CT of the chest was performed on 117
consecutively registered ICU patients with-
in the 2-month period October and Novem-
ber 2007. Thirteen of the patients were ex-
cluded because they had not undergone chest
radiography within 24 hours before CT. Two
patients were excluded because overlying
structures completely obscured the lung bas-
es; one patient, because the lung bases were
completely omitted; and one patient, because
a chest tube was inserted between the imag-
ing studies. The other 100 ICU patients (200
hemithoraces; 60 men and boys, 40 women
and girls; mean age, 54.3 years; range, 14–91
years) formed the final study group. Diffuse
airspace disease was present in 40 patients.
Chest CT depicted 117 cases of pleural ef-
fusion: 16 large, 46 medium, and 55 small.
Eight isolated right pleural effusions, 11 iso-
lated left pleural effusions, and 49 bilateral
pleural effusions were found. Fifteen cases
of effusion were loculated.
Sensitivity and Specificity
All readers had similar accuracy in detect-
ing effusions without a significant difference
according to level of experience. The read-
ers with the highest sensitivity had the lowest
specificity, and thus all readers’ results were
at different positions on similar ROC curves
(Fig. 4). With the results for all readers aver-
aged, the overall sensitivity in the diagnosis of
pleural effusion on bedside chest radiographs
was 66%. The specificity was 89% with a fit-
ted area under the ROC curve of 0.852 (Table
2). Exclusion of the 15 loculated effusions did
not change the diagnostic accuracy.
Because the four readers had similar diag-
nostic accuracy in diagnosing effusion, the
composite results for all readers were used
to determine overall accuracy in the detec-
tion of effusion on the basis of the size of the
effusion and to analyze the accuracy of in-
dividual radiographic characteristics of effu-
sion. Overall, 53% of the small pleural ef-
fusions, 71% of the moderate effusions, and
almost all (92%) of the large pleural effu-
sions were correctly diagnosed. The speci-
ficity was high, approximately 89%, for all
pleural effusions (Table 3).
False-Negative Findings of Effusion
On average, 34% (161/468, the denomina-
tor being the total number of effusions mul-
tiplied by four readers) of the pleural effu-
sions were missed on chest radiographs. Most
(64%) of the missed effusions were small,
33% were moderate, and 3% were large. For-
ty-five percent (73/161) of the missed effu-
sions were interpreted as parenchymal opaci-
ties rather than effusions, and the other 55%
(88/161) were interpreted as not basilar opaci-
ties. In 59% (95/161) of the cases of false-neg-
ative findings of effusions interpreted as not
basilar opacity, evidence of diffuse alveolar

AJR:194, February 2010 409

 Kitazono et al.

Fig. 4—Graph shows hibiting some kind of homogeneous opacity,

1.0 receiver operating
characteristic curves
and was the finding most likely to indicate
0.9 the presence of small and moderate-sized ef-
for detection of pleural
0.8 effusions of all sizes by fusions (Table 4). Atelectasis, however, can
readers with varying occasionally mimic the pleural veil sign on
True-Positive Fraction

0.7 levels of experience.

chest radiographs, accounting for most of the
Downloaded from www.ajronline.org by 112.215.238.249 on 12/19/17 from IP address 112.215.238.249. Copyright ARRS. For personal use only; all rights reserved

