European Journal of Internal Medicine 66 (2019) 29–34

Contents lists available at ScienceDirect

European Journal of Internal Medicine

journal homepage: www.elsevier.com/locate/ejim

Original Article

Lung ultrasonography in pulmonary tuberculosis: A pilot study on T

diagnostic accuracy in a high-risk population
⁎
M. Montuoria, , F. Casellab, G. Casazzac, F. Franzettia, P. Pinib, C. Invernizzib, D. Torzillob,
G. Rizzardinid, M. Gallia, C. Cogliatib
a
Infectious Diseases Unit, Department of Biomedical and Clinical Sciences "Luigi Sacco" University of Milan, ASST-FBF-Sacco, Italy
b
Department of Internal Medicine, ASST-FBF-Sacco, Italy
c
Dipartimento di Scienze Biomediche e Cliniche “Luigi Sacco”, Università degli Studi di Milano, Italy
d
First Division of Infectious Diseases, ASST-FBF-Sacco, Milano, Italy

A R T I C LE I N FO A B S T R A C T

Keywords: Objectives: The validity of lung ultrasound (LUS) in the diagnosis of interstitial or focal lung pathologies is well
Pulmonary tuberculosis documented, we assessed its accuracy in the diagnosis of pulmonary tuberculosis (PTB).
Diagnostic imaging Methods: Sonographic signs suggestive of PTB and their diagnostic accuracy were evaluated in patients admitted
Lung ultrasonography with clinical suspicion of PTB. Consolidations, subpleural nodules, pleural thickenings or irregularities and
pleural effusion were assessed. LUS signs significantly associated with PTB in the univariate analysis (p < .05)
were entered in a multivariate logistic regression model.
Results: PTB was confirmed in 51 out of 102 patients. Multiple consolidations (OR 3.54, 95%CI 1.43–8.78),
apical consolidations (OR 9.65, 95%CI 3.02–30.78), superior quadrant consolidations (OR 4.01, 95%CI
1.76–9.14), and subpleural nodules (OR 5.29, 95%CI 2.27–12.33) were significantly associated with PTB di-
agnosis. Apical consolidation (OR 9.67, 95%CI 2.81–33.25, p 0.003) and subpleural nodules (OR 5.30, 95%CI
2.08–13.52, p 0.005) retained a significant association in a multivariate model, with an overall accuracy of
0.799.
Conclusions: Our data suggest a possible role of LUS in the diagnosis of PTB, a high burden pathological con-
dition for which the delay in diagnosis still represents a critical point in the control of the disease.

1. Introduction WHO [3].

Chest X-Ray (CXR), the mainstay for the radiological identification,
According to the World Health Organization (WHO), in 2017, 10 is burdened by low specificity, radiation exposure and it's not promptly
million individuals became ill with tuberculosis (TB) and 1.6 million available in health systems with limited resources [4].
died [1]. The epidemiology of TB varies substantially around the world Bedside ultrasound is a safe, portable, versatile and cost-effective
with highest rates observed in sub-Saharan Africa, India, and the is- imaging modality [5]. The WHO has acknowledged that it should be
lands of Southeast Asia. Missed or late diagnosis of TB is still significant available worldwide to assist the clinician in the diagnostic process [6].
for both low- and high-income countries [2]. In particular, lung ultrasonography (LUS) has been applied in the di-
The most common available tests present many pitfalls and micro- agnosis of several lung pathologies as pneumothorax [7], interstitial
biologic confirmation is often lacking. Regarding pulmonary tubercu- lung diseases [8], pleuritis and, of course, pleural effusion [9]. Its di-
losis (PTB), smear microscopy has low sensitivity, culture methods take agnostic accuracy in recognition of pneumonia is demonstrated to be
several weeks for the results and rapid molecular tests are not widely similar to computed tomography (CT) [10]. Thus, LUS is a potential
available although they are the recommended methods according to useful diagnostic tool for the diagnosis of PTB, particularly in

Abbreviations: TB, Tuberculosis; PTB, Pulmonary tuberculosis; WHO, World Health Organization; CXR, Chest X-Ray; LUS, Lung ultrasonography; CT, Computed
tomography; TST, Tuberculin Skin Test; HIV, Human immunodeficiency virus; US, Ultrasound; LR+, Likelihood ratio; LR-, Negative likelihood ratio; PPV, Positive
predictive value; NPV, Negative predictive value; CI, Confidence intervals; OR, Odds ratios; ROC, Receiver-operating characteristic
⁎
Corresponding author at: Infectious Diseases Unit, Department of Biomedical and Clinical Sciences "Luigi Sacco" University of Milan, ASST-FBF-Sacco, Via dei
Sormani 12, 20144 Milano (MI), Italy.
E-mail address: michelemontuori6@gmail.com (M. Montuori).

