a b s t r a c t

Objective: The differential diagnoses of patients presenting with chest pain (CP) and shortness of breath (SOB) are broad and non-specific. We aimed to 1) determine how use of Point-of-care ultrasound impacted emer- gency physicians’ differential diagnosis, and 2) evaluate the accuracy of POCUS when compared to chest radio- graph (CXR) and composite final diagnosis.

Methods: We conducted a prospective observational study in a convenience sample of patients presenting with CP and SOB to the Emergency Department (ED). Treating physicians selected possible diagnoses from a pre- indexed list of possible diagnoses of causes of CP and SOB. The final composite diagnosis from a chart review was determined as the reference standard for the diagnosis. The primary analysis involved calculations of sensi- tivity and specificity for POCUS identifiable diagnoses in detecting cause of CP and SOB. Additional comparative accuracy analysis with CXRs were conducted.

Results: 128 patients with a mean age of 64 +- 17 years were included in the study. Using a reference standard of composite final diagnoses, POCUS had equal or higher specificity to CXR for all indications for which it was used, except for pneumonia. POCUS correctly identified all patients with pneumothorax, pleural effusion and pericar- dial effusion. In patients with a normal thoracic ultrasound, CXR never provided any actionable clinical informa- tion. Adding POCUS to the initial evaluation causes a significant narrowing of the differential diagnoses in which the median differential diagnosis from 5 (IQR 3-6) to 3 (IQR 2-4) p b 0.001.

Conclusion: In evaluation of patients with CP and SOB, POCUS is a highly feasible diagnostic test which can assist in narrowing down the differential diagnoses. In patients with a normal thoracic ultrasound, the added value of a CXR may be minimal.

(C) 2018

Introduction

Chest pain (CP) and shortness of breath (SOB) are among the most common chief complaints in the Emergency Department (ED) [1]. The initial management can be challenging due to the broad differential di- agnoses, which often includes life-threatening conditions requiring rapid identification and management [2]. With many possible

* Corresponding author at: Center for Ultrasound Research and Education, 326 Cambridge Street, Suite 410, Boston, MA, 02114 USA.

E-mail address: [email protected] (H. Shokoohi).

etiologies, understanding the cause of CP and SOB is essential for targeted interventions, and a timely management of patients. Typically, a chest radiograph (CXR) is part of the initial work up for these patients. However, previous studies have challenged the diagnostic utility of CXR for these indications, suggesting that in the evaluation of CP, CXRs yield clinically significant information in as few as 12% of studies [3-5]. There- fore, the limited sensitivity and specificity of CXR in the evaluation of CP and SOB raises questions about the diagnostic accuracy and initial diag- nostic tests in patients with CP and SOB.

The use of point-of-care (POCUS) in the evaluation of CP and SOB is growing rapidly and in different clinical settings it shows promise as ac- curate point-of-care diagnosis is made possible [4-6]. Results of studies have shown that POCUS has excellent diagnostic accuracy for the

https://doi.org/10.1016/j.ajem.2018.10.059

0735-6757/(C) 2018

pathologies most commonly encountered in patients presented with CP and SOB, and often higher than CXR [6-10]. POCUS most often diagnoses pneumothorax, pleural effusion, lung consolidation, pulmonary edema, and pulmonary embolism [11-16]. For example, with respect to pleural effusion, Unluer et al. have reported a sensitivity and specificity of bed- side ultrasound as 93% and 93%, using CT as the gold standard [1]. With respect to pulmonary edema, Martindale et al. found 74% agreement with CT findings for lung ultrasound (versus 58% with CXR) [8]. In crit- ical care cases, Bedside lung ultrasound has been reported to yield a di- agnosis for patients with acute respiratory failure in 90.5% percent of cases [1,6,7,17].

Despite being a rapidly evolving initial Diagnostic modality, further studies are needed to determine whether POCUS actually increases the proportion of patients with undifferentiated CP and SOB who are correctly diagnosed and treated in the ED. Previous studies of the utility and efficacy of POCUS have largely focused on specific patient popula- tions including ventilated patients, trauma patients and children, who likely have significant difference in incidence of disease, and concomi- tant pathology [8,18,19].

