Pediatr Radiol (2004) 34: 895900

DOI 10.1007/s00247-004-1251-3

Joaquim Bosch-Marcet
Xavier Serres-Creixams
Amalia Zuasnabar-Cotro
Xavier Codina-Puig
Margarita Catala`-Puigbo
Jose L. Simon-Riazuelo

Received: 21 February 2004

Revised: 1 April 2004
Accepted: 11 May 2004
Published online: 9 September 2004

Springer-Verlag 2004
J. Bosch-Marcet (&)
A. Zuasnabar-Cotro M. Catala`-Puigbo
J. L. Simon-Riazuelo
Department of Paediatrics,
Avda. Francesc Ribas s/n,
Hospital General de Granollers, 08400
Granollers, Barcelona, Spain
E-mail: 8543jbm@telefonica.net
Tel.: +34-93-8425039
Fax: +34-93-8425036
X. Serres-Creixams
Department of Diagnostic Imaging,
Hospital General de Granollers,
Barcelona, Spain
X. Codina-Puig
Emergency Department,
Hospital General de Granollers,
Barcelona, Spain

ORIGINAL ARTICLE

Comparison of ultrasound with plain

radiography and CT for the detection
of mediastinal lymphadenopathy in children
with tuberculosis

Abstract Background: Lymphadenopathy, with or without parenchymal abnormality, is the radiological

hallmark of primary tuberculosis
(TB) in children. However, lymph
node enlargement may pass undetected on plain chest radiographs.
Ultrasonography provides complementary information to that obtained by radiographs. Objective: To
assess the clinical value of US for the
detection of mediastinal lymphadenopathy in children with a positive
intradermal tuberculin test. Materials and methods: Thirty-two children
with a mean age of 6 years and a
positive Mantoux test underwent
chest radiography (frontal and lateral) and US (suprasternal and left
parasternal access routes). Chest CT
was performed at the discretion of
the attending physician in six cases.

Introduction
Tuberculosis (TB) has re-emerged as a serious public
health problem in developed countries, particularly
among young adults and children. The diagnosis of TB
in children is often dicult to conrm, because Mycobacterium TB is cultured only in a small percentage of
cases [1, 2]. Whereas the diagnosis of active TB in adults
is mainly bacteriological, in children it is usually epidemiological and indirect. In the absence of a positive
culture, the strongest evidence for TB in a child is recent
exposure to an adult with active disease [3]. Indirect
diagnostic techniques, such as the tuberculin skin test,

Results: Eleven children had clinical

symptoms and 90% a recent contact
with a person with active TB. In
90.5% of children with chest radiographic images compatible with TB,
coincident ndings in the mediastinal US study were found. By comparison, 66.7% of those with normal
chest radiography had evidence of
mediastinal lymphadenopathy on
the US scan. In all cases but one, US
and CT ndings agreed. Conclusions: Mediastinal US is useful for
the detection of enlarged lymph
nodes in children with a positive
tuberculin reaction and normal chest
radiography.
Keywords Mediastinum
TB Lymphadenopathy
Radiography Ultrasound
CT Children

chest radiography and physical examination oer supportive information [4].

Central to the clinical diagnosis of childhood TB is
the chest radiograph and the presence of lymphadenopathy with or without parenchymal involvement is the
single most important diagnostic feature [4]. The nodal
enlargement typically involves the hilar and paratracheal
nodes, with bilateral hilar lymphadenopathy identied
in about 25% of cases. Dierent studies have documented right-sided predominance of lymphadenopathy
and parenchymal changes [5]. Both frontal and lateral
views are necessary to evaluate lymphadenopathy. Of
interest is the fact that enlarged lymph nodes may be

896

detected by ultrafast CT in 60% of children with

tuberculous infection and normal ndings on chest
radiography [6]. However, this technique would not be
available in most cases and the cost is very high. We
report the clinical value of US to detect mediastinal
lymph node involvement in children with a positive
intradermal tuberculin skin test.

