a b s t r a c t

Objective: In this study, emergency physicians and on-call radiologists were compared regarding identification of fatal injuries on Computed tomographic scans in patients with trauma.

Materials and Methods: Multiply injured patients who were older than 18 years and underwent CT scanning were included in the study. The CT scans were interpreted by the responsible emergency physician. At the same time, these images were also evaluated by the on-call radiologist. Final evaluation was done 1 week later by a radiology instructor who knew the clinical follow-up of the patient.

Results: The study included 156 patients. The mean age of the patients included in the study was found to be

41.6 years. Less than half (33.5%) of the patients were female and 86.5% were male. A total of 482 CT scans were performed in the patients. Regarding brain CTs, the concordance rate for emergency physicians was 98%, whereas it was 94% for on-call radiologists. Regarding thoracic CTs, the concordance rate for emergency physi- cians was 91%, whereas was 93% for on-call radiologists. There was a perfect concordance (? value N 0.75) for on-call radiologists and emergency physicians in terms of brain and thoracic CTs. Regarding abdominal-pelvic CTs, the concordance rate for emergency physicians was 97%, whereas it was 98% for on-call radiologists. Moderate concordance (? range = 0.40-0.75) was detected for emergency physicians in terms of identification of liver, spleen, kidney, and intra-abdominal/Retroperitoneal hemorrhages. There was a perfect concordance (? value N 0.75) for pelvic fractures. Conclusion: In this study, it was shown that emergency physicians were successful in identifying fatal injuries on trauma CT images after a short-term training on interpretation of trauma CTs.

(C) 2016

1. Introduction

Trauma is a major health problem throughout the world. Trauma- related deaths rank as the first cause of death especially among the population aged 1 to 45 years and show Trimodal distribution. The second period, in which trauma-related deaths peak, is the period when patients with trauma are in the emergency department (ED). The emergency physicians in this period should master every stage of approach to patients with trauma. Rapid resuscitation of patients with trauma, rapid identification of fatal injuries, and appropriate approach to these injuries are lifesaving [1-3].

* Corresponding author. Tel.: +90 5306426185.

E-mail addresses: [email protected], [email protected] (Z.A. Kartal), [email protected] (N. Kozaci), [email protected] (B. Cekic), [email protected] (I. Beydilli), [email protected] (M. Akcimen), [email protected] (D.S. Guven), [email protected] (I.E. Toslak).

Computed tomography (CT) is a fast and descriptive diagnostic tool in patients with trauma. With its increasing availability, it has become a common tool used in the diagnostic management of patients with trauma in recent years. However, the interpretation of CT scans requires high levels of knowledge and skill; otherwise, interpretation errors arise and patient management is adversely affected by this situation. To provide an acceptable level of an accurate diagnosis, the radiologists who evaluate the images need to be informed about the patient’s condition. However, some studies on this subject reported that it was impossible for a ra- diologist to simultaneously evaluate CT scans because of numerical shortage of radiologists [4].

Emergency physicians are critical in the management of patients with trauma and are primarily responsible for the patient’s medical care. They are also the group of physicians with the most comprehensive information on the clinical condition of the patients with trauma because the overall evaluation and clinical follow-up of patients with trauma are performed by emergency physicians. The interpretation of fatal injuries by emergency physicians on CT scans is very important in terms of

http://dx.doi.org/10.1016/j.ajem.2016.08.044

0735-6757/(C) 2016

early detection of these injuries, ensuring earlier intervention to injuries and prevention of misdiagnosis [4].

As in many countries, there are serious issues in simultaneous reporting of scans by a radiologist also in our country. Therefore, emer- gency physicians must be able to identify fatal injuries that can be seen on CTs of patients with trauma. In this study, we aimed to demonstrate the success of emergency physicians in identifying fatal injuries on CT scans after a short-term training.

1. Materials and methods

This study was initiated in the ED of Antalya Training and Research Hospital (TRH) after ethics committee approval. Both theoretical and practical trainings on interpreting CT scans were provided to emergency physicians by radiology clinic instructors. The body was divided into 3 parts in this training, namely, head-spine, thorax, and pelvic. Trauma- related life-threatening conditions were identified for each part. These trainings were given on weekly training day of ED for a total of 8 hours for 2 consecutive weeks. The practical training was given for a total of 16 hours in 2 days.