https://doi.org/10.1016/j.ejim.2019.06.002
Received 1 February 2019; Received in revised form 15 May 2019
Available online 22 June 2019
0953-6205/ © 2019 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved.
M. Montuori, et al. European Journal of Internal Medicine 66 (2019) 29–34

geographic areas and in situations where radiological or laboratory

equipments are not readily available. [11,12]
Consolidations, subpleural nodules, pleural thickenings, fibrosis,
pleural effusion and pneumothorax together with miliary pattern have
been reported in descriptive studies on US in PTB [11,13,14].
Nevertheless, the diagnostic accuracy of LUS in diagnosing PTB has
never been studied. Aim of this study is to assess LUS signs associated
with PTB and evaluate their diagnostic accuracy.

2. Methods

This interventional non-pharmacologic prospective study was con-

ducted at the Luigi Sacco Hospital, ASST Fatebenefratelli-Sacco, Milan,
Italy, a university-affiliated hospital serving as a referral centre for the
treatment of infectious diseases. The study is consistent with the prin-
ciples of the Declaration of Helsinki on clinical research involving
human subjects and according to the quality standards of Good Clinical
Practice and it was approved by the Ethics Committee Milano Area 1
(project approval no. 40082/2017). Written informed consent was
obtained from each patient included in the study.
All consecutive patients admitted to the Department of Infectious
Diseases presenting with clinical suspicion of PTB in a 15-months
period were included. According to international guidelines suspicion
of PTB was defined by the presence of at least one of the following
clinical scenarios: 1) any patients with a cough of ≥ 2–3 weeks duration
with at least one additional symptom including fever, night sweats,
weight loss or hemoptysis; 2) any patient at high risk for TB with un-
explained illness including respiratory symptoms of ≥ 2–3 weeks
duration; 3) any patient with HIV infection and unexplained cough and
fever; 4) any patient at high risk for TB with a diagnosis of community Fig. 1. Convex probe placed coronally on the supraclavicular region for vi-
acquired pneumonia who has not improved after seven days of treat- sualization of the lung apex.
ment; 5) any patient at high risk for TB with incidental finding on CXR
suggestive of TB even if symptoms are minimal or absent [15]. CXR was
Department experienced in clinical ultrasonography, blinded to the
considered suggestive of PTB in case of infiltrates with or without ca-
patient clinical information and radiological examinations. Sonography
vitation in the upper lobes or involving the superior segments of the
was conducted using a portable US device Sonosite M-Turbo (Fujifilm
lower lobes [16]. Risk factors for TB were: any known contact or family
SonoSite, Inc., USA) with a 5–2 MHz convex probe. Patients were stu-
member affected by TB, history of previous TB or any TB treatment or
died in the supine position for evaluation of the anterior thorax and in
positive TST results, high risk congregate settings, substance abuse,
seated position for evaluation of posterior and lateral thorax. If the
country of birth, immunosuppressive therapy and immunodeficiency-
patient was not able to maintain the seated position the exam was
associated diseases with particular attention to HIV status.
performed in the supine position and lateral decubitus to assess the
We considered exclusion criteria a previously known interstitial
posterior areas. Patients were examined by longitudinal and oblique
lung disease, active PTB diagnosed before the admission to the hospital,
scans. During the exam of the posterior regions, the patients were asked
LUS contraindications (subcutaneous emphysema, burns localized on
to raise the arms above the head in order to displace the shoulder blade
the thorax or skin diseases that interfere with the methodology) or re-
and uncover the area of the lung which is usually masked by the bone.
fusal of consent. Primary assessment of the enrolled patients consisted
The examination of the lung apexes was performed by applying the
in routine evaluation with medical history and physical examination.
probe vertically between the clavicle and the trapezius muscle ante-
According to WHO [17], PTB refers to any bacteriologically con-
riorly (Fig. 1) and directly on the cranial part of the trapezius muscle on
firmed or clinically diagnosed case of TB involving the lung par-
the back. The whole surface of the chest was sistematically analized. To
enchyma or the tracheobronchial tree. Miliary TB and patients with
describe and record echographic signs each hemithorax was divided
both PTB and extrapulmonary TB were classified as case of PTB. A
following anatomical lines in 6 areas (4 anterior and 2 posterior areas)
bacteriologically confirmed PTB case is one from whom sputum/
as illustrated in Fig. 2.
broncho-alveolar lavage had a positive smear microscopy, culture or
Taking into consideration the descriptive studies [11,13,14], the
nucleic acid amplification tests for M. tuberculosis. Samples were sent to
following US findings were considered and recorded: consolidation
the Microbiology Laboratory of L. Sacco Hospital and examined for
(single or multiple); subpleural nodules (circular or ellipsoidal hy-
acid-fast bacilli using the Ziehl-Neelsen method. The presence of M.
poechoic subpleural lesions < 15 mm (measured as extension on the
tuberculosis complex was determined also by real-time polymerase chain
pleural line and as depth)); irregularities or focal thickenings of the
reaction (PCR), after DNA extraction, using Xpert MTB/RIF assay
pleura; presence of pleural effusion. Apical consolidations were defined
(Cepheid, Sunnyvale, CA). Mycobacterial cultures were performed on
as visible by scanning the supraclavicular region or within the first two
both solid and liquid media, using Lowenstein-Jensen medium and
intercostal spaces or the body of the first thoracic vertebra posteriorly.
BACTEC MGIT (Mycobacteria Growth Indicator Tube; BD, Sparks, MD,
In the first half of enrolled populations patients with cavitations at
USA), respectively. In clinically diagnosed PTB bacteriological con-
CXR/CT images underwent a second non-blind US evaluation to iden-
firmation was not obtained but the physician diagnosed active PTB on
tify possible US characteristics of cavitated lesions. Thus, in the second
clinical/radiological basis and treated the patient with a full course
half, we blindly searched for recurrent US characteristics associated
anti-TB chemotherapy.
with cavitated consolidations.
Patients underwent LUS within 3 days after admission to the ward.
Electronic images of each examination were acquired and stored. At
The examination was performed by a physician of the Internal Medicine