Materials and methods

A retrospective review of the medical records of 32 children, 17 boys and 15 girls, with a mean age of 6 years
(range 4 months to 17 years), who had a positive intradermal tuberculin skin test was made. These patients had
been referred to our Department of Paediatrics for workup studies and eventual treatment between 1994 and 2000.
None of the patients had been exposed to BCG vaccination. All patients underwent a thorough history (including
exposure tracing), physical examination, frontal and lateral chest radiographs, and sonographic study of the
mediastinum. The radiographic ndings considered representative of TB included nonspecic localized inltrates, hilar adenitis, localized hyperaeration, atelectasis,
segmental lesions, cavitation, calcication, and localized
pleural eusion. CT of the chest was performed in selected
patients at the discretion of the physicians in charge.
Ultrasonography of the mediastinum was performed
with high-resolution equipment (Logiq 700, General
Electric) using a 5-MHz convex probe. The presence of
one or more masses with an ovoid or round shape and
hypoechoic appearance in the anterior or middle mediastinum was recorded. The anterior mediastinum included the prevascular region, occupied by the thymus
gland and the middle mediastinum, the right paratracheal, supra-aortic, aortopulmonary, and subcarinal regions. On US the normal thymus has a bilobulated
appearance and homogeneous echotexture with some
echogenic strands. It is hypoechoic relative to the thyroid gland and has a smooth, well-dened margin due to
its brous capsule. It is a soft organ that does not
compress neighbouring vascular structures, a characteristic that can help the radiologist to dierentiate it
from mediastinal masses. The normal thymus can vary
considerably in position, extension, size and conguration. In small children, the organ can extend from the
cervical region to the diaphragm. During respiration and
particularly when the child is crying, the thymus can be
above the manubrium and simulate a cervical mass.
The mediastinum was accessed via the suprasternal
and left parasternal approaches [7]. When using the
suprasternal approach, the patient was placed in a
supine decubitus position with a cushion under the
back and the neck slightly extended. The transducer
was placed above the manubrium and titled caudally.
To obtain an oblique sagittal view, the probe was

placed laterally to encounter the space between the

trachea and the sternocleidomastoid muscle. For the
left parasternal approach, the patient was placed in a
left lateral decubitus position to move the mediastinum
downwards and increase the size of the anatomic
acoustic window. Five standard sonographic slices
were used to visualize the complete anterior and middle regions of the mediastinum. Three sonographic
slices were obtained with the suprasternal approach
(oblique coronal, coronal, and oblique parasagittal)
and two with the left parasternal approach (axial and
parasagittal views). The oblique coronal view through
the suprasternal approach was used to visualize the
paratracheal region and to study the aortopulmonary
region; the coronal view was useful for visualizing the
vessels, particularly the SVC; the oblique parasagittal
view visualized the aortopulmonary region. The axial
and parasagittal views through the left parasternal
approach were used to study the subcarinal and prevascular regions.
In all the cases, the number and size (long axis) of
lymph nodes were determined. The following groups
were established arbitrarily: no adenopathy or lymph
nodes <10 mm in diameter (negative, group 0); a single
lymph node >10 mm (positive +, group 1); a single
lymph node >15 mm (positive ++, group 2); a
single lymph node >20 mm (positive +++, group 3);
more than one lymph node >15 mm (positive ++++,
group 4). In the case of clearly matted nodes, the size of
the whole mass was considered. When possible, the size
of each of its components was measured.
For each patient, the results of chest radiography, US
of the mediastinum, and chest CT were compared.

Results
Of the 32 patients who had US studies of the mediastinum, 90% had recent contact with a person with conrmed pulmonary TB. Only 11 (34.4%) children had
clinical manifestations such as fatigue, low-grade fever,
mild cough, weight loss, night sweats, chills, and failure
to thrive. The remaining 21 children were asymptomatic,
but with a positive tuberculin skin test. Pulmonary
radiographic ndings were suggestive of TB in 21 children, negative in nine, and uncertain in two. With regard
to US of the mediastinum, there were ve children in
group 0, 15 in group 1, two in group 2, four in group 3,
and six in group 4. CT of the chest was performed in six
children. Details of ndings of chest radiography,
mediastinal US and chest CT are shown in Table 1 and
in Figs. 1, 2.
In the group of nine children with normal ndings on
chest radiography, US of the mediastinum conrmed
lymphadenopathy in six cases (66.7%) and was negative
in the remaining three. One of these three patients had a

