a b s t r a c t

Background: Chest radiography is the initial choice for thoracic imaging. However, the wide availability of com- puted tomography (CT) has led to a substantial increase in its use in the emergency department (ED). We eval- uated the utility of chest CT after a chest X-ray in patients presenting to the ED with non-traumatic thoracic emergencies, and determined if the diagnosis and management decision changed after CT.

Methods: The study enrolled 500 consecutive patients with both chest X-rays and CT who presented to the ED with non-traumatic complaints. Chest X-rays and CT images obtained within 12 h before any definitive treatment were randomly evaluated in consensus by two radiologists blinded to the clinical information.

Results: The chest X-ray and CT image findings were concordant in 49.2% of the 500 patients and this concordance was negatively correlated with patient age. Leading diagnosis and management decisions based on the chest ra- diograph changed after CT in 35.4% of the study group and this finding was also correlated with age. In 55% of 205 patients, pneumonic infiltrations were undiagnosed with radiography. Pulmonary edema was the most specific (93.3%) and sensitive (85.4%) radiography finding. Posteroanterior chest radiographs taken in the upright posi- tion had higher concordance with CT than anteroposterior (AP) radiographs taken in the supine position.

Conclusions: Chest CT may be an appropriate imaging choice in patients presenting to the ED for non-traumatic reasons, particularly for elderly patients and when the radiograph is taken with the AP technique in a supine position.

(C) 2017

Introduction

Chest radiography is the initial choice for thoracic imaging because it offers simplicity, a large amount of information, and lower costs [1]. thoracic ultrasonography has been shown to be an imaging modality comparable with portable chest X-rays since it can be used bedside, has high accuracy and no exposure to radiation as well as minimal low cost. Nevertheless, the usefulness of computed tomography (CT) in thoracic radiology is recognized. The remarkable development and wide availability of CT have led to a substantial increase in its use in the emergency department (ED) [2]. Recent studies have reported that CT is more useful than chest radiography for detecting traumatic pathologies, while the clinical utility of chest CT for non-traumatic emergencies has not been described [2,3].

CT also has disadvantages, such as the relatively higher exposure to ionizing radiation, higher costs, and risk of nephrotoxicity when

* Corresponding author at: Department of Radiology, Faculty of Medicine, Dokuz Eylul University Hospital, Mithatpasa Cad., 35340 Inciralti-Izmir, Turkey.

E-mail address: [email protected] (N.S. Gezer).

contrast medium is used [2]. Furthermore, a CT examination in ED con- ditions is time-consuming and may slow ED throughput [2]. Therefore, the indications for chest CT should be reevaluated [4]. Medical knowl- edge, routine workload, health policies, and patient expectations all play roles in the physician’s selection of radiological examinations. However, the fundamental role of radiological and laboratory examina- tions is to contribute to the diagnosis and patient management [5].

This study evaluated the utility of chest CT after a chest radiograph in patients presenting to the ED with non-traumatic thoracic emergencies and determined if the diagnosis and management decisions changed after the CT examination. These data will better help clinicians decide whether to order chest CT in the ED for non-trauma patients.

Methods

This retrospective study was approved by the ethics committee of our institution. The study enrolled 500 consecutive patients (225 fe- males, 275 males; mean age 70 [range 18 -97] years) who presented to the ED with non-traumatic thoracic emergencies and had both a chest X-ray and subsequent chest CT within 12 h. The patients whose

http://dx.doi.org/10.1016/j.ajem.2016.12.058

0735-6757/(C) 2017

CT examination were obtained >= 12 h after X-ray were not included in the study for the reason that new problems might have been added after this period of time. The patients who were younger than 18 years or who received definitive treatment before the chest CT were not included in the study since the treatment would influence im- aging findings. Chest X-rays were taken with the posteroanterior (PA) technique in the upright position or with the anteroposterior (AP) tech- nique in a supine position using a portable machine. CT imaging was performed using a multi-detector CT scanner (Brilliance 64 Philips; Philips Medical Systems(C), Eindhoven, The Netherlands). The routine scanning parameters were voltage 120 kVp, current 120 mA, and slice thickness of 2 mm. CT was performed without intravenous or oral contrast.