The CT scans were performed using Hitachi, Tokyo, Japan 16-slice to- mography that was located in the emergency tomography unit of ED. Brain and cervical CT scans were unenhanced, and thoracic and abdom- inal CT scans were enhanced using intravenous Contrast agents. Bone windows of thoracic and abdominal CT scans were used for the evalua- tion of thoracic and lumbar spines.

The patients of this study were collected in a 1-year period between July 1, 2013 and June 30, 2014. The patients who were admitted to Antalya TRH ED because of trauma, who had multiple trauma symp- toms, and who underwent brain, spinal, thoracic, and Abdominal CT scans were included in this study. The exclusion criteria were as follows: patients who were evaluated in other centers and referred to Antalya TRH, patients who were admitted to ED from other centers, patients who did not undergo CT scans because of clinical instability, and pa- tients who were younger than 18 years.

A 3-page study form was completed for each patient included in the study. Life-threatening conditions for each system were tabulated for CT interpretations. The first page of the form was consent. Mechanism of trauma, vital signs and physical examination findings, laboratory results, and clinical outcome were recorded in the second page. The tables of life-threatening conditions in CT interpretations were available on the third page. Pathologies including fatal conditions in head and spinal trauma were asked in box A, chest trauma in box B, and ab- dominal trauma in box C (Annex 1, third page of study form).

The mechanisms of trauma, physical examination findings, clinical findings, outcome (discharged, hospitalized, died, etc), and performed studies were recorded. The CT scans were first evaluated by the emer- gency physicians and registered into the study form. During the assess- ment of the patients in the ED, CT scans were evaluated through Picture Archiving Communication Systems by on-call radiologists, and findings were recorded on SARUS hospital information management system. The SARUS records were decisive in terms of the patient’s diagnosis and treatment plan. radiologic findings were interpreted finally by a single radiology instructor using the patient’s entire clinical follow-up and results.

The emergency physicians, on-call radiologists, and radiology instructor did not know each other during the assessment of CT scans.

Data analysis

In this study, statistics were performed from findings available in the study form. Additional findings detected by physicians were not included in the statistics. Analysis of the data collected in the study was performed using Statistical Package for the Social Sciences 21 statis- tical software package (IBM Corporation, Chicago, IL).

Sensitivity, specificity, positive predictive value (PPV), negative pre- dictive value (NPV), positive likelihood ratio (PLR), negative likelihood ratio (NLR), and ? coefficient were calculated for interpretations of emergency physicians and on-call radiologists based on the final inter- pretations of the radiology instructor. Interrater agreement was graded according to ? values. A ? value of greater than 0.75 was considered as perfect concordance, 0.75 to 0.40 as moderate concordance, and less than 0.40 as poor concordance. To determine the statistical significance and assumptions of predictions, P b .05 with 95% confidence intervals was considered significant in all analyses. In addition, PLR greater than 10 and NLR less than 0.1 were considered significant. Frequency distri- butions and ?² tests were used for demographic analysis.

1. Results

A total of 179 forms were completed for the study. One hundred fifty-six patient forms were included in the study. Twenty patients were excluded because of being younger than 18 years, and 2 patients were excluded because they had CT scans performed in other centers.

The mean age of the patients included in the study was found to be

41.6 +- 17.2 years. Of all patients included in the study, 21 (13.5%) were female and 135 (86.5%) were male. A total of 482 CT scans, including 120 brain CTs, 142 thoracic CTs, 138 abdominal CTs, and 82 spinal CTs,

were performed in 156 patients.

The mechanisms of trauma in study patients were as follows: motor vehicle accidents in 36 (23%) patients, extravehicular traffic accidents in 42 (27%) patients, gunshot wounds in 3 (2%) patients, Stab wounds in 11 (7%) patients, fall from height in 46 (30%) patients, crushing injuries be- tween objects in 2 (1%) patients, and other reasons in 16 (10%) patients. Forty-six (29.5%) of 156 patients were discharged from the ED after 12 hours of follow-up. Forty-seven (30%) patients were hospitalized to intensive care unit, and 58 (37%) patients were hospitalized to the

service. Two (1.3%) patients died in the ED.

