Esophagography After Pneumomediastinum Without CT Findings of Esophageal Perforation: Is It Necessary?
Abstract
OBJECTIVE. The purpose of our study was to determine the necessity of fluoroscopic esophagography in patients with pneumomediastinum on CT but without CT findings of esophageal perforation.
MATERIALS AND METHODS. From January 1, 2006, through December 31, 2010, there were 4305 fluoroscopic esophagography examinations including 533 with CT identified from a search of our PACS. Patients with pneumomediastinum on CT who were subsequently referred for emergent fluoroscopic esophagography to exclude esophageal perforation were enrolled. Fluoroscopic esophagography examinations performed within 3 days of CT were included. Patients with a history of esophageal disease were excluded. As a result, 103 patients were enrolled in the study; patients were divided into groups on the basis of whether there was additional clinical history of esophageal damage (trauma group) or not (nontrauma group). Images were reviewed by two board-certified radiologists blinded to the clinical data and radiologic reports for the presence or absence of esophageal perforation. A positive result on CT was defined as esophageal injury or periesophageal infiltration that coexisted with periesophageal air. A positive fluoroscopic esophagography result was defined as oral contrast medium leakage from the esophagus.
RESULTS. Esophageal perforation was diagnosed in 15 of the 103 patients. The CT findings were significantly correlated with esophageal perforation