The chest radiographs and CT images were randomly evaluated in consensus by two radiologists blinded to the clinical information. Chronic findings, such as chronic bronchitis or emphysema, were not considered. Emergency pathology findings including pulmonary edema, pneumonic infiltration, pleural effusion, atelectasis, pneumo- thorax, and fractures were noted as imaging findings in each modality separately. The radiologists also recorded nonspecific suspicious radio- graphic findings such as hilar or mediastinal widening that required fur- ther investigation. The concordance of the detection of pathology by chest radiography and CT imaging were assessed, after which miscon- ceptions were determined. For each imaging finding, we compared the sensitivity and specificity of radiography with those for CT. After the radiological evaluation, each patient’s hospital records were investi- gated for clinical details to assess the impact of chest CT (after chest X- ray) on the patient’s diagnosis and management in the ED. Patients whose hospital records were missing or incomplete were excluded from the study.

Mediastinal enlargement, Pulmonary nodules (b 3 cm), and masses (N 3 cm) detected by radiography were recorded because they required further investigation with CT. The concordance of the detection of medi- astinal masses, pulmonary nodules, and masses by chest radiography and CT imaging was evaluated. The sensitivity and specificity of radiog- raphy for the detection of nodules and masses compared to CT were also analyzed.

The statistical analyses were performed using SPSS 20.0 for Win- dows (IBM, NY, USA). Continuous data are expressed as the mean +- standard deviation, whereas categorical data are presented as the num- ber of patients and percentages. Continuous variables were compared between groups using the independent samples t-test. Correlations be- tween different variables were examined using the Pearson correlation coefficient. Subgroup analysis also compared the PA and AP radiography groups for concordance with CT, impact on diagnosis/management, sensitivity, specificity, positive/negative predictive value, and positive/ negative likelihood ratio. A p value b 0.05 indicated statistical significance.

Results

Chest radiography was taken with the PA technique in 56% of the pa- tients and with the AP technique for the remaining 44% patients. The pa- tients in the PA group were significantly younger than those in the AP group (67.3 +- 16.9 vs. 75.2 +- 11.3 years, p b 0.001). According to the ra- diologists’ consensus decision, 208 (41.6%) of the radiographs were nor- mal or had chronic findings. Among the remaining 292 (58.4%) radiographs, the most frequently detected finding was pleural effusion (n = 133, 45.6%), followed by pulmonary edema (n = 120, 41.1%)

and pneumonic infiltration (n = 34, 33.9%). Of the 500 CT images, 129 (25.8%) were normal or had chronic findings. Among the remaining 371 (74.2%) CT images, the most frequent finding was pneumonic infil- tration (n = 205, 55.3%) followed by pleural effusion (n = 168, 45.3%) and pulmonary edema (n = 110, 29.7%). Table 1 summarizes the imag- ing findings.

Table 1

Characteristics of study population, imaging findings detected by radiography and CT

Age (years)	70 (+-15.2)
Gender Female	225 (45)
Male	275 (55)

Type of chest radiography

AP 222 (44)

PA 278 (56)

Total number of abnormal radiographs* 292 (58)

Chest radiography findings Pleural effusion	133 (45.6)
Pulmonary edema	120 (41.1)
Pneumonic infiltration	99 (33.9)
Pulmonary mass	25 (8.6)
Pneumothorax	5 (1.7)
Mediastinal widening	4 (1.4)
Multiple nodules	4 (1.4)
Solitary nodule	3 (1)
Atelectasis	3 (1)
Costal fracture	2 (0.7)
Hilar widening	2 (0.7)
Total number of abnormal CTs*	371 (74)
Chest CT findings Pneumonic infiltration	205 (55.3)
Pleural effusion	168 (45.3)
Pulmonary edema	110 (29.7)
Pulmonary mass	25 (6.7)
Solitary nodule	23 (6.2)
Multiple nodules	10 (2.7)
Pneumothorax	6 (1.6)
Costal fracture	6 (1.6)
Atelectasis	4 (1.1)
Mediastinal mass	1 (0.3)
Mediastinal lymphadenopathy	1 (0.3)
Pulmonary nodule(s) or mass(s) suggesting malignancy	33 (11.3)