30
M. Montuori, et al. European Journal of Internal Medicine 66 (2019) 29–34

Fig. 2. Anatomical Lines identifying superior and inferior areas anteriorly (left panel) and posteriorly (right panel).
R4: fourth rib.
T3: third thoracic vertebra.

the end of each exam, the physician was required to fill a predefined 4. Results
form where an accurate description and localization of the pathological
findings were outlined. One hundred-ten consecutive patients with clinical suspicion of PTB
were enrolled. Eight patients have been excluded: 2 of them had a
known interstitial lung disease while in 6 patients the diagnosis of PTB
3. Calculations predated the day of admission. One hundred-two patients were thus
included and completed the study. PTB was confirmed in 51 (44 by
Continuous variables were reported as mean (standard deviation) or culture and 7 on clinical basis) with a disease prevalence of 50% in our
median (range), as appropriate. Categorical data were expressed as population.
counts (percentages). For group comparison, the Student's t-test or Study population characteristics are reported in Table 1. Ninety-six
Mann-Whitney test were used for continuous variables, as appropriate, (94.1%) of the patients had at least one risk factor for TB. PTB patients
and the Fisher exact test was used for categorical variables. were younger, more frequently foreign born and living in overcrowded
Diagnostic accuracy of each echographic sign was assessed and conditions (homelessness, incarceration, refugee camp). We observed
sensitivity, specificity, positive likelihood ratio (LR+), negative like- longer duration of symptoms in the PTB patients and a lower white
lihood ratio (LR-), positive predictive value (PPV) and negative pre- blood cells count. Out of the overall population, 28 patients (27%)
dictive value (NPV), with their 95% confidence intervals (95% CI), tested HIV positive, of which 11 have been diagnosed with PTB. Dis-
were calculated. Then a logistic regression approach was adopted to charge diagnosis are shown in Table 2. The most frequent non-PTB
find the combination of echographic signs with the best diagnostic diagnosis was bacterial pneumonia, followed by pulmonary neoplasia,
accuracy. First, univariate logistic models were fitted to assess the as- pneumocystosis and aspergillosis. In 10 patients no definite diagnosis
sociation between each of the sign and PTB diagnosis. Then, a multi- was obtained, TB was nonetheless excluded.
variate analysis was performed by considering only the signs sig- Univariate analysis of LUS findings showed significant association
nificantly associated with PTB diagnosis at univariate analysis and with PTB diagnosis for multiple consolidations, apical consolidations,
adopting a stepwise strategy in order to find the best model. Results of superior quadrant involvement and subpleural nodules. These variables
logistic regression analysis were reported as odds ratios (OR) with 95% were then included in the multivariate model where apical consolida-
CI. The c-statistic, which can be interpreted as the area under the re- tion (OR 9.67, 95%CI 2.81–33.25, p 0.003) and subpleural nodules (OR
ceiver-operating characteristic (ROC) curve, was used to assess the 5.30, 95%CI 2.08–13.52, p 0.005) were found to be independently
overall accuracy of univariate and multivariate models. Finally, in correlated with the diagnosis of PTB (Table 3, Fig. 3). Calculation of
order to translate the results of the models into clinical practice, diag- sensitivity and specificity were performed for each echographic sign
nostic accuracy was assessed considering only the echographic signs (Table 4) and for the model constructed by using the two variables
that were found to be statistically significant at the multivariate ana- significantly correlated with PTB in the multivariate analysis (apical
lysis. Sensitivity and specificity were calculated for combinations of consolidation and subpleural nodules) (Table 5). We observed a speci-
signs, according to the following two scenarios: positive patients are all ficity of 96% when the apical consolidations and subpleural nodules
those with a positive result in at least one of the signs; positive patients were found in the same patient, with sensitivity of 31%. Sensitivity of
are all those with positive results in all the signs. 86% and specificity of 63% were attained when the test was considered
P values < .05, two sided, were considered statistically significant. positive with the detection of at least one of those signs. The c-statistic
All the statistical analyses were performed using SAS statistical software for the model with two variables was 0.799.
(release 9.4). We analysed the characteristics of subpleural nodules. Maximal
pleural extension and depth measured 8 ± 3 mm (range 3–15) x
8 ± 2 (range 3–14) in TB patients and 7 ± 3 mm (range 4–14) x
3.1. Sample size
6 ± 2 mm (range 3–11) in non-TB. Eighteen out of 37 TB patients
presenting subpleural nodules had bilateral involvement. In twenty-one
We were interested in the assessment of the ability of echographic
patients nodules were confined to superior quadrants, while 14 patients
signs in ruling-out PTB. We expected that the sign with the highest
presented diffuse nodules (superior and inferior quadrants). In only two
sensitivity would have 80% sensitivity. Assuming a PTB prevalence of
PTB patients (5%) subpleural nodules were limited to inferior quad-
45%, the inclusion of at least 100 consecutive patients (45 with PTB)
rants. Five out of the 17 patients without TB (29%) presented nodules
would have provided an estimate of the anticipated sensitivity with
just in the inferior quadrant while in 11 the superior quadrants were
95% CI from 65% to 90%, that can be considered sufficiently precise.