0.6
false-positive findings in this study.
0.5

0.4
Parenchymal Opacities
CT showed 146 pulmonary opacities, 65%
0.3 (381/584) of which were predicted by the read-
Attending
0.2 Fellow
ers on chest radiographs with a specificity of
0.1
4th-year Resident 89% (193/216) for all readers combined. Indi-
2nd-year Resident
vidual readers had sensitivities varying from
0.0
56% to 77% and specificities varying from
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
False-Positive Fraction
81% to 96% with no significant trend based on
level of experience. The area under the ROC
TABLE 2:  Reader Accuracy in Diagnosis of Pleural Effusion curve varied from 0.815 (second-year resi-
dent) to 0.872 (fourth-year resident) (Table 5).
Reader Sensitivity (%) Specificity (%) Area Under ROC Curve
Most of the misdiagnosed pulmonary opaci-
Attending 59 (69/117) 95 (79/83) 0.844 ties (93%, 150/161) on chest radiographs were
Fellow 62 (72/117) 90 (75/83) 0.843 erroneously predicted to be normal lung bases
Fourth-year resident 76 (89/117) 86 (71/83) 0.878 with no effusion or discrete opacity.
Second-year resident 66 (77/117) 93 (77/83) 0.855
Discussion
Average 66 (309/468)a 89 (295/332)a 0.852 Bedside supine and semierect chest radiog-
Note—Values in parentheses are raw numbers. ROC = receiver operating characteristic. raphy of ICU patients has low and interme-
aTotal number of effusions multiplied by four readers.
diate sensitivity in the detection of small and
moderate effusions (53% and 71%, respective-
TABLE 3: Average Reader Accuracy for Detecting Pleural Effusions ly) but is highly sensitive and specific in the
of Different Sizes With Bedside Chest Radiography
detection of large effusions (92% and 89%).
Size of Effusion Sensitivity (%) Specificity (%) Accordingly, large effusions usually can be
Small 53 (117/220) 89 (295/332) excluded with high confidence, but small and
Moderate 71 (131/184) 89 (295/332) medium effusions often are not identified be-
cause they are misdiagnosed as parenchymal
Large 92 (59/64) 89 (295/332)
opacities (45%) or are not seen (55%).
Total (n = 117) 66 (309/468)a 89 (295/332)a The 66% overall sensitivity for all pleural
Note—Values in parentheses are raw numbers. effusions in our study is slightly lower than
aTotal number of effusions multiplied by four readers.
previously reported sensitivities ranging from
67% to 82% [6, 7]. However, the previous
lung disease was seen on chest radiographs, Characteristics of Pleural Effusion studies may have shown falsely elevated sen-
but the presence of an independent basilar Meniscus, apical cap, lateral band, sub- sitivities due to the methods of verifying ef-
opacity was not predicted. pulmonic opacity, layering opacity, and gra- fusion, including decubitus radiography and
dient opacity all were highly specific but not ultrasound, which are not as sensitive as CT
False-Positive Findings of Effusion sensitive signs of pleural effusion (Table 4). in the detection of small effusions. Exclud-
Eleven percent (36/327) of the suspected In total, 73% (342/468) of all patients with ing the small effusions in our study increases
effusions were false-positive findings, most pleural effusions had at least one of the high- the sensitivity to 78% (190/244) in the detec-
of which (64%, 209/327) were homogeneous ly specific signs on chest radiographs (apical tion of effusion, and this rate is comparable
opacities on chest radiographs that proved to cap, lateral band, meniscus, layering opac- with previously reported values.
be atelectasis on CT scans. Approximately ity, gradient opacity, or subpulmonic opac- Diagnostic accuracy for effusion did not
one of three (31%, 11/36) false-positive find- ity). Complete opacification of a hemitho- differ significantly between readers, whose
ings of effusion had a false meniscus sign, rax with ipsilateral mediastinal shift was levels of experience varied. Although there
which was usually caused by pleural thick- found in only three patients, and CT showed was only one reader in each category, our re-
ening or a large pericardial fat pad (Fig. 5). that all three of these patients had ipsilater- sults suggest that the ability to diagnose pleu-
The attending radiologist made one false- al effusions and atelectasis. Homogeneous ral effusions on bedside radiographs is a basic
positive finding of meniscus, and each of the opacity was the single most sensitive find- skill that can be learned quickly during resi-
other readers made three or four false-posi- ing of pleural effusion on bedside chest ra- dency. The results also suggest that because
tive findings of meniscus. diographs, 68% (319/468) of effusions ex- of suboptimal visualization on bedside radio-

410 AJR:194, February 2010

 Chest Radiography of Pleural Effusions

causes homogeneously increased attenuation

indistinguishable from that of pleural fluid on
chest radiographs. It accounted for most of the
false-positive findings (73%) in this study.
The readers’ accuracy in detecting paren-
chymal opacities was rather low, only 65%
Downloaded from www.ajronline.org by 112.215.238.249 on 12/19/17 from IP address 112.215.238.249. Copyright ARRS. For personal use only; all rights reserved