Of 120 patients with head injury, 31 (26%) had positive CT findings. The detected pathologies were as follows: Skull fractures in 14 (12%) pa- tients, Brain edema in 21 (18%) patients, herniation in 8 (7%) patients, Intraparenchymal hemorrhage/contusion in 18 (15%) patients, subarach- noid hemorrhage (SAH) in 17 (14%) patients, and subdural/epidural hematoma (SDH/EDH) in 16 (13%) patients. Brain CT finding was normal in 89 (74%) patients. The concordance and discordance rates for emer- gency physicians regarding detection of pathologies on brain CTs were 98% (118 patients) and 2% (2 patients), respectively. However, the concordance and discordance rates for on-call radiologists were 94% (112 patients) and 6% (8 patients), respectively. Compared with the final report, there was a perfect concordance (? value N 0.75) regarding brain CT interpretations of both on-call radiologists and emergency phy- sicians. Data on brain CT evaluation of emergency physicians and on-call radiologists are shown in Tables 1 and 2.

Ten (7%) of 149 patients with spinal trauma had vertebral fractures.

Regarding spinal CTs, emergency physicians had a concordance rate of 96%, discordance rate of 4%, and ? value of 0.73 (perfect concordance). However, on-call radiologists had a concordance rate of 99%, discor- dance rate of 1%, and ? value of 0.89 (perfect concordance). There

Table 1

The interpretation success of the emergency physicians on brain CT

	Spes.	Sens.	NPV	PPV	PLR	NLR	Area
Brain edema	97	81	96	95	27	0.19	0.89
Parenchymal hemorrhage/contusion	96	100	100	91	25	0.0	0.98
SDH/EDH	99	88	98	93	88	0.12	0.93
SAH	98	94	99	89	47	0.16	0.96
Skull fractures	99	93	99	93	99	0.17	0.96
Herniation	100	88	99	100	88	0.01	0.94
Presence of pathological finding	98	100	100	94	45	0.0	0.99
Spinal vertebral fractures	94	96	99	64	16	0.04	0.93

Abbreviations: Area, area under the curve; Sens., sensitivity; Spes., specificity.

Table 2

The interpretation success of the on-call radiologists on brain CT

Table 4

The interpretation success of the on-call radiologists on thorax CT

	Spes.	Sens.	NPV	PPV	PLR	NLR	Area			Spes.	Sens.	NPV	PPV	PLR	NLR	Area
Brain edema	97%	86%	96%	95%	28.56	0.14	0.91		Pneumothorax	100%	87%	95%	97%	87	0.13	0.93
Parenchymal	98%	78%	94%	88%	39	0.22	0.88		Hemothorax	100%	82%	98%	100%	82	0.18	0.91
hemorrhage/contusion									pulmonary contusion/laceration	94%	81%	88%	90%	13.5	0.2	0.93
SDH/EDH	100%	94%	99%	100%	94	0.06	0.97	Presence of pathological finding 99% 97% 93% 93% 33 0.1 0.93
SAH	100%	88%	98%	100%	88	0.01	0.94
Skull fractures	100%	100%	100%	100%	100	0.001	1
Herniation	100%	100%	100%	100%	100	0.001	1
Presence of pathological finding	96%	87%	96%	87%	19	0.13	0.91	In studies on the management of patients with trauma, CT has been
Spinal vertebral fractures	90%	99%	99%	90%	25	0.1	0.95	reported to be a very effective method in the determination of injury [6].

were minor vertebral fractures with no effect on clinical management in patients who were interpreted false negatively by both physicians.

Of 142 patients with Thoracic trauma, 71 (50%) had positive CT findings. The detected pathologies were as follows: pneumothorax in 39 (27.5%) patients, hemothorax in 17 (12%) patients, and pulmonary contusion/laceration in 57 (40%) patients. There was Mediastinal emphysema in 1 patient and diaphragmatic rupture in 1 patient. It was seen that the emergency physicians had diagnosed 1 patient with diaphragmatic rupture and 1 patient with mediastinal emphysema. None of the patients had major cardiovascular pathology or pericardial effusion. Regarding thoracic CTs, emergency physician had a concordance rate of 91% (129 patients) and discordance rate of 9% (13 patients). However, the concordance and discordance rates for on-call radiologists were 93% (121 patients) and 7% (10 patients), respectively. Compared with the final report, there was a perfect concordance (? value N 0.75) regarding thoracic CT interpretations of both on-call radiologists and emergency physicians. However, preexisting lung pathology in 1 patient was interpreted false positively as pulmonary contusion by both physi- cians. The false-negative interpretations for pneumothorax and hemotho- rax were minimal with no effect on clinical management of patients. Data on thorax CT evaluation of emergency physicians and on-call radiologists are shown in Tables 3 and 4.