31
M. Montuori, et al. European Journal of Internal Medicine 66 (2019) 29–34

Table 1 involved. Five of these latter presented a previous PTB diagnosis. In

Baseline population characteristics. non-TB patients nodules were frequently found in the context of diffuse
PTB non PTB p Value pleural irregularities. A diffuse interstitial pattern and multiple sub-
pleural nodules were identified in the two cases presenting miliary TB.
n = 51 n = 51 On chest US pleural effusion was inversely associated with PTB and
mainly represented in patients with a different discharge diagnosis (e.g.
Sex M, No. (%) 37 (72) 30 (59) 0,2105
Age, years, Median (IQR) 34 (24–49) 49 (39–60) 0,0001 pneumonia, lung cancer). Pleural fluid characteristics did not differ
Origin, No. (%) between TB and non-TB patients: in 5 out of 10 TB patients and in 5 out
Italy 9 (18) 29 (57) of 13 non-TB patients' pleural fluid presented septations and locula-
Others 42 (82) 22 (43) < 0,0001 tions.
Risk factors
In the first half of population we performed a second non-blind
HIV +, No. (%) 11 (22) 17 (33) 0,2672
CD4+ cell/mm3, Median 150 (74–279) 255 (158–749) 0,0776 ultrasonography in 13 patients with cavitations at CXR/CT. Six of them
(IQR) had markedly hypoechogenic or anechogenic areas inside a consolida-
Smoke, No. (%) 21 (41) 24 (47) 0,8417 tion, without bronchograms or signs of vascularization at color Doppler.
Type II Diabetes Mellitus, 6 (12) 4 (8) 0,7409
In the second half of population we checked the suspicion of cavitations
No. (%)
Alcohol abuse, No. (%) 15 (30) 11 (22) 0,496
in all 58 patients. The presence of hypoechogenic or anechogenic areas
Active IV abuse, No. (%) 4 (8) 8 (16) 0,3573 inside a consolidation, without bronchograms or signs of vasculariza-
Malnutrition, No. (%) 3 (6) 0 (0) 0,2426 tion at color Doppler was identified in 12 cases. Two of them were not
High risk congregate 14 (28) 3 (6) 0,0065 confirmed by CT scan. Radiology found cavitations in 24 patients out of
settings, No. (%)
52, 18 by CRX and 6 more by CT. Out of these 24, 15 were TB patients
Chronic kidney disease, No. 1 (2) 0 (0) 0,9999
(%) and 9 non-TB.
Immunosuppressive 1 (2) 4 (8) 0,3624
therapy, No. (%) 5. Discussion
Active cancer, No. (%) 0 (0) 5 (10) 0,0564
Previous Tuberculosis, No. 7 (14) 14 (27) 0,1406
(%)
To our knowledge this is the first study dealing with the diagnostic
Symtoms, No. (%) accuracy of LUS in the diagnosis of PTB in adults. We assessed the
Fever 34 (66) 35 (69) 0,9999 echographic signs associated with PTB showing that apical consolida-
Cough 37 (72) 35 (69) 0,8282 tion and subpleural nodules were independently correlated with the
Sputum production 19 (37) 16 (31) 0,6769
diagnosis of PTB. They had a good sensitivity when considered in-
Haemoptysis 13 (25) 15 (29) 0,8247
Generalized weakness 19 (37) 18 (35) 0,9999 dividually and an excellent specificity when present simultaneously.
Weight loss 13 (25) 10 (20) 0,6362 Ultrasound is a potent tool for the diagnosis, screening, and mon-
Night sweats 12 (24) 10 (20) 0,8102 itoring of treatment response for a broad and still expanding range of
Thoracic pain 16 (32) 17 (33) 0,9999
infectious diseases [18]. In areas with high prevalence of TB the de-