of the opacities being correctly predicted on

chest radiographs. Most of these misdiag-
nosed opacities (93%) were simply not seen
and were interpreted as a normal lung base
with no effusion or focal opacity. On CT
scans, however, many of these missed opaci-
ties were visualized as small areas of subseg-
A B
mental atelectasis or focal ground-glass opac-
Fig. 5—75-year-old woman with false-positive finding of pleural effusion. ities, which are findings of dubious clinical
A, Bedside chest radiograph shows apparent meniscus sign at left costophrenic angle, typically scored 4 or 5
for probable or definite pleural effusion. Heterogeneous opacity at left lung base also was predicted. significance. In a small percentage of cases
B, CT scan shows large pericardial fat pad curving up lateral margin of chest wall, producing false appearance (7%), the opacity was mistaken for effusion.
of meniscus. Parenchymal infiltrate and dependent atelectasis are evident at left lung base. Several limitations to our study should
be noted. The retrospective design in which
TABLE 4:  Reader Accuracy in Identifying Features of Pleural Effusion
only patients undergoing chest CT were se-
Frequency (%) lected as participants may have biased the
Small Moderate Large Sensitivity Specificity sample to include patients who are the most
Finding Effusion Effusion Effusion (%) (%) ill and more likely to have pleural effusions.
Apical capping 2 (4/220) 8 (14/184) 9 (6/64) 5 (24/468) 100 (24/24) The high prevalence of both pleural effusion
and pulmonary opacity in this ICU sample
Lateral band 4 (9/220) 12 (22/184) 31 (20/64) 11 (52/468) 100 (52/52)
prevents generalization of these results to
Subpulmonic opacity 2 (4/220) 2 (4/184) 9 (6/64) 3 (14/468) 88 (14/16) populations of patients who are not critically
Meniscus 13 (28/220) 16 (29/184) 41 (26/64) 18 (83/468) 89 (83/94) ill. However, patients who are not critically
Layering opacity 9 (19/220) 22 (41/184) 23 (15/64) 16 (75/468) 95 (75/79) ill can undergo standard upright radiography.
In addition, the 200 hemithoraces were treat-
Gradient opacity 19 (41/220) 32 (58/184) 55 (35/64) 29 (134/468) 91 (134/147)
ed as separate entities, and the radiographic
Any kind of homogeneous opacity 43 (94/319) 67 (123/184) 91 (58/64) 59 (275/468) 86 (275/319) findings of large ipsilateral pleural effusion
might have biased the readers toward pre-
TABLE 5:  Reader Accuracy in Diagnosing Pulmonary Opacities dicting the presence of a small contralateral
Area Under ROC effusion in equivocal cases.
Reader Sensitivity (%) Specificity (%) Curve Bedside supine and semierect chest radiog-
Attending 61 (89/146) 94 (51/54) 0.842 raphy of ICU patients is reliable for differen-
tiating large pleural effusions from pulmonary
Fellow 56 (82/146) 96 (52/54) 0.852
opacities. Small and medium effusions, how-
Fourth-year resident 77 (112/146) 85 (46/54) 0.872 ever, often are misdiagnosed as parenchymal
Second-year resident 67 (98/146) 81 (44/54) 0.815 opacities (45% in this study) or are not seen
Average 65 (381/584) 89 (193/216) 0.863 (55% of cases in this study). Pleural effusion is
present in most ICU patients in whom effusion
is suspected (89% specificity in this study).
graphs, there is a limit to diagnostic accuracy pneumonia, atelectasis, and contusion, which Large effusions can be excluded with high
that is difficult to overcome, even for the most cause parenchymal opacities and often con- confidence on supine and semierect chest ra-
experienced readers. tain a mixture of aerated and unaerated lung diographs. A considerable portion of pulmo-
All six characteristic findings of effusion that causes heterogeneously increased attenu- nary opacities are misdiagnosed (34% in this
were highly specific for the presence of ef- ation. In this study, the presence of a homoge- study), most of which are simply not seen.
fusion. However, the extrapulmonary signs, neous opacity seen in isolation or superim- An opacity is infrequently mistaken for an ef-
including meniscus, apical cap, lateral band, posed on a heterogeneous opacity was a fusion (7% of cases in this study).
and subpulmonic opacity, are uncommon moderately specific indicator of the presence
findings. Only 27% of effusions had at least of pleural fluid, having 86% specificity. This References
one of these signs, and most of these effu- finding was the most sensitive (68%) for 1. Light RW. Pleural diseases. Philadelphia, PA:
sions were moderate or large. pleural effusion on bedside chest radiographs Lea and Febiger, 1983
The homogeneously increased attenuation and was the finding most likely to indicate 2. Ovenfors CO, Hedgecock MW. Intensive care
caused by pleural fluid often differentiates the presence of small and moderate-sized ef- unit radiology: problems of interpretations. Radi-
pleural effusion from conditions such as fusions. However, atelectasis occasionally ol Clin North Am 1978; 16:407–439

AJR:194, February 2010 411

 Kitazono et al.

3. Rose CC, Delbridge TR, Mosesso VN. The por-
table chest film. Emerg Med Clin North Am 1991;
9:767–788
4. Emamian SA, Kaasbol MA, Hegedus V. Supine
bedside chest x-ray as source for iatrogenic com-
plication. Iatrogenics 1991; 1:165–168
Downloaded from www.ajronline.org by 112.215.238.249 on 12/19/17 from IP address 112.215.238.249. Copyright ARRS. For personal use only; all rights reserved
AJR 1984; 142:59–64
6. Ruskin JA, Gurney JW, Thorsen MK, Goodman
LR. Detection of pleural effusions on supine chest
radiographs. AJR 1987; 148:681–683
7. Emamian SA, Kaasbol MA, Olsen JF, Pedersen
JF. Accuracy of the diagnosis of pleural effu-
8. Mergo PJ, Helmberger T, Didovic J, Cernigliaro J,
Ros PR, Staab EV. New formula for quantification
of pleural effusions from computed tomography. J
Thorac Imaging 1999; 14:122–125
9. Eng J. ROC analysis: Web-based calculator for
ROC curves. Baltimore: Johns Hopkins Universi-