The objective of this study was to 1) determine how use of POCUS in- fluenced emergency physicians’ differential diagnosis, and 2) compare the US findings to chest radiograph and composite final diagnosis to as- sess relative sensitivities and specificities of each imaging modality in this patient population. Our hypothesis was that the use of POCUS would narrow the overall differential diagnosis of SOB and CP, and in- crease the proportion of patients who receive a correct presumptive di- agnosis in the ED. We also investigated the hypothesis that POCUS would exhibit a comparable diagnostic accuracy compared with CXR in respect to pneumothorax, pleural effusion, pneumonia, and pulmo- nary edema.

Methods

Inclusion and exclusion criteria

This was a single-center prospective observational study conducted in an ED at Massachusetts General Hospital (MGH), a tertiary, academic hospital. The ED has about 110,000 visits per year, and 35,000 of these result in admission to the hospital. Admitted patients with CP and SOB are generally admitted to the medical or cardiac wards. Exceptions are unstable patients with acute ST elevation MI and those with a complete heart block who are referred directly to the cardiology catheterization labs.

In the ED, initial assessment and imaging testing are typically ordered based on symptoms and signs. Thus, symptoms of CP and SOB were cho- sen as inclusion criteria rather than a set of specific presumptive diagno- ses. Patients were included if they were over the age of 18, had a chief complaint of CP and/or SOB, and had a CXR ordered by the treating emer- gency physician (EP) for clinical reasons. Patients were excluded if they were referred from an outside hospital with a known diagnosis, if the EP was aware of the results of any diagnostic imaging prior to data collec- tion, if the patient was known or found to be pregnant during the course of care, if the patient was unavailable to complete any portion of the ul- trasound exam, or if the patient was unable to provide informed consent for any reason, including clinical instability.

This study was conducted in accordance with the Helsinki Declara- tion and approved by the hospital ethics committee and Internal Review Board. Informed consent was obtained from each patient prior to the study enrollments.

Data collection

Study investigators were physicians-sonographers (PS) with speci- fied ultrasound Fellowship training. If a patient was found to meet eligi- bility criteria, the PS asked the EP to select possible diagnoses from a list of common cardiopulmonary conditions. The primary diagnosis was

defined as the most likely diagnosis as determined by the treating phy- sician. After the primary clinical assessment, treatment and further di- agnostic tests were ordered by the EP based on clinical needs.

The complete list of differential diagnoses included: 1) asthma/ chronic obstructive pulmonary disease, 2) acute coronary syndrome,

3) pulmonary embolism, 4) pericarditis, 5) aortic dissection, 6) malig- nancy, 7) musculoskeletal-related chest pain, 8) upper respiratory ill- ness, and 9) other category which includes diagnoses such as non- specific chest pain, gastro-intestinal etiologies, and anxiety. Of the dif- ferential diagnoses, 5 are validated sonographic diagnoses: pneumotho- rax, pneumonia, pleural effusion, pulmonary edema, and pericardial effusion.

A focused ultrasound examination of the heart and lungs was then performed by the PS who was blinded to any additional imaging ob- tained at that encounter (CXR, CT, MRI), and unblinded to other relevant clinical information. The PS disclosed their final imaging findings to the EP. The EP was then re-surveyed on the differential diagnoses after re- ceiving the US data. This design was implemented to simulate the real clinical scenario at our institution, in which EPs frequently perform their own POCUS, in order to stratify their differential diagnoses and fur- ther imaging studies.

Additional clinical data was obtained retrospectively including pa- tient demographics and final composite diagnosis, and recorded by the PS. As in regular clinical encounters, the EP had access to all available patient data including all available imaging, in order to inform their clin- ical decisions.

A sample size of 128 was targeted in order to achieve 80% power to detect a difference of 0.5 in mean number of differentials, assuming a standard deviation of 2 differential diagnoses, and using a paired t-test (with a 0.05 two-sided significance).

Ultrasound protocol

Emergency ultrasound fellows or attendings with ultrasound fellow- ship training performed and interpreted scans. At all times, participating patients received the Standard care.