Pharyngodynia 2 (4) 1 (2) 0,9999
Lynphadenopathy 4 (8) 1 (2) 0,3624
tection of echographic signs may suffice to initiate a treatment for ex-
No symptoms 1 (2) 2 (4) 0,577 trapulmonary TB without histologic confirmation [19]. Point-of-care
Duration of symptoms, 30 (14–62) 15 (5–45) 0,009 ultrasonography is increasingly used in resource-limited settings be-
Median (IQR) cause of its diagnostic accuracy associated with non-invasiveness, re-
Positive lung physical 30 (59) 26 (51) 0,5508
peatability and easy transportability. LUS has emerged as a pivotal part
examination, No. (%)
Laboratory findings of point of care ultrasound: its sonographic signs are simple to learn and
CRP, mg/dL, Median (IQR) 49 (21–99) 38 (8–102) 0,4257 analysis of artefacts and morphologic images allows accurate diagnosis
Leukocytes, cell/mm3, 7700 9710 0,0092 in many lung pathologies [9]. The WHO states that plain radiography
Median (IQR) (5730–9530) (7700–11,660) and ultrasonography, singly or in combination, meet up to 90% of all
Hb, g/dL, Mean (SD) 12,0 (2) 12,3 (1,8) 0,2896
LDH, UI/l, Median (IQR) 216,5 207 0,8127
imaging needs in developing countries [20].
(178,5–257) A screening test with accuracy in the order of CXR for TB ab-
normalities but without the need for CXR equipment would be an
PTB: Pulmonary Tuberculosis; IV: Intravenous; CRP: C reactive protein; Hb: highly desirable asset to facilitate TB screening and a way to improve
Hemoglobin; LDH: Lactate dehydrogenase; IQR: interquartile range; SD: stan- TB case detection [4] which is a key component of the End TB Strategy
dard deviation. [21].
In the literature, with the exception of case reports [11,13,22] and a
Table 2 paediatric study [23], there is only one study describing the echo-
Discharge diagnosis. graphic picture of PTB in adult patients already diagnosed with the
n (%) disease [14]. The authors showed that subpleural nodules were the
most frequent echographic sign in patients with active PTB in Sub-Sa-
Pulmonary Tuberculosis 51 (50)
haran Africa, being often bilateral and randomly distributed. In the
Pneumonia 19 (18,6)
Lung lesions of unknown origin 10 (9,8) same study consolidations identified by LUS were described as irregular
Pneumocystosis 3 (2,9) consolidations of relatively homogeneous texture, indistinguishable
Lung cancer 3 (2,9) from bacterial pneumonia [14].
Aspergillosis 3 (2,9) In our study apical consolidations and subpleural nodules were the
Bronchiectasis 2 (1,9)
Upper respiratory infections 2 (1,9)
two echographic signs independently correlated with the diagnosis of
Hematologic malignancy 2 (1,9) PTB. The association of LUS finding of subpleural nodules with PTB is of
Atypical mycobacterial diseases 2 (1,9) particular interest because, as previously described [14], subpleural
Pleural disease other than TB 2 (1,9) nodules are often missed by CXR. Among our 17 patients with sub-
Extrapulmonary TB 1 (0,9)
pleural nodules but no PTB diagnosis, 7 had a previous TB infection and
Miscellaneous 2 (1,9)
6 were immunodepressed. Furthermore 7 of them had pulmonary dis-
TB: Tuberculosis. ease detected on CT scan such as emphysema with bronchiectasis,
pulmonary fibrosis, pulmonary neoplasm and aspergillosis. Distribution