5. Woodring JH. Recognition of pleural effusion on sion on supine chest x-ray. Eur Radiol 1997; ty. http://www.jrocfit.org. Updated May 17, 2006.
supine radiographs: how much fluid is required? 7:57–60 Accessed March 20, 2008

F O R YO U R I N F O R M AT I O N
Mark your calendar for the following ARRS annual meetings:
May 2–7, 2010—Manchester Grand Hyatt San Diego, San Diego, CA
May 1–6, 2011—Hyatt Regency Chicago, Chicago, IL
April 29–May 4, 2012—Vancouver Convention Center, Vancouver, BC, Canada
April 14–April 19, 2013—Marriott Wardman Park Hotel, Washington, DC

412 AJR:194, February 2010

 This article has been cited by:

1. Pascal Selot, Daniel Ferreiro Carballal, Maria Bonarek, Stéphane Henriot, Julie Rodriguez, Aymeric Le Neindre. 2017. À propos
d’un cas de pneumopathie, l’échographie pulmonaire comme aide au diagnostic et au contrôle de la réponse à la kinésithérapie.
Kinésithérapie, la Revue . [CrossRef]
2. Bryan K. Moffett, Tanmay S. Panchabhai, Raul Nakamatsu, Forest W. Arnold, Paula Peyrani, Timothy Wiemken, Juan
Downloaded from www.ajronline.org by 112.215.238.249 on 12/19/17 from IP address 112.215.238.249. Copyright ARRS. For personal use only; all rights reserved

Guardiola, Julio A. Ramirez. 2016. Comparing posteroanterior with lateral and anteroposterior chest radiography in the initial
detection of parapneumonic effusions. The American Journal of Emergency Medicine 34:12, 2402-2407. [CrossRef]
3. J. M. Karhu, T. I. Ala-Kokko, L. K. Ahvenjärvi, E. Rauvala, P. Ohtonen, H. P. T. Syrjälä. 2016. Early chest computed
tomography in adult acute severe community-acquired pneumonia patients treated in the intensive care unit. Acta
Anaesthesiologica Scandinavica 60:8, 1102-1110. [CrossRef]
4. Nilam J. Soni, Ricardo Franco, Maria I. Velez, Daniel Schnobrich, Ria Dancel, Marcos I. Restrepo, Paul H. Mayo. 2015.
Ultrasound in the diagnosis and management of pleural effusions. Journal of Hospital Medicine 10:12, 811-816. [CrossRef]
5. David Chiao, Michael Hanley, Juan M. Olazagasti. 2015. CT Volumetric Analysis of Pleural Effusions. Academic Radiology
22:9, 1122-1127. [CrossRef]
6. A. P. Nayate, J. G. Dubroff, J. E. Schmitt, I. Nasrallah, R. Kishore, D. Mankoff, D. A. Pryma. 2015. Use of Standardized
Uptake Value Ratios Decreases Interreader Variability of [18F] Florbetapir PET Brain Scan Interpretation. American Journal
of Neuroradiology 36:7, 1237-1244. [CrossRef]
7. Judith F. Katz, James S. Bezreh, E. Kent Yucel. 2015. Lung ultrasound in the intensive care unit: an idea that may be too good
to be true. Intensive Care Medicine 41:2, 379-380. [CrossRef]
8. G. Thommi, J.C. Shehan, K.L. Robison, M. Christensen, L.A. Backemeyer, M.T. McLeay. 2012. A double blind randomized
cross over trial comparing rate of decortication and efficacy of intrapleural instillation of alteplase vs placebo in patients with
empyemas and complicated parapneumonic effusions. Respiratory Medicine 106:5, 716-723. [CrossRef]
9. C. Schülke, N. Roos, B. Buerke, W. Heindel. 2011. Thoraxradiologie auf der Intensivstation. Medizinische Klinik -
Intensivmedizin und Notfallmedizin 106:2, 96-102. [CrossRef]
10. ANUPAMA G. BRIXEY, YIFENG LUO, VASILEIOS SKOURAS, ANNA AWDANKIEWICZ, RICHARD W. LIGHT.
2011. The efficacy of chest radiographs in detecting parapneumonic effusions. Respirology 16:6, 1000-1004. [CrossRef]
11. T. Dorman. 2011. Differentiation of Pleural Effusions From Parenchymal Opacities: Accuracy of Bedside Chest Radiography.
Yearbook of Critical Care Medicine 2011, 90-92. [CrossRef]
12. M.L. Rosado de Christenson. 2011. Differentiation of Pleural Effusions From Parenchymal Opacities: Accuracy of Bedside
Chest Radiography. Yearbook of Diagnostic Radiology 2011, 18-19. [CrossRef]