Studies included focused cardiac and thoracic examinations. The fo- cused cardiac examination was conducted using a phased array probe and included a protocol to obtain at least 2 of the 4 basic cardiac views (subxiphoid view, parasternal long view, parasternal short view and apical four chamber view). This scanning protocol is consistent with previously published assessments of SOB or undifferentiated respiratory failure, including the Rapid Assessment of Dyspnea with Ultrasound: RADiUS, which incorporates 2 cardiac views and 6 lung views [20]. Other similar protocols include SEARCH 8E’s which assesses: empty thorax, edematous lung, extended focused assessment with sonography for trauma (E-FAST), effusion, equality (the ratio between left and right ventricle), ejection fraction, exit (aorta) and entrance (inferior vena cava [IVC]) and endocardial movement [21,22]. Cardiac images were interpreted for presence or absence of: 1) pericardial effusion, 2) de- creased LV systolic function, and 3) RV strain.

The focused thoracic examination was conducted using a curviLinear probe and included six thoracic zones: anterior (bilaterally between sternum and anterior axillary line), lateral (bilaterally between anterior and posterior axillary line), and posterior (bilaterally between the pos- terior axillary line and spine). Pulmonary images were interpreted by the study investigators for the presence of Lung sliding, pleural effusion, consolidation, and B-lines.

We defined a POCUS diagnosis of pneumonia as the presence of a lobar consolidation or multiple patchy consolidation with hyperechoic artifacts consistent with air bronchograms. In a majority of the cases a disruption of pleural line extending inferiorly with air bronchogram, and irregular consolidations (shred sign), with or without pleural effu- sion were defined as pneumonia. Isolated interstitial consolidation without additional artifacts were not considered as a pneumonia. Pul- monary edema was defined as the presence of diffuse B-lines (>=3) in

ultrasound findings“>both lung fields, with or without pleural effusion. The lack of pleural sliding as compared with the opposite side was considered as pneumothorax.

We look for the presence of pleural effusion from right and left upper quadrant views and confirmed with the presence of spine sign.

Statistical analyses

The differential diagnoses before and after POCUS were compared using Wilcoxon signed-rank tests. Sonographic diagnosis and CXR diag- nosis were compared using the McNemar test. Kappa scores were assessed to measure interrater reliability of sonographic diagnoses (bedside ultra-sonographer versus expert reviewer). A two-sided p b

0.05 was considered to indicate statistical significance. All statistical analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC).

Results

128 patients presenting with chest pain and/or shortness of breath were included in the study. The mean age was 64 +- 17 years. 55% were male (71/128) and 45% were female (57/128). 34% (44/128) pre- sented with chest pain, 37% (47/128) with shortness of breath, and 29% (37/128) with both CP and SOB. Baseline characteristics of the patients are summarized in the Table 1.

Of these patients, 27% (35/129) were ultimately diagnosed with con- ditions that can be accurately assessed via rapid POCUS (pneumothorax, pericardial effusion, pneumonia, pulmonary edema, and pericardial effusion).

Ultrasound findings

The proportions of patients with specific ultrasound findings are provided in Table 2. In these patients the most common positive POCUS findings were diffuse B-lines and Pleural effusions in lung ultra- sound and a decreased ejection fraction and RV dilation. (Table 2). POCUS took an average of 10.0 min to be completed.

Of the 128 lung Ultrasound studies, 43% (55/128) were normal, and 57% (73/128) were abnormal for pathology. Of the 55 normal lung

Table 1

Patient demographics.

Patient characteristics (n = 128) Number (%) Mean age 64

Female 57 (45)

Presenting complaint

Chest pain only 44 (34)

Shortness of breath only 47 (37)

Both chest pain and shortness of breath 37 (29) Comorbidities

Coronary artery disease 44 (34)

Chronic obstructive lung disease 33 (26)

Asthma 17 (13)

Congestive heart failure 30 (23)

Malignancy 22 (17)

Interstitial lung disease/pulmonary fibrosis 14 (11) Most common final (composite) diagnosis

Asthma/COPD 20

Pulmonary edema 19

Pneumonia 14

ACS 14

Pericarditis 5

URI 3

Malignancy 2

Pneumothorax 1

Pleural effusion 1

Atypical CP and Other Dx 68

Table 2

cardiac and lung ultrasound findings (n = 128).