897

Table 1 Reults of chest roentgenograms, mediastinal ultrasonography, and chest CT scans in 32 children with positive intradermal
tuberculin skin test
Case

Sex and age

Chest X-ray

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32

Female, 7 years
Male, 2 years
Male, 2 years
Male, 7 months
Female, 14 years
Female, 16 months
Female, 11 years
Female, 3 years
Male, 16 years
Male, 18 months
Male, 13 years
Male, 8 years
Female, 2 years
Male, 22 months
Female, 4 years
Female, 12 years
Female, 11 years
Male, 4 months
Male, 2 years
Female, 17 years
Female, 14 months
Male, 14 months
Female, 14 months
Male, 7 months
Male, 3 years
Male, 14 years
Male, 2 years
Female, 15 years
Female, 13 months
Female, 15 months
Male, 12 years
Male, 5 years

Normal
Compatible
Compatible
Normal
Compatible
Compatible
Compatible
Compatible
Compatible
Compatible
Doubtful
Normal
Compatible
Compatible
Compatible
Compatible
Doubtful
Compatible
Compatible
Compatible
Normal
Compatible
Compatible
Compatible
Compatible
Normal
Normal
Compatible
Normal
Normal
Normal
Compatible

TB
TB
TB
TB
TB
TB
TB
TB
TB
TB
TB
TB
TB
TB
TB
TB
TB
TB
TB
TB

TB

Mediastinal lymph nodes

by ultrasonography

Mantoux (mm)

Chest CT scan

Present
Present
Present
Present
Present
Present
Present
Present
Present
Present
Present
Present
Present
Present
Present
Present
Present
Present
Present
Present
Present
Present
Present
Absent
Present
Present
Present
Present
Absent
Absent
Absent
Absent

20
16
20
10
10
10
16
10
10
10
10
10
10
28
14
10
10
10
10
22
16
14
10
12
9
20
10
10
10
10
10
14

ND
ND
ND
ND
Conrmatory
ND
Conrmatory
ND
ND
ND
Normal
ND
ND
ND
ND
Conrmatory
ND
ND
ND
ND
ND
ND
ND
ND
ND
Conrmatory
ND
ND
ND
ND
Normal
ND

(+)
(++++)
(+)
(++)
(+)
(+)
(++++)
(+)
(++)
(++++)
(+++)
(++++)
(+)
(+++)
(++++)
(+++)
(+)
(++++)
(+)
(+)
(+++)
(+)
(+)
(+)
(+)
(+)
(+)

ND not done

normal chest CT scan and in the other two, CT examination was not performed. All patients but two with
compatible radiological ndings of TB had visible
mediastinal lymph nodes on US. Therefore, 90.5% (19
out of 21) of patients with pathologic images in the chest
Fig. 1 Results of chest radiography

radiographs, had visible mediastinal lymphadenopathy

in the ultrasonographic study. In the two patients with
doubtful radiological images, ultrasonography conrmed the diagnosis of tuberculous lymphadenopathy in
two.

898

Fig. 2 Results of mediastinal US

The chest CT examination, which was carried out in

six patients, conrmed the results of US in four. In one
patient with normal radiographic ndings and absence
of mediastinal adenopathy, the CT scan was also negative, whereas in the other patient with uncertain radiographic ndings and mediastinal lymphadenopathy in
the ultrasound examination, the CT scan was negative.
This patient, however, was given antituberculous treatment and his clinical symptoms resolved and radiological images cleared. In ve of six (83.3%) patients a
concordance between results of mediastinal ultrasonography and CT examination was observed.
The case of a patient with lymphadenopathy in the
right paratracheal region using the suprasternal approach is shown in Fig. 3. In this case, results of US
were conrmed by CT (Fig. 3). In the case of a 14-yearold patients with active TB involving the left upper lobe,
the suprasternal approach revealed a lymph node,
1.8 cm in diameter, in the aortopulmonary region
(Fig. 4). A lymphadenopathy in the subcarinal space
was detected in a patient with normal chest roentgenogram using the left parasternal approach (Fig. 5).