Of 138 patients with abdominal and/or pelvic thoracic trauma, 39 (28%) had positive CT findings. The detected pathologies were as follows: liver injury in 8 (5.8%) patients, spleen injury in 8 (5.8%) patients, kidney injury in 2 (1.4%) patients, intra-abdominal or retro- peritoneal hemorrhage in 20 (14.5%) patients, and pelvic fracture in 20 (14.5%) patients. None of the patients had pancreatic injury or urinary bladder rupture. Regarding abdominal CTs, emergency physicians had a concordance rate of 97% (134 patients) and discordance rate of 3% (4 patients). However, the concordance and discordance rates for on-call radiologists were 98% (135 patients) and 2% (3 patients), respec- tively. Data on abdominal CT evaluation of emergency physicians and on-call radiologists are shown in Tables 5 and 6.

1. Discussion

Emergency departments, where the Initial diagnosis and treatment for patients with trauma are performed, are the entry points to health care services and are very crowded especially in Developing countries. Rapid and accurate detection of Organ injuries and life-threatening conditions in patients with trauma in the emergency environment is very critical for patients’ lives [5].

Table 3

The interpretation success of the emergency physicians on thorax CT

Spes. Sens. NPV PPV PLR NLR Area

Pneumothorax 97% 100% 100% 93% 97 0.0 0.99

Hemothorax 97% 94% 99% 80% 16 0.06 0.96

Therefore, CT has become a widely used Diagnostic modality in EDs. Because of leading to high-dose radiation exposures, CT imaging is not recommended unless it is necessary [7]. However, it is used frequently because it can rapidly detect organ injuries, show exact localizations of the injuries, visualize many areas of the body, and provide images with different sections. Because most hospitals have CT scanners, there is no need to transport patients with severe injuries to long distances. In today’s medicine, CT has become essential as a primary screening tool in acute illnesses and patients with trauma [8,9].

Despite technological developments, there is no system for the in- terpretation of CT images yet. Simultaneous interpretation of CT images by a radiologist is not possible in many institutions because of shortage of radiologists in Japan. That is why most emergency physicians in Japan interpret the CT images of the patients with trauma and plan treatments [4]. Although Hunter and colleagues [10] reported that CT scans in the EDs could only be interpreted by radiologists in major medical institu- tions and university hospitals, Torreggiani et al [11] stated that simulta- neous radiologist interpretation was not possible in many institutions and that radiology interpretations could be reported after 48 hours. Therefore, emergency physicians complain about the inadequacy of current reporting system. These problems will continue in the long term, as long as effective measures are taken. One of the most important problems regarding CT interpretation is the erroneous interpretation of the images. Several studies on interpretation errors have been con- ducted. In a study by Tieng et al [12], 10% major discordance rate and 20% minor discordance rate were detected for on-call radiologists in the whole-body CT interpretations performed in the ED.

In a study by Agostini et al [13], whole-body CT scans of multiply in- jured patients were reevaluated, and splenic rupture, thoracic vertebral fractures, and EDH that changed the patient management completely but did not affect mortality have been identified in patients. In a study in a primary trauma center, Chung et al [14] compared CT interpreta- tions of emergency physicians and radiologists and found a 2% major discordance rate between emergency physicians and radiologists regarding whole-body CT interpretations, including thoracic, abdominal, and pelvic CTs. In our study, the discordance rates for on-call radiologists regarding interpretation of CT images of the patients were as follows: 6% for brain CTs, 7% for thoracic CTs, and 2% for abdominal-pelvic CTs. There- fore, emergency physicians who manage patients should keep in mind that on-call radiologists could make mistakes in the CT interpretations. These errors can be reduced by increasing the CT knowledge of inter- vening physicians. Indeed, in a study by Arentz et al [15] in patients with trauma, it was shown that surgery consultants accurately identi- fied acute injuries on brain, thoracic, abdominal, and pelvic CT scans.