Present Absent Indeterminate

Abnormal cardiac ultrasound	46 (36%)
Decreased ejection fraction	36	92	0
Right ventricular dilation	10	107	11
Pericardial effusion	8	119	1
Pericardial effusion (significant)	2	125	1
Enlarged aortic root	6	99	23
Abnormal lung ultrasound	73 (57%)
Absence of lung sliding	3	125	0
right pleural effusion	27	96	5
Left pleural effusion	31	93	4
Diffuse B-lines	57	71	0
Irregular pleural line/sub pleural	60	68	0
Hepatization/air bronchogram	21	107	0

ultrasound studies, 87% (48/55) were associated with a normal chest ra- diograph. The abnormal CXRs associated with normal US were positive for: 5 with possible atelectasis, 1 with known speculated lesion, 1 with mild pulmonary edema that needed no treatment. (Fig. 1).

The inclusion of heart and lung ultrasound to the conventional diag- nostic evaluation of patients with CP and SOB significantly narrowed down the list of differential diagnoses in their initial ED evaluations. EPs were presented with a list of 14 potential diagnoses from which 5 diagnoses were directly assessed using ultrasound. The diagnoses ame- nable to US diagnosis were included: 1) pneumothorax, 2) pleural effu- sion, 3) pneumonia, 4) pulmonary edema, and 5) pericardial effusion. On average, the EP initially considered 5 diagnoses in their initial differ- ential based upon history and physical exam and of these, an average of 2 (IQR 1-4) of the 5 diagnoses were conducive to sonographic evalua- tion. After the performance of ultrasound, the median number of differ- ential diagnoses entertained decreased from 5 (IQR 3-6) to 3 (IQR 2-4), p b 0.001 (Table 3). Following POCUS evaluation there was a statistically significant decrease in all sonographically identifiable diagnoses. These results are consistent with previous studies in which point-of-care ul- trasound was found to be a feasible and accurate diagnostic tool in pa- tients presenting with dyspnea [19-21].

Interestingly, the differential of non-sonographic diagnoses also de- creased following ultrasound. This may be because when a compelling positive finding was considered the likely sole cause of symptoms, other diagnoses became much less likely (Occam’s razor) (Table 3).

Does CXR add value if the thoracic ultrasound is negative?

We found variable agreement between the CXR and ultrasound find- ings, with kappa scores ranging from 1.0 (100% agreement) for pneu- mothorax, to 0.40 (40% agreement) for pneumonia. When thoracic ultrasound was negative, the CXR was also negative 87% of the time. In only 13% of cases (7/55) did CXR find something that ultrasound did not. 5/7 discrepant positive x-rays revealed possible atelectasis on chest radiographs. Of these, 2/7 were diagnosed with minor exacerba- tion of COPD and were discharged home, 2/7 required additional work-up of chest pain to rule out acute coronary syndrome, and 1/7 pa- tients was felt to have a post-viral cough, observed overnight without evidence of desaturation and discharged the following day. For the re- maining 2 discrepant CXRs, 1/7 revealed mild pulmonary edema in a pa- tient with chest pain. He was admitted for blood pressure management. No diuretics were administered. 1/7 revealed a known speculated lung lesion, information that was already known to the provider team. This patient was ultimately admitted for acute exacerbation of chronic ob- structive pulmonary disease (COPD), a condition that could not be ruled in or out by POCUS. In no patients with a normal lung ultrasound did the chest x-ray provide any new or actionable information. This sug- gests that a normal lung ultrasound is highly predictive of having a

Fig. 1. Lung and chest radiograph results of 128 patients with chest pain and/or shortness of breath.

negative chest radiograph, or a positive CXR not requiring further inter- vention in the emergency context. POCUS may provide a higher propor- tion of actionable data, or in other words higher specificity without a clinically significant decrease in sensitivity. Given that ultrasound can be done faster than CXR at a lower cost and with no radiation, an ultrasound-first approach for evaluation of patients with chest pain or shortness of breath with a follow-up CXR only in patients with any pos- itive findings on lung ultrasound may prove to be a reasonable approach.