Data are presented as mean +- SD or n (%); * Abnormal findings: Radiography or CT which is not evaluated as normal or which is presenting findings other than chronic and sequel Fibrotic changes, chronic obstructive pulmonary disease or emphysema. AP, anteroposterior; PA, posteroanterior; CT, computed tomography.

Mediastinal widening was also recorded in four patients, and may be a sign of a malignant pathology that needs further investigation with CT. In those cases, CT examination showed that one patient had an actual mediastinal mass, which was subsequently diagnosed as thymic carci- noma; one patient had a right perihilar central pulmonary mass; and one patient had multiple lymphadenopathy related to lymphoma while one of them was a false image and the mediastinum was normal. With CT imaging, pulmonary masses or nodules that were highly suspicious for malignancy and required further investigation were de- tected in 33 (6.6%) of the 500 patients. In comparison, radiography de- tected a mass or nodule in only eight (24%) of these patients. The radiographic findings were concordant in 246 (49.2%) of the 500 pa- tients and this concordance was negatively correlated with patient age (p = 0.025, r = -0.1). The leading diagnosis and management deci- sions according to chest radiography changed after CT in 177 (35.4%) patients of the study group and this finding was also correlated with age (p = 0.002, r = 0.138). With radiography, pneumonic infiltration was undiagnosed in 113 (55%) of 205 patients and pulmonary edema in 15 (14%) of 109 patients; in addition, 22 patients were misdiagnosed with pneumonic infiltration, while 5 patients were misdiagnosed with pulmonary edema. Radiography failed to diagnose pneumothorax in four of six patients. There were six patients with costal fractures and three of them had accompanying Pleural effusions compatible with he- mothorax, but only two of them were diagnosed with radiography, two were reported as normal, and two were only diagnosed with pleural

effusions.

Pulmonary edema was the most sensitive (85.4%) and specific (93.3%) finding of radiography. The sensitivity of radiography for de- tecting atelectasis, pneumothorax, pleural effusion, mass, pneumonic infiltration, and fractures was 75%, 66.7%, 60.1%, 60%, 37.7%, and 33.3%,

respectively, whereas the specificities for these were 100%, 99.8%, 90.4%, 97.9%, 92.8%, and 100%, respectively. The sensitivity of radiography for detecting solitary and multiple pulmonary nodules was 8.7% and 30%, respectively, while its specificity was 99.8% for both. Table 2 compares the sensitivity and specificity of radiography with CT for each imaging finding. Subgroup analysis showed that PA radiography had higher con- cordance (p = 0.01, r = 115) with CT, and CT had a higher impact on diagnosis/management when taken after AP radiography (p = 0.018, r = 106).

Discussion

In this study, the radiographic findings were concordant in 49.2% of the cases, while the leading diagnosis and management decisions changed after CT in 35.4% of the patients presenting to the ED with non-traumatic thoracic emergencies. We suggest that this situation could be specific to the ED, because 44% of the radiographs in our study had to be taken with the AP technique, which has lower radio- graphic quality, and this group was also significantly older. We also sug- gest that the negative correlation of age with our results was associated with the diagnostic challenge due to fibrotic changes, accompanying chronic pulmonary diseases, or cardiac failure, which frequently occur in elderly patients.

Pneumonic infiltration

The sensitivity and specificity of radiography for detecting pneu- monic infiltration was 37.7% and 92.8%, respectively. Our results suggest that radiography has significant limitations for detecting pneumonic in- filtrations in patients presenting to the ED, and CT is important for diag- nosis and management decisions. In contrast to our results, Syrjala et al.