32
M. Montuori, et al. European Journal of Internal Medicine 66 (2019) 29–34

Table 3
Diagnostic accuracy of LUS signs.
Univariate Multivariatea

n (%) PTB (%) non PTB (%) OR (95% CI) p-value c-statistic OR (95% CI) p-value c-statistic

Consolidations 73 (72) 40 (77) 33 (64) 1,98 (0,82-4,78) 0,1274 0,569 0,799

Multiple consolidations 31 (30) 22 (43) 9 (18) 3,54 (1,43-8,78) 0,0064 0,627
Apical consolidation 27 (26) 23 (45) 4 (8) 9,65 (3,02-30,78) < 0,0001 0.686 9,67 (2,81-33,25) 0,003
Superior consolidation 53 (52) 35 (69) 18 (35) 4,01 (1,76-9,14) 0,001 0,667
Subpleural nodules 54 (53) 37 (73) 17 (33) 5,29 (2,27-12,33) < 0,0001 0,696 5,30 (2,08-13,52) 0,005
Pleural irregularities 70 (69) 37 (73) 33 (65) 1,44 (0,62-3,34) 0,3943 0.539
Pleural effusion 23 (23) 10 (20) 13 (25) 0,71 (0,28-1,82) 0.4782 0.529
Cavitations (n = 58 PTB = 30) 12 (21) 9 (30) 3 (11) 3,57 (8,85-14,92) 0,081 0,596

PTB: Pulmonary Tuberculosis; OR: Odds ratio; CI: Confidence Interval; AUC: Area Under Curve.
a
Only the variables found to be statistically significant after stepwise strategy are reported in the table.

Fig. 3. (a,b) Single subpleural nodule (c) multiple subpleural nodules (d,e,f) apical consolidations.

Table 4
Diagnostic accuracy of LUS signs.
SE CI 95% SP CI 95% LR+ LH- PPV CI 95% NPV CI 95%

Consolidations 78,4 64,7-88,7 35,3 22,4-49,9 1212 0,945-1554 0,611 0,322-1661 54,8 42,7-66,5 62,1 42,3-79,3
Multiple consolidations 43,1 29,3-57,8 82,4 69,1-91,6 2444 1249-4784 0,69 0,527-0,905 71 52–85,8 59,2 46,8-70,7
Apical consolidation 45,1 31,1-59,7 92,2 81,1-97,8 5,75 2,14-15,449 0,596 0,459–0,774 85,2 66,3-95,8 62,7 50,7-73,6
Superior consolidation 68,6 54,1-80,9 64,7 50,1-77,6 1944 1284-2946 0,485 0,308-0,763 66 51,7-78,5 67,3 52,5-80,1
Subpleural nodules 72,5 58,3-84,1 66,7 52,1-79,2 2176 1425-3323 0,412 0,253-0,67 68,5 54,4-80,5 70,8 55,9–83
Pleural irregularities 72,5 58,3-84,1 35,3 22,4-49,9 1121 0,861-1460 0,778 0,435-1390 52,9 40,6-64,9 56,3 37,7-73,6
Pleural effusion 19,6 9,8-33,1 74,5 60,4-85,7 0,769 0,372-1592 1079 0,874-1331 43,5 23,2-65,5 48,1 36,7-59,6

SE: Sensitivity; CI: Confidence interval; SP: Specificity; LR: Likelihood ratio; PPV: Positive predictive value; NPV: Negative predictive value.

Table 5
Diagnostic accuracy of the model based on LUS signs.
TP FP FN TN SE CI 95% SP CI 95% LR+ CI 95% LR- CI 95% PPV CI 95% NPV CI 95%

Apical consolidation OR subpleural 44 19 7 32 86 74–94 63 48–76 2,32 1,6-3,36 0,22 0,11-0,45 69,80% 57–80,8% 82,10% 66,5-92,5%
nodules
Apical consolidation AND subpleural 16 2 35 49 31 19–46 96 87–100 8 1,94- 0,71 0,59-0,87 88,89% 65,3-98,6% 58,30% 47,1–69%
nodules 33,03

TP: True positive; FP: False positive; FN: False negative; TN: True negative; SE: Sensitivity; SP: Specificity; LR: Likelihood ratio; PPV: Positive predictive value; NPV:
Negative predictive value.