Table 5

The interpretation success of the emergency physicians on abdominal CT

	Spes.	Sens.	NPV	PPV	PLR	NLR	Area
Liver injury	97%	88%	99%	64%	29	0.19	0.92
Spleen injury	97%	88%	99%	64%	29	0.002	0.92
Kidney injury	100%	50%	99%	100%	5	0.005	0.75
The abdominal/retroperitoneal	97%	70%	95%	82%	23	0.003	0.84
hemorrhage

Pulmonary contusion/laceration	85%	95%	96%	80%	6.3	0.06	0.90	Pelvic fracture	98%	100%	100%	90%	50	0.0	0.99
Presence of Pathological finding	86%	99%	98%	88%	7	0.001	0.92	Presence of pathological finding	96%	100%	100%	91%	25	0.9	0.98

Table 6

The interpretation success of the on-call radiologists on abdominal CT

	Spes.	Sens.	NPV	PPV	PLR	NLR	Area
Liver injury	99%	100%	100%	89%	100	0.001	1.0
Spleen injury	100%	88%	99%	100%	88	0.12	0.94
Kidney injury	100%	100%	100%	100%	100	0.001	1.0
The abdominal/retroperitoneal hemorrhage	98%	90%	98%	90%	45	0.1	0.94
Pelvic fracture	100%	95%	100%	100%	95	0.05	0.98
Presence of pathological finding	98%	97%	99%	95%	48	0.03	0.98

Several studies have been conducted comparing the emergency physicians and radiologists regarding radiologic imaging techniques. Most of these studies are related to ultrasound and x-rays [16,17]. How- ever, studies on CT interpretations by emergency physicians in both patients with and without trauma are being conducted in recent years. There are still very few studies. In these few studies, it was found that the ability of emergency physicians to interpret CTs was good [18,19].

Brain CT is one of the most requested tests in the EDs, and emergency physicians often plan the management of the patients without waiting for the radiologist’s interpretation. Therefore, the accuracy of the inter- pretation made by the emergency physicians is very important. In a study, Dolatabadi et al [5] found that emergency physicians had 86.5% sensitivity, 81.4% specificity, 86.9% PPV, 86.9% NPV, 4.6 PLR, and 0.16 NLR in brain CT interpretations based on Radiology reports. In a study by Khan et al [20], the concordance rate, discordance rate, and ? value of emergency physicians in brain CT interpretations were found to be 87.14%, 12.86%, and 0.64, respectively. Unlike our study, these studies also included patients with trauma. Indeed, in a study by Harding et al [21], the discordance rate of emergency physicians in brain CT interpre- tations was found to be 2%. Similarly, the discordance rate of emergency physicians in brain CT interpretations was 2% in our study. Unlike other studies, the ? values of both emergency physicians and on-call radiolo- gists were calculated for brain edema, intracranial hemorrhage, and skull fractures in our study, and the ? values were found to be greater than 0.75 for all, indicating perfect concordance. The higher rates in our study may be due to inclusion of only patients with trauma.

Although more than 130 000 severe spinal injuries are reported each year in Europe, only 15 000 of these injuries include cervical spine injury. Standard (anteroposterior and lateral) cervical radiographs can visualize vertebral fractures; however, it was found that 23% to 57% of cervical ver- tebral fractures were overlooked on standard radiographs compared with multidetector CT. Therefore, multidetector CT is recommended as first-choice imaging modality in severe spinal traumas [22]. In a study by Van Zyl et al [23], the specificity, sensitivity, and NLR values of emer- gency physicians in accurate diagnosis of cervical vertebral injuries on CTs were found to be 76%, 87%, and 0.18, respectively; however, it was seen that the sensitivity of emergency physicians in excluding clinically significant cervical injuries was not sufficient. Although the interpreta- tion of emergency physicians for vertebral injuries was at moderate level (? = 0.74) in our study, perfect concordance (? = 0.86) was found for on-call radiologists. It was seen that emergency physicians had higher false-positive rates, and this might be due to emergency physicians interpreting vascular structures as fractures. However, the cervical vertebra fracture that was false negatively interpreted by emergency physicians was at the tip of the spinous process and was not affecting the clinical condition of the patient.