[6] reported that the clinical significance of CT findings for patients with negative chest radiography was unclear. However, in support of our findings, Pandharipande et al. [7] more recently reported that the lead- ing diagnosis changed in 163 of 387 (42%) cases and admission deci- sions changed in 72 of 387 cases (19%) admitted to the ED with chest pain or dyspnea [7]. The Infectious Diseases Society of America/Ameri- can Thoracic Society Consensus guidelines on the management of com- munity-acquired pneumonia in adults require a chest radiograph for the routine evaluation of patients who are likely to have pneumonia [8]. They reported that a pneumonic infiltration rarely is absent in radio- graphs and CT might be more sensitive. However, the guidelines suggest treating such conditions presumptively with antibiotics and then re- peating the imaging in 24-48 h [8]. In comparison, we found that pneu- monic infiltrations were undiagnosed with radiography in more than half of the study group population (55%) and this is not a rare situation (Fig. 1).

According to our results, undiagnosed pneumonia is an important issue. The guidelines require a demonstrable infiltrate on a chest X-ray or another imaging technique with or without supporting microbiolog- ical data for the diagnosis of pneumonia, in addition to suggestive clin- ical features such as fever, cough, sputum production, and chest pain [8].

In comparison, the physical detection of rales or bronchial breath sounds is less sensitive and specific than chest radiographs [9]. Further- more, the clinical features and physical examination may be lacking or altered in elderly patients, and routine laboratory tests are frequently falsely negative and are often nonspecific [8]. Because we demonstrated that the concordance of radiography with CT is negatively correlated with age and the clinical features and physical examination findings are often absent in Elderly people, we believe that it is very important to prove pneumonia radiologically. A Prompt diagnosis of pneumonia with the use of CT in the ED would also prevent unnecessary tests and interventions from being performed. On the other hand, 22 patients di- agnosed with pneumonia using radiography did not actually have pneu- monia. Pneumonia has to be differentiated from non-pneumonia entities considering the potential side effects of empirical antibiotic treatment. Also, there are legitimate concerns about the overuse of an- tibiotics in primary care and the development of resistance. We suggest that, in addition to unnecessary drug use, treating these patients as if they have pneumonia will prevent clinicians from searching for the ac- tual infection focus.

Another important issue is the criteria for severe community-ac- quired pneumonia, which suggests the need for intensive care unit treatment. A multilobar infiltration is one of the minor criteria for sever- ity, indicating that not only the presence of a pneumonic infiltration but also its distribution is important (Fig. 2). Therefore, we believe that CT can play an important role in the management of pneumonia.

Pleural effusion

The accurate detection of a pleural effusion is important, because it is one of the clinical indications for more extensive diagnostic testing for pneumonia [8]. Furthermore, it is a complementary finding in the diagnosis of cardiac failure. In our series, however, the specificity of ra- diography for detecting a pleural effusion was high (90.4%), while its sensitivity was unsatisfactory (60.1%). We suggest that this diagnostic challenge was mostly due to the high frequency (44%) of use of the AP technique, which had to be used for immobile patients in the ED. In the supine position, fluid is dispersed evenly throughout the lung fields so that non-massive fluid levels could not be detected in AP radiographs.

Pulmonary edema

In our series, pulmonary edema was the most sensitive (85%) and specific (93%) finding detected by radiography. Remember, however, that pulmonary edema may mask a pneumonic infiltration on radiogra- phy, so CT may be helpful when the clinical findings support pneumonia.

Mass and pulmonary nodules

Radiography identified only 8 of 33 (24%) masses or nodules that were highly suspicious for malignancy and required further investigation.