33
M. Montuori, et al. European Journal of Internal Medicine 66 (2019) 29–34

of the nodules seems to be different in the PTB and non PTB population. interest.
The superior quadrant involvement of the subpleural nodules seems to
be associated with active PTB or history of TB disease in our population. Acknowledgements
In previous studies subpleural nodules were also described in crypto-
coccosis, aspergillosis, sarcoidosis, cytomegalovirus and pneumocystis We would like to thank Maria Luigia Malerba for her help in plan-
pneumonia [24]. It is possible to hypothesize that in patients with ning US examinations.
symptoms compatible with PTB and without the confounding presence
of chronic pulmonary disease the detection of subpleural nodules with References
superior involvement may narrow the differential diagnosis and in-
crease the pre-test probability of PTB diagnosis. [1] World Health Organization. Global tuberculosis report. 2018.
Apical consolidations reached a high specificity with a lower sen- [2] Sreeramareddy CT, Panduru KV, Menten J, Van Den Ende J. Time delays in diag-
nosis of pulmonary tuberculosis: a systematic review of literature. 2009. https://
sitivity. We systematically visualized the lung apexes by performing a doi.org/10.1186/1471-2334-9-91.
specific US projection usually described only in US-guided procedures [3] Cudahy P, Shenoi SV. Diagnostics for pulmonary tuberculosis. Postgrad Med J
on brachial plexus. The only three false positives cases showed parti- 2016;92:187–93. https://doi.org/10.1136/postgradmedj-2015-133278.
[4] World Health Organization. Systematic screening for active tuberculosis : Principles
cular conditions: one was diagnosed with pneumonia with multiple and recommendations. World Health Organization; 2013.
consolidations in Hodgkin's lymphoma; one patient had residual scar [5] Bouhemad B, Zhang M, Lu Q, Rouby J-J. Clinical review: bedside lung ultrasound in
tissue from a spontaneous pneumothorax; the third one was a HIV critical care practice. Crit Care 2007;11:205. https://doi.org/10.1186/cc5668.
[6] Training in diagnostic ultrasound: essentials, principles and standards. Report of a
positive patient with multiple supraclavicular lymphadenopathies in
WHO Study Group. World Health Organ Tech Rep Ser 1998. 875:i-46; back cover.
myeloproliferative disease. [7] Soldati G, Testa A, Sher S, Pignataro G, La Sala M, Silveri NG. Occult traumatic
The high specificity of apical consolidation was further increased by pneumothorax: diagnostic accuracy of lung ultrasonography in the emergency de-
partment. Chest 2008;133:204–11. https://doi.org/10.1378/chest.07-1595.
the concomitant presence of subpleural nodules, indicating LUS as a
[8] Tardella M, Gutierrez M, Salaffi F, Carotti M, Ariani A, Bertolazzi C, et al.
useful rule-in test in these cases. Ultrasound in the assessment of pulmonary fibrosis in connective tissue disorders:
In our study LUS had a reliable sensitivity for PTB diagnosis when correlation with high-resolution computed tomography. J Rheumatol
considering the presence of at least one between apical consolidation 2012;39:1641–7. https://doi.org/10.3899/jrheum.120104.
[9] Volpicelli G, Elbarbary M, Blaivas M, Lichtenstein DA, Mathis G, Kirkpatrick AW,
and subpleural nodules. Moreover, it should be emphasized that no et al. International evidence-based recommendations for point-of-care lung ultra-
patient with PTB had completely negative ultrasound examination. sound. Intensive Care Med 2012;38:577–91. https://doi.org/10.1007/s00134-012-
Considering the 7 patients who did not show either apical consolidation 2513-4.
[10] Reissig A, Copetti R, Mathis G, Mempel C, Schuler A, Zechner P, et al. Lung ul-
or subpleural nodules, 4 of them had pleural effusion, 5 had pleural trasound in the diagnosis and follow-up of community-acquired pneumonia: a
irregularities, 3 had multiple consolidations. prospective, multicenter, diagnostic accuracy study. Chest 2012;142:965–72.
These data suggest that the implementation of other sonographic https://doi.org/10.1378/chest.12-0364.
[11] Di Gennaro F, Pisani L, Veronese N, Pizzol D, Lippolis V, Saracino A, et al. Potential
signs besides apical consolidation and subpleural nodules may further diagnostic properties of chest ultrasound in thoracic tuberculosis-a systematic re-