Table 3

Effects of point-of-care ultrasound on the differential diagnoses for CP and SOB.

Relative sensitivities of CXR and POCUS

Comparison of the relative sensitivities and specificities for the two tests revealed overlapping confidence intervals for all studies, except in the study of pneumonia. The sensitivity and specificity of CXR and POCUS were: 38% (95% CI 13-70%) and 96% (95% CI 90-99%) versus

89% (95% CI 54-100) and 74% (95% CI 64-82%) respectively (Table 4). Although the specificity of CXR in evaluation of pneumonia was higher than for POCUS (statistically significant) the sensitivity of POCUS was higher (clinically significant) (CXR 38% versus POCUS 89%). A driver of this discrepancy may be the high proportion of indeterminant studies for pneumonia (16 CXRs versus 30 POCUS). The inability to assess for pathology on indeterminate studies likely decreases the actual sensitiv- ity of these exams in practice, though not the NPV. However, clearly de- fining indeterminate scans (with pathophysiologic rationale) may

	(%)		(%)
Acute coronary syndrome		78		78	1.0
Pneumonia		77		49	b0.0001
Pulmonary edema		71		29	b0.0001
Musculoskeletal pain		51		52	0.85
Pericardial effusion		51		4	b0.0001
Pleural effusion		50		22	b0.0001
Pulmonary embolism		45		46	0.85
Pneumothorax		43		1	b0.0001
Asthma/COPD		39		28	0.03
Pericarditis		36		35	0.84

Presumptive Differential Diagnoses

Pre-POCUS

Differential Diagnoses

Post-POCUS Differential Diagnoses

p-Value

prove an essential component of ultrasound interpretation going for- ward, in order to maintain and improve the sensitivity of these studies. In practice, it is advantageous for POCUS, as a rapid bedside exam, to have a higher NPV, thus helping clinicians to safely avoid unnecessary further diagnostic work up. Therefore, as in this study, we argue that conservative definitions should be favored going forward, acknowledg- ing the clinical use of ultrasound as a “rule-out” test, with the potential to eliminate the need for CXR.

Upper respiratory tract infection

28 28 1.0

Limitations

In our analysis, CXR and POCUS diagnoses were compared to the composite final discharge diagnosis, as selected by the treating physi-

Aortic dissection 12 12 1.0

Malignancy 11 11 1.0

Other 9 9 1.0

Total number of ddx 5 (IQR 3-6) 3 (IQR 2-4) p b

0.001

IQR = interquartile range.

cian. Use of the composite final diagnosis had certain advantages, such as incorporating other imaging findings and clinical information. How- ever, in most cases physicians selected a single leading diagnosis at the end of the encounter as the primary etiology of their symptoms, without necessarily coding incidental findings. It is possible that in a

Table 4

Relative sensitivity and specificity of CXR and ultrasound with respect to final composite diagnoses.

	CXR				Ultrasound
	Sensitivity	Specificity	Indeterminate		Sensitivity	Specificity	Indeterminate
Pneumothorax	1.0	1.0	0		1.0	1.0	0
	(0.17-1.00)	(0.96-1.00)	(0)		(0.17-1.00)	(0.96-1.00)	(0)
Pneumonia	0.38	0.96	16		0.89	0.74	30
	(0.13-0.70)	(0.90-0.99)	(13%)		(0.54-1.00)	(0.64-0.82)	(23%)
Pleural effusion	1.0	0.80	0		1.0	0.71	2
	(0.17-1.00)	(0.73-0.86)	(0)		(0.17-1.00)	(0.63-0.78)	(2%)
Pulmonary edema	0.81	0.92	6		0.78	0.93	14
	(0.56-0.94)	(0.86-0.97)	(5%)		(0.52-0.93)	(0.86-0.97)	(11%)

Note: There were zero instances in which pericardial effusion or atelectasis were the final composite discharge diagnosis.

proportion of the false positive studies, patients in fact had a component of either atelectasis, pneumonia, pleural effusion, CHF or pericardial ef- fusion that was not captured in a parsimonious discharge diagnosis. In other words, we suspect that the specificities of both CXR and POCUS were underestimated using this method of analysis. As an example, al- though 8 POCUSs were positive for pericardial effusion, pericardial effu- sion was never coded as the principal diagnosis.