Table 2

Predictive values of chest radiography findings for the diagnosis of non-traumatic emergencies

	Sensitivity % (95% CI)	Specificity % (95% CI)	PPV % (95% CI)	NPV % (95% CI)	LR+ (95% CI)	LR- (95% CI)
Pulmonary edema	85.4 (77.5-91.4)	93.3 (90.4-95.6)	78.3 (69.9-85.3)	95.8 (93.2-97.6)	12.8 (8.8-18.7)	0.16 (0.1-0.25)
Atelectasis	75 (19.4-99.4)	100 (99.3-100)	100 (29.2-100)	99.8 (98.9-100)	NAa	0.25 (0.05-1.4)
Pneumothorax	66.7 (22.3-95.7)	99.8 (98.9-100)	80 (28.4-99.5)	99.6 (98.5-99.9)	329.3 (42.9-2528)	0.3 (0.1-1.04)
Pleural effusion	60.1 (52.3-67.6)	90.4 (86.7-93.3)	75.9 (67.8-82.9)	81.7 (77.4-85.6)	6.2 (4.4-8.9)	0.4 (0.37-0.5)
Pulmonary mass	60 (38.7-78.8)	97.9 (96.2-99)	60 (38.7-78.8)	97.9 (96.2-99)	28.5 (14.3-56.9)	0.4 (0.2-0.7)
Pneumonic infiltration	37.7 (31.1-44.7)	92.8 (89.2-95.5)	78.8 (69.4-86.4)	67.8 (63-72.4)	5.3 (3.4-8.3)	0.7 (0.6-0.75)
Costal fracture	33.3 (4.3-77.7)	100 (99.3-100)	100 (15.8-100)	99.2 (98-99.8)	NAa	0.7 (0.4-1.2)
Multiple nodules	30 (6.7-65.2)	99.8 (98.8-100)	75 (19.4-99.4)	98.6 (97.1-99.4)	147 (16.7-1293.8)	0.7 (0.5-1.05)
Solitary nodule	8.7 (1.1-28)	99.8 (98.8-100)	66.7 (9.4-99.2)	95.8 (93.6-97.4)	41.5 (3.9-440.9)	0.9 (0.8-1.04)
a NA: Not available.

rare findings”>Fig. 1. An AP radiograph of a 63-year-old man showed increased aeration and interstitial markings, but a pneumonic infiltration was not suspected (A). Consecutive axial CT images demonstrated a pneumonic infiltration in the posterobasal segment of the left lower lobe (B).

The sensitivity of radiography for detecting a mass was 60%, while the sensitivity for detecting solitary and multiple pulmonary nodules were quite low (8.7% and 30%, respectively). Although this is an extremely valuable CT finding for the patient’s long-term survival, we must remem- ber that the incidental detection of malignancies is not the goal of CT im- aging in the ED. However, CT may be useful as patients sometimes visit the ED unaware of a primary malignancy, but because of its acute pulmo- nary complications such as a pleural effusion, atelectasis, pneumothorax, and opportunistic infections.

Other rare findings

Radiography was less sensitive for detecting costal fractures, show- ing only two of six fractures. Although costal fractures may be classified as traumatic pathologies, these patients were not excluded from this study because they did not recall a history of trauma. Therefore, physi- cians should always remember that costal fractures can occur without awareness in elderly patients without significant trauma. This also dem- onstrated the utility of CT imaging in non-traumatic pathologies. Simi- larly, Langdorf et al. [2] reported that 60% of costal fractures were only identified by CT, even in patients with known thoracic blunt trauma. Exadaktylos et al. [10] reported that chest CT found Significant injuries in more than half of the patients with a normal chest radiograph, and 20% of the injuries identified with radiographs were more serious with CT, leading to Changes in management. In our series, radiography was also less sensitive for detecting pneumothorax, because it was not

detected in four of six cases. Three of them were elderly patients with AP X-rays. One patient was a 19-year-old male with PA radiography, but the pneumothorax could still not be identified (Fig. 3). In support of our findings, pneumothorax is the pathology most commonly missed in the literature [3].

One limitation of this study is its retrospective nature, so information on the diagnosis and management of the patients was based on results obtained from the hospital medical reports. Another limitation of the study was the duration between X-ray and CT imaging in some cases. In the majority of the patients, chest radiography and subsequent chest CT were obtained within a few hours. However, in a small propor- tion of the study group the time between X-ray and CT imaging was up to 12 h. For this reason there was a potential risk that some diagnoses (such as pulmonary edema) that could’ve been absent at the time of the X-ray and subsequently developed by the time the CT was done.