Undiagnosed abdominal injuries constitute most of preventable cause

of death due to trauma [6]. Abdominal traumas account for 15% to 20% of all trauma-related deaths. These deaths are due to hemorrhage-induced hypovolemic shock in the early stages. Abdominal injuries can result in solid organ, hollow organ, diaphragm, and retroperitoneal injuries. Focused sonographic assessment (focused assessment with sonography for trauma-FAST) has been used since 1980 as part of the routine examination in the evaluation of patients with abdominal trauma in EDs. Conducted studies have revealed that emergency physicians are

successful in detecting intra-abdominal free fluid by FAST [24-26]; however, CT has become the criterion standard for the determination of abdominal organ injuries. The CT is a superior method compared with other methods in terms of detection of organ-specific and retro- peritoneal hemorrhages. In a study, Solid organ injuries were detected by CT in 29% of patients who were evaluated as not having free fluid (hemoperitoneum) by previous bedside ultrasonography (FAST) [26].

Studies on the ability of emergency physicians to interpret abdominal CTs revealed that the discordance rates of emergency physicians and radiologists regarding CT interpretations were 16.7% and 12.2%, respec- tively [27]. However, unlike our study, nontraumatic acute abdominal injuries were evaluated in that study. Considering intra-Abdominal pathologies, although the concordance and discordance rates of emer- gency physicians regarding abdominal-pelvic CT interpretations were 96% and 4%, respectively, they were found to be 98% and 2% for on-call radiologists. However, ? values of emergency physicians were in the 0.40 to 0.75 range for liver, spleen, kidney injuries, and intra- abdominal free fluids; and moderate concordance was detected. The ? value of emergency physicians for the detection of pelvic fractures was 0.96, demonstrating a perfect concordance. There was a perfect con- cordance (? N0.75) for on-call radiologists regarding the detection of liver injury, spleen injury, intra-abdominal/retroperitoneal hemorrhage, and pelvic fractures in abdominal-pelvic CT interpretations. Moderate concordance rates (?, 0.50-0.75) were found for kidney injuries. These results show us that emergency physicians need a little more long- term training for abdominal CT interpretations.

Chest injuries (thoracic injuries) are common in patients with multiple trauma. Thoracic traumas accompany about 10% to 15% of all traumas and can be fatal if not diagnosed. Thoracic traumas account for 25% of trauma-related deaths [6,28,29]. Airway obstruction, Tension pneumothorax, open pneumothorax, flail chest, massive hemothorax, and cardiac tamponade were described as injuries that should be identi- fied during primary examination in patients with chest trauma. On secondary examination, simple pneumothorax, hemothorax, pulmonary contusion, Tracheobronchial injuries, blunt cardiac injury, traumatic aortic injuries, diaphragm injuries, and Esophageal rupture should be identified [6].

Accurate diagnosis in thoracic trauma-related injuries is dependent

on accurately knowing the clinical data and radiographic findings. Life- threatening conditions such as tension pneumothorax, hemothorax, flail chest, and mediastinal abnormalities can be recognized on chest radiographs. However, alternative diagnostic methods should be re- ferred in case life-threatening conditions could not be distinguished on poor quality radiographs [29]. Emergency physicians use bedside ultrasonography along with examination for early diagnosis. Pericardial effusion, pleural effusion, and pneumothorax can be identified by ultra- sound in the evaluation of patients with chest trauma in the ED. For this purpose, extended FAST (e-FAST) protocol was expanded by adding thoracic trauma examination to the FAST examination that is used in the initial assessment of patients with trauma. However, in a study by Uz et al [30], they found lower sensitivity of e-FAST compared with CT in the detection of hemothorax and abdominal injuries, despite the successful implementation of e-FAST protocol by emergency physicians. Thoracic CT has begun to take an important role gradually in the evaluation of patients with chest injuries. This is because the thoracic CT is more accurate and sensitive than chest radiographs in detecting lesions that especially involve vessels, heart, pericardium, airways, mediastinum, chest wall, spine, and diaphragm [29]. In this case, the ability of emergency physician, trauma, or thoracic surgeon to interpret a CT seems to be very important in avoiding serious errors that may affect patient management. In our study, compared with final findings, concordance and discordance rates of emergency physicians regarding thoracic CT interpretations were found to be 91% (129 patients) and

9% (13 patients).

However, false-positive result was due to erroneous interpretation of preexisting lung pathology as pulmonary contusion in 1 patient.

Pneumothorax and hemothorax pathologies were minimal in false- negative results and did not affect the clinical management. In our study, perfect concordance (? N0.75) was determined for the presence of Pathological findings in thoracic CT interpretations of emergency physicians. This success of emergency physicians may depend on management of patients with trauma by themselves and interpretation of CT images with this information.