Another consideration is that a significant proportion of CXRs and POCUS were indeterminate (technically limited studies). CXRs were interpreted by radiologists within the context of regular clinical use. Therefore, this number is likely reflective of actual practice, in which a proportion of CXRs do not provide information which can be used to guide clinical decision making. A limitation of our study is that the POCUS exams were performed by sonographers who had special ultra- sound training, rather than the treating emergency physician. There- fore, we cannot generalize the findings among emergency physicians without special ultrasound training. Therefore, it is possible that our rate of indeterminate scans underestimates the rate that would be ascertained in regular clinical practice by physicians with less training in POCUS.

Alternatively, another limitation to the study is that the sonographer was not directly involved in the patient’s care, which could have limited the role of the POCUS in the physician’s differential diagnosis. The main advantage of point of care ultrasound is that the clinician performs POCUS in the context of the clinical background and makes informed Clinical decisions based on US interpretation. Therefore, had the ultraso- nographer been the EP, it is possible that there may have been a lower rate of indeterminate scans. Finally, as sonographic interpretation con- tinues to evolve, there remains variability in the way in which several pathologies are defined.

Changing the definition of pneumonia, for example to include scans

that were positive for B-lines only, would significantly change our rela- tive sensitivities and specificities.

Another consideration is that both CXR and POCUS findings were compared to the composite diagnosis as the gold standard. This had the advantage of incorporating all of the clinical information, including Clinical impression, blood work, all imaging, and other studies. How- ever, previous studies have largely used CT as a gold standard. Because not all patients presenting to the ED with CP or SOB required a CT, this information was used to guide the composite diagnosis in some but not all cases. In the absence of CT, either the CXR or POCUS may have in- troduced clinical bias on the composite diagnosis that is absent from the CT study which is conventionally interpreted by the radiologist, acting as a relatively independent third party.

Conclusion

In patients presenting to the ED with chest pain or SOB, use of ultra- sound is a highly feasible diagnostic test which can narrow down the differential diagnosis. In initial evaluation of patients with CP and SOB, CXR and POCUS had comparable accuracy in diagnosing pneumothorax, pleural effusion, and pulmonary edema. In patients with a normal

thoracic ultrasound, CXR never provided any actionable clinical infor- mation. Considering the benefits of POCUS in reducing cost of care, the speed of care delivery, and as a radiation free test, these data may sug- gest that the incorporation of ultrasound into the initial evaluation of ED patients with chest pain or shortness of breath may reduce the need for CXR in some cases and provide complementary information in others.

Financial support

None.

Conflicts of interest

No conflicts of interest.