Conclusions

PA radiography had higher concordance with CT than AP radiogra- phy. Chest CT after a chest radiograph frequently changed the diagnosis and management of patients presenting to the ED with non-traumatic thoracic problems, and this finding was negatively correlated with age. Therefore, we suggest that chest CT is an appropriate choice for im- aging in patients presenting to the ED for non-traumatic thoracic prob- lems, particularly for elderly patients and when the X-ray is taken with the AP technique in a supine position.

Fig. 2. A PA chest radiograph of 44-year-old woman showed enlargement of the cardiac silhouette and increased interstitial markings, although a pneumonic infiltration was not suspected

(A). Axial CT images showed bilateral pneumonic infiltrations in the anterior and posterior segments of the right upper lobe (B) and in the basal segments of the left lower lobe (C).

Fig. 3. The PA radiograph of a 19-year-old man was evaluated as normal (A). Axial (B) and coronal reformatted (C) CT images showed a left-sided Spontaneous pneumothorax located in the apical and anterobasal regions of the lung.

Acknowledgements

We are thankful to Atalay Ekin for his contribution to the prepara- tion of this manuscript.

References

1. Cherney AR, Richardson DM, Greenberg MR, Choo EK, McGregor AJ, Safdar B. Preva- lence and clinical import of thoracic injury identified by chest computed tomogra- phy but not chest radiography in blunt trauma: multicenter prospective cohort study. Ann Emerg Med 2016;68:133-4.
2. Langdorf MI, Medak AJ, Hendey GW, Nishijima DK, Mower WR, Raja AS, et al. Prev- alence and clinical import of thoracic injury identified by chest computed tomogra- phy but not chest radiography in blunt trauma: multicenter prospective cohort study. Ann Emerg Med 2015;66:589-600.
3. Nguyen BM, Plurad D, Abrishami S, Neville A, Putnam B, Kim DY. Utility of chest computed tomography after a “normal” chest radiograph in patients with thoracic Stab wounds. Am Surg 2015;81:965-8.
4. Obata S. Indications for chest CT: retrospective study of cases with normal chest CT. Nihon Igaku Hoshasen Gakkai Zasshi Nippon Acta Radiol 1995;55:180-3.
5. Cantrill SV, Karas SJ, American College of Emergency Physicians. Cost-effective diag- nostic testing in emergency medicine : guidelines for appropriate utilization of clin- ical laboratory and radiology studies. Dallas, Tex.: American College of Emergency Physicians; 1994
6. Syrjala H, Broas M, Suramo I, Ojala A, Lahde S. High-resolution computed tomogra- phy for the diagnosis of community-acquired pneumonia. Clin Infect Dis 1998;27: 358-63.
7. Pandharipande PV, Reisner AT, Binder WD, Zaheer A, Gunn ML, Linnau KF, et al. CT in the emergency department: a real-time study of changes in physician decision mak- ing. Radiology 2016;278:812-21.
8. Mandell LA, Wunderink RG, Anzueto A, Bartlett JG, Campbell GD, Dean NC, et al. In- fectious Diseases Society of America/American Thoracic Society consensus guide- lines on the management of community-acquired pneumonia in adults. Clin Infect Dis 2007;44:27-72.
9. Wipf JE, Lipsky BA, Hirschmann JV, Boyko EJ, Takasugi J, Peugeot RL, et al. Diagnosing pneumonia by physical examination: relevant or relic? Arch Intern Med 1999;24: 1082-7.
10. Exadaktylos AK, Sclabas G, Schmid SW, Schaller B, Zimmermann H. Do we really need routine computed tomographic scanning in the Primary evaluation of blunt chest trauma in patients with “normal” chest radiograph? J Trauma 2001;51: 1173-6.