Discussion
Most TB infections in children and adolescents are
asymptomatic when the tuberculin skin test is positive.
In the present series, only 34.4% of patients had nonspecic symptoms, such as fever, cough, weight loss, and
failure-to-thrive pattern in young infants. All patients
were referred for evaluation because of the tuberculin
skin test and in 90% of them, a recent exposure to an
adult with active disease was present. It should be noted
that there were two patients aged between 15 and
17 years of age. Despite the fact that these patients were
adolescents, they were referred for evaluation to our
department because 18 years is the upper age limit assigned to pediatrics by our health care system. On the

Fig. 3 Right paratracheal lymphadenopathy. a Suprasternal,

oblique coronal US section. The echogenic line originated in the
right upper lobe is displaced by the mass. b Axial CT section
conrming the US ndings. IA innominate artery, LBV left
brachiocephalic vein, AO aorta, TR trachea, RUL right upper
lobe, LN lymph node

other hand, the fact that 32 patients with a tentative

diagnosis of TB had been referred for work-up studies
during the study period indicates that although signicant progress has been made in the control of TB in
developed countries, this communicable disease has not
yet been eradicated. Furthermore, children with primary
tuberculous infection are the reservoir from which future
cases will emerge.
One of the major practical problems in diagnosing
TB in children is that isolation of Mycobacterium TB
from gastric aspirates or sputum is dicult [8]. Sputum
for acid-fast stain and culture is rarely available from
infants and children. Optimal collection of gastric aspirates requires hospitalization to sample the swallowed
secretions that accumulate overnight. However, the

899

Fig. 4a, b 14-year-old patient with active TB. Suprasternal oblique

parasagittal (a) and suprasternal oblique coronal (b) US sections
showing a lymph node 1.8 cm in diameter in the aortopulmonary
region involving the left upper lobe. IA innominate artery, LBV left
brachiocephalic vein, RUL right upper lobe, TR trachea, AO aorta,
LN lymph node, LPA left pulmonary artery, LB left bronchus, LC
left carotid artery; LS left subclavian artery, LB left bronchus, RPA
right pulmonary artery, LA left atrium

sensitivity of acid-fast stain for gastric contents is usually below 10%. The low yield of positive cultures from
gastric aspirates is a result of the small number of
organisms in primary TB in childhood and possible
inadequate techniques for collection of gastric washings.
Therefore, the diagnosis is frequently based solely on

Fig. 5 Positive US with negative radiograph. a Normal frontal

chest radiograph. b Left parasternal axial US section in the same
patient shows lymphadenopathy in the subcarinal space. TH
thymus gland, AO aorta, RPA right pulmonary artery, LC left
carotid artery, LS left subclavian artery, LA left atrium, LB left
bronchus, LBV left brachiocephalic vein, LPA left pulmonary
artery, LN lymph node

detecting typical radiographic abnormalities in a child

with a reactive tuberculin skin test and with history of
contact of an infectious case. The Mantoux method is
helpful in supporting the diagnosis. Although a reaction
of 10 mm induration is the usual cut-point for dening
a signicant reaction, a reaction of 5 mm is considered
signicant for symptomatic children and for recent
contacts with infectious cases [9]. However, a negative

900

reaction in a child who has signs and symptoms compatible with TB does not rule out the diagnosis. In the
present series, indurations ranged between 9 and 28 mm.
Lymphadenopathy, with or without parenchymal
abnormality, is the radiological hallmark of primary TB
in children [10]. Children less than 3 years of age show a
higher prevalence of lymphadenopathy and a lower
prevalence of parenchymal abnormalities compared with
children 415 years [5]. In early childhood, lymphadenopathy as the sole radiological manifestation of disease
was seen in 49% of cases versus 9% in late childhood
and adolescence according to data reported by Leung
et al. [11]. Bronchi in infants are of smaller calibre and
more easily compressed by enlarging hilar lymph nodes.
As the hilar lymph nodes enlarge, bronchial obstruction
may occur and signs of air trapping may develop. Although hilar lymphadenopathy may be the only suggestive nding of TB in the chest radiographs, in the
present study, 66.7% of patients with chest radiographs
considered unrevealing showed mediastinal lymphadenopathy in the ultrasound examination. Lymph nodes
can sometimes be dicult to visualize on frontal plain
radiographs. Occasionally, lymphadenopathy is visible
only on the lateral lm [12]. Apical-lordotic views may
aid in visualizing lesions obscured by the heart. When no
lymphadenopathy is present on the standard radiographic examination of the chest, special imaging techniques such as CT may be of particular value [13]. It has
been shown that CT scan may reveal mediastinal adenopathies which are not evident on the chest radiograph
[6]. Ultrafast CT scanning, however, is costly, not
available in many institutions, includes radiation, and
may require the use of sedation in young children. In
contrast, US is much less expensive, the use of sedatives