The limitation of our study was that this study only included life- threatening conditions in patients with trauma. It did not include non- life-threatening conditions and other preexisting pathologies. incidental findings identified by radiologists were excluded from the study, even if they were clinically significant. These are the limitations of our study.

In conclusion, it was seen that emergency physicians were very successful especially in thoracic and brain CT interpretations, and moderately successful in abdominal and vertebral CT interpretations after CT interpretation training for identification of life-threatening conditions. Both on-call radiologists and emergency physicians can make mistakes in the interpretation of CT images. Error rates are close to each other, and these interpretation errors are not sufficient to affect the patient’s management. However, interpretation of CT images by emergency physicians can reduce interpretation errors of on-call radiologists. It is possible for emergency physicians to identify life- threatening conditions with short-term CT training.

Acknowledgments

ZAK is responsible for the integrity of the work as a whole from inception to published article. NK is responsible for the integrity of the work as a whole from inception to published article and has a contribu- tion to interpretation of data. BC has a contribution to conception and design of article. IB has a contribution of drafting the article. DSG is responsible for analyzing the data. MA and IET have a contribution to data acquisition.

References

1. Brunet PH, Cameron PA. Section 21: trauma, chapter 250: trauma in adults. In: JE T, JS S, MD C, JO M, RK C, GD M, editors. Tintinalli’s emergency medicine. 7th. New York: McGraw-Hill CO; 2010.
2. Markopoulou A, Argyriou G, Charalampidis E, Koufopoulou A, Nestor A, Nanas S, et al. Time-to-treatment for critically ill-polytrauma patients in emergency department. Health Sci J 2013;7:81-9.
3. Hui CM, Mac Gregor JH, Tien HC, Kortbeek JB. radiation dose from initial trauma assessment and resuscitation: review of the literature. Can J Surg 2009;52(2):147-52.
4. Ikegami Y, Suzuki T, Nemoto C, Tsukada Y, Hasegawa A, Shimada J, et al. Establishment and implementation of an effective rule for the interpretation of computed tomography scans by emergency physicians in blunt trauma. World J Emerg Surg 2014;9:40.
5. Dolatabadi A, Baratloo A, Rouhipour A, Abdalvand A, Hatamabadi H, Forouzanfar M, et al. Interpretation of computed tomography of the head: emergency physicians versus radiologists. Trauma Mon 2013;18(2):86-9.
6. Flohr TG, Bruder H, Stierstorfer K, Petersilka M, Schmid TB, McCollough CH. Image reconstruction and Image quality evaluation for a dual source CT scanner. J Med Phys 2008;35:5882-97.
7. Tien HC, Tremblay LN, Rizoli SB, Gelberg J, Spencer F, Caldwell C, et al. Radiation expo- sure from diagnostic imaging in severely injured trauma patients. J Trauma 2007;62: 151-6.
8. Broder J, Warshauer DM. Increasing utilization of computed tomography in the adult emergency department. J Emerg Radiol 2006;13:25-30.
9. Smith CB, Barrett TW, Berger CL, Zhou C, Thurman RJ, Wrenn KD. Prediction of blunt traumatic injury in High-acuity patients: Bedside examination vs computed tomogra- phy. Am J Emerg Med 2011;29:1-10.
10. Hunter TB, Krupinski EA, Hunt KR, Erly WK. Emergency department coverage by academic department of radiology. J Acad Radiol 2000;7:165-70.
11. Torreggiani WC, Nicolaou S, Lyburn ID, Harris AC, Buckley AR. Emergency radiology in Canada: a national survey. Can Assoc Radiol J 2002;53:160-7.
12. Tieng N, Grinberg D, Li SF. Discrepancies in interpretation of ED body computed tomographic scans by radiology residents. Am J Emerg Med 2007;25(1):45-8.
13. Agostini C, Durieux M, Milot L, Kamaoui I, Floccard B, Allaouchiche B, et al. Value of double reading of whole body CT in polytrauma patients. J Radiol 2008;89(3 Pt 1): 325-30.