References

1. Unluer EE, Karagoz A. Bedside lung ultrasound versus chest X-ray use in the emer- gency department. Interv Med Appl Sci 2014 Dec;6(4):175-7.
2. Turner JP, Dankoff J. Thoracic ultrasound. Emerg Med Clin North Am 2012 May;30 (2):451-73.
3. Goldschlager R, Roth H, Solomon J, Robson S, et al. Validation of a clinical decision rule: chest X-ray in patients with chest pain and possible acute coronary syndrome. Emerg Radiol 2014 Aug;21(4):367-72.
4. Zanobetti M, Poggioni C, Pini R. Can chest ultrasonography replace standard chest radiography for evaluation of acute dyspnea in the ED? Chest 2011 May;139(5): 1140-7.
5. Harel-Sterling M, Diallo M, Santhirakumaran S, Maxim T, Tessaro M. Emergency de- partment resource use in pediatric pneumonia: point-of-care lung ultrasonography versus chest radiography. J Ultrasound Med 2018 Jul;37(7):1687-96.
6. Zanobetti M, Scorpiniti M, Gigli C, Nazerian P, Vanni S, Innocenti F, et al. Point-of- care ultrasonography for evaluation of acute dyspnea in the ED. Chest 2017 Jun; 151(6):1295-301.
7. Huang D, Ma H, Xiao Z, Blaivas M, et al. Diagnostic value of cardiopulmonary ultra- sound in elderly patients with acute respiratory distress syndrome. BMC Pulm Med 2018 Aug 13;18(1):136.
8. Martindale JL, Noble VE, Liteplo A. Diagnosing pulmonary edema: lung ultrasound versus chest radiography. Eur J Emerg Med 2013 Oct;20(5):356-60.
9. Rempell JS, Noble VE. Using lung ultrasound to differentiate patients in acute dys- pnea in the prehospital emergency setting. Crit Care 2011;15(3):161.
10. Kajimoto K, Madeen K, Nakayama T, Tsudo H, et al. Rapid evaluation by lung- cardiac-inferior vena cava (LCI) integrated ultrasound for differentiating heart fail- ure from pulmonary disease as the cause of acute dyspnea in the emergency setting. Cardiovasc Ultrasound 2012 Dec 4;10(1):49.
11. Volpicelli G, Elbarbary M, Blaivas M, Lichtenstein DA, et al. International evidence- based recommendations for point-of-care lung ultrasound. Intensive Care Med 2012 Apr;38(4):577-91.
12. Laursen CB, Sloth E, Lambrechtsen J, et al. Focused sonography of the heart, lungs, and deep veins identifies missed life-threatening conditions in admitted patients with acute respiratory symptoms. Chest 2013;144:1868-75.
13. Reissig A, Copetti R, Mathis G, et al. Lung ultrasound in the diagnosis and follow-up of community-acquired pneumonia: a prospective, multicenter, diagnostic accuracy study. Chest 2012;142:965-72.
14. Patel CJ, Bhatt HB, Parikh SN, Jhaveri BN, Puranik JH. Bedside lung ultrasound in emergency protocol as a diagnostic tool in patients of acute respiratory distress pre- senting to emergency department. J Emerg Trauma Shock 2018 Apr-Jun;11(2): 125-9.
15. Orso D, Ban A, Guglielmo N. Lung ultrasound in diagnosing pneumonia in childhood: a systematic review and meta-analysis. J Ultrasound 2018 Jun;21.
16. Balk DS, Lee C, Schafer J, Welwarth J, Hardin J, Novack V, et al. Lung ultrasound com- pared to chest X-ray for diagnosis of pediatric pneumonia: a meta-analysis. Pediatr Pulmonol 2018 Aug;53(8):1130-9.
17. Lichtenstein DA, Meziere GA. Relevance of lung ultrasound in the diagnosis of acute respiratory failure: the BLUE protocol. Chest 2008 Jul;134(1):117-25.
18. Acosta CM, Maidana GA, Jacovitti D, Belaunzaran A, et al. Accuracy of transthoracic lung ultrasound for diagnosing anesthesia-induced atelectasis in children. Anesthe- siology 2014 Jun;120(6):1370-9.
19. Laursen CB, Sloth E, Lassen AT, Christensen RD, et al. Point-of-care ultrasonography in patients admitted with respiratory symptoms: a single-blind, randomised con- trolled trial. Lancet Respir Med 2014 Aug;2(8):638-46.
20. Manson W, Hafez NM. The rapid assessment of dyspnea with ultrasound: RADiUS. Ultrasound Clin 2011;6(2):261-76.
21. Ahn JH, Jeon J, Toh HC, Noble VE, et al. SEARCH 8Es: a novel point of care ultrasound protocol for patients with chest pain, dyspnea or symptomatic hypotension in the emergency department. PLoS One 2017 Mar 29;12(3):e0174581.
22. Yang JX, Zhang M, Liu ZH, Ba L, et al. Detection of lung atelectasis/consolidation by ultrasound in Multiple trauma patients with mechanical ventilation. Crit Care 2009;1(1):13-6.