or contrast medium is not necessary, and can be easily

obtained both in the hospital and in primary care settings. Although subcarinal adenopathy has recently been
reported to be the most common site of lymphadenopathy in children with TB [14], we have detected small
adenopathies more frequently in the paratracheal region
and aortopulmonary window because of a better echographic access. In the subcarinal region, we have documented large lymph nodes due to limitations in the
echographic access and artifacts (e.g., the oesophagus).
As far as we are aware, no previous study regarding
the usefulness of mediastinal ultrasonography for the
diagnosis of lymphadenopathy in TB in children has
been published. For this reason, the present results
cannot be compared to those reported by others.

Conclusions
In the present series of 32 patients with positive tuberculin skin test, 90.5% of those with chest radiographic
images compatible with TB had coincident ndings in
the mediastinal ultrasonographic study. On the other
hand, 66.7% of those with normal chest radiography
had evidence of mediastinal lymphadenopathy on US.
In all cases but one, US and CT ndings agreed. In view
of the usefulness of US of the mediastinum for the
diagnosis of lymphadenopathy in children with TB, this
non-invasive method could also be of value in the control and follow-up of children receiving antituberculous
chemotherapy.
Acknowledgement We thank Marta Pulido, MD, for editing the
manuscript and for editorial assistance.

References
1. Vallejo JG, Ong LT, Starke JR (1994)
Clinical features, diagnosis, and treatment of TB in infants. Pediatrics 94:17
2. American Thoracic Society (1990)
Diagnostic standards and classication
of TB. Am Rev Respir Dis 142:725735
3. Hilman BC (1993) Pulmonary TB and
tuberculous infection in infants, children, and adolescents. In: Hilman BC
(ed) Pediatric respiratory disease: diagnosis and treatment. Saunders, Philadelphia, pp 311319
4. Snider DE, Rieder HL, Combs D et al
(1988) TB in children. Pediatr Infect Dis
7:271278
5. Burroughs M, Beitel A, Kawamura A
et al (1999) Clinical presentation of TB
in culture-positive children. Pediatr Infect Dis 18:440446

6. Delacourt C, Mani TM, Bonnerot V

et al (1993) Computed tomography with
normal chest radiograph in tuberculous
infection. Arch Dis Child 69:430432
7. Lucaya J, Strife J (2001) Pediatric chest
imaging: chest imaging in infants and
children. Springer, Berlin Heidelberg
New York, pp 125
8. Starke JR (1988) Modern approach to
the diagnosis and treatment of TB in
children. Pediatr Clin N Am 35:441464
9. American Thoracic Society and the
Centers for Disease Control and Prevention (2000) Diagnostic standards
and classication of TB in adults and
children. Am J Respir Crit Care Med
161:13761395

10. Omlor GJ (2001) Pulmonary lymphadenopathy. Pediatr Infect Dis 20:437

438
11. Leung AN, Muller NL, Pineda PR et al
(1992) Primary TB in childhood:
radiographic manifestations. Radiology
182:8791
12. Smuts NA, Beyers N, Gie RP et al
(1994) Value of the lateral chest radiograph in TB in children. Pediatr Radiol
24:478480
13. Vallejo JG, Starke JR (1996) Intrathoracic TB in children. Semin Respir Dis
11:184195
14. Andronikou S, Joseph E, Lucas S et al
(2004) CT scanning for the detection of
tuberculous mediastinal and hilar
lymphadenopathy in children. Pediatr
Radiol 34:232236