14. Chung JH, Strigel RM, Chew AR, Albrecht E, Gunn ML. Overnight resident interpreta- tion of torso CT at a level 1 trauma center an analysis and review of the literature. Acad J Radiol 2009;16(9):1155-60.
15. Arentz C, Griswold JA, Halldorsson A, Quattromani F, Dissanaike S. Best poster award: accuracy of surgery residents interpretation of computed tomography scans in trauma. Am J Surg 2008;196(6):809-12.
16. Summers SM, Scruggs W, Menchine MD, Lahham S, Anderson C, Amr O, et al. A prospective evaluation of emergency department bedside ultrasonography for the detection of acute cholecystitis. Ann Emerg Med 2010;56(2):114-22.
17. Fox JC, Solley M, Anderson CL, Zlidenny A, Lahham S, Maasumi K. Prospective evaluation of emergency physician performed bedside ultrasound to detect acute appendicitis. Eur J Emerg Med 2008;15(2):80-5.
18. Hochhegger B, Alves GR, Chaves M, Moreira AL, Kist R, Watte G, et al. Interobserver agreement between radiologists and radiology residents and emergency physicians in the detection of PE using CTPA. Clin Imaging 2014;38(4):445-7.
19. Klein EJ, Koenig M, Diekema DS, Winters W. Discordant radiograph interpretation between emergency physicians and radiologists in a pediatric emergency depart- ment. Pediatr Emerg Care 1999;15(4):245-8.
20. Khan A, Qashqari S, Al-Ali AA. Accuracy of non-contrast CT brain interpretation by emergency physicians: a cohort study. Pak J Med Sci 2013;29(2):549-53.
21. Harding J, Craig M, Jakeman N, Young R, Jabarin C, Kendall J. Emergency physician interpretation of head CT in trauma and suspected subarachnoid haemorrhage-is it viable? An audit of current practice. Emerg Med J 2010;27(2):116-20.
22. Imhof H. In: Marincek B, Dondelinger RF, editors. Imaging of spinal injuries, section 2; traumatic injuries: imaging and intervention, chapter 2-5. Emergency radiology: imaging and intervention. Berlin Heidelberg: Springer-Verlag; 2007. p. 2-5 [141-155].
23. Van Zyl HP, Bilbey J, Vukusic A, Ring T, Oakes J, Williamson LD, et al. Can emergency physicians accurately rule out clinically important Cervical spine injuries by using computed tomography? Can J Emerg Med 2014;16(2):131-5.
24. Shojaee M, Faridaalaee G, Sabzghabaei A, Safari S, Mansoorifar H, Arhamidolatabadi A, et al. Sonographic detection of abdominal free fluid: emergency residents vs radiology residents. Trauma Mon 2013;17(4):377-9.
25. Brenchley J, Walker A, Sloan JP, Hassan TB, Venables H. Evaluation of Focussed Assessment with Sonography in Trauma (FAST) by UK emergency physicians. Emerg Med J 2006;23(6):446-8.
26. Chiu WC, Cushing BM, Rodriguez A, Ho SM, Mirvis SE, Shanmuganathan K, et al. Abdominal injuries without hemoperitoneum: a potential limitation of Focused abdominal sonography for Trauma (FAST). J Trauma 1997;42:617.
27. Kang MJ, Sim MS, Shin TG, Jo IJ, Song HG, Song KJ, et al. Evaluating the accuracy of emergency medicine resident interpretations of abdominal CTs in patients with non-traumatic abdominal pain. J Korean Med Sci 2012;27:1255-60.
28. Ebrahimi A, Yousefifard M, Mohammad Kazemi H, Rasouli HR, Asady H, Moghadas Jafari A, et al. Diagnostic accuracy of chest ultrasonography versus chest radiography for identification of pneumothorax: a systematic review and meta-analysis. Tanaffos 2014;13(4):29-40.
29. Gavelli G, Napoli G, Bertaccini P, Battista G, Fattori R. In: Marincek B, Dondelinger RF, editors. Chapter 2-6; imaging of thoracic injuries, section 2; traumatic injuries: imaging and intervention, emergency radiology: imaging and intervention. Berlin Heidelberg: Springer-Verlag; 2007. p. 2-6 [155-17].
30. Uz I, Yuruktumen A, Boydak B, Bayraktaroglu S, Ozcete E, Cevrim O, et al. Impact of the practice of “extended Focused Assessment with Sonography for Trauma (e-FAST)” on clinical decision in the emergency department. Ulus Travma Acil Cerrahi Derg 2013; 19(4):327-32.