increase the rule-out value of LUS. view. Int J Environ Res Public Health 2018;15. https://doi.org/10.3390/
As previously shown [14] hypoechogenic or anechogenic areas in- ijerph15102235.
[12] Sippel S, Muruganandan K, Levine A, Shah S. Review article: use of ultrasound in
side a consolidation, without bronchograms or signs of vascularisation the developing world. Int J Emerg Med 2011;4:72https://doi.org/10.1186/1865-
were the US findings more frequently associated with cavitations. 1380-4-72.
However these signs does not appear to be enough sensitive probably [13] Heuvelings CC, Bélard S, Janssen S, Wallrauch C, Grobusch MP, Brunetti E, et al.
Chest ultrasonography in patients with HIV: a case series and review of the lit-
due to the high number of lesions not reaching the pleura and the
erature. Infection 2016;44:1–10. https://doi.org/10.1007/s15010-015-0780-z.
presence of air in cavitation. [14] Agostinis P, Copetti R, Lapini L, Badona Monteiro G, N'Deque A, Baritussio A. Chest
ultrasound findings in pulmonary tuberculosis. Trop Dr 2017;47:320–8. https://doi.
6. Conclusions org/10.1177/0049475517709633.
[15] Taylor Z, Nolan CM, Blumberg HM, American Thoracic Society, Centers for Disease
Control and Prevention, Infectious Diseases Society of America. Controlling tu-
Low cost, easy to use and to transport, rapid learning curve, absence berculosis in the United States. Recommendations from the American Thoracic
of radiation exposure and repeatability are recognized advantages of Society, CDC, and the Infectious Diseases Society of America. MMWR Recomm
Reports Morb Mortal Wkly Report Recomm Reports 2005;54:1–81.
ultrasonography making it a useful point of care diagnostic and mon- [16] Daley C, Gotway M, Jasmer R. Radiographic Manifestations of Tuberculosis: A
itoring tool. Primer for Clinicians. Francis J. Curry National Tuberculosis Center: San Francisco;
Our data suggest that chest ultrasonography may be a promising 2019. (2nd Edition).
[17] WHOeditor. WHO | definitions and reporting framework for tuberculosis. 2015.
tool to support clinical, radiological and microbiological data in the [18] Brunetti E, Heller T, Richter J, Kaminstein D, Youkee D, Giordani MT, et al.
diagnosis of PTB, a high burden pathological condition for which the Application of ultrasonography in the diagnosis of infectious diseases in resource-
delay in diagnosis represents a critical point in the control of the dis- limited settings. Curr Infect Dis Rep 2016;18:6https://doi.org/10.1007/s11908-
015-0512-7.
ease. [19] Heller T, Mtemang'ombe EA, Huson MAM, Heuvelings CC, Bélard S, Janssen S, et al.
Future studies are warranted to assess a possible role of LUS as a Ultrasound for patients in a high HIV/tuberculosis prevalence setting: a needs as-
triage/screening test in high burden countries. Furthermore the use of sessment and review of focused applications for Sub-Saharan Africa. Int J Infect Dis
2017;56:229–36. https://doi.org/10.1016/j.ijid.2016.11.001.
LUS as a follow-up examination to monitor therapy efficacy may be
[20] AIDE-MEMOIRE for Diagnostic Imaging Services 1999.
another attractive application to be investigated. [21] Uplekar M, Weil D, Lonnroth K, Jaramillo E, Lienhardt C, Dias HM, et al. WHO's
new end TB strategy. Lancet 2015;385:1799–801. https://doi.org/10.1016/S0140-
Funding 6736(15)60570-0.
[22] Hunter L, Bélard S, Janssen S, van Hoving DJ, Heller T. Miliary tuberculosis: so-
nographic pattern in chest ultrasound. Infection 2016;44:243–6. https://doi.org/
This research did not receive any specific grant from funding 10.1007/s15010-015-0865-8.
agencies in the public, commercial, or not-for-profit sectors. [23] Heuvelings CC, Bélard S, Andronikou S, Jamieson-Luff N, Grobusch MP, Zar HJ.
Chest ultrasound findings in children with suspected pulmonary tuberculosis.
Pediatr Pulmonol 2019. https://doi.org/10.1002/ppul.24230.
Declaration of Competing Interest [24] Yuan A, Yang P-C, Chang D-B, Yu C-J, Lee Y-C, Kuo S-H, et al. Ultrasound-guided
aspiration biopsy of small peripheral pulmonary nodules. Chest 1992;101:926–30.
https://doi.org/10.1378/chest.101.4.926.
The authors have no conflict of interest to disclose. On behalf of all
authors, the corresponding author states that there is no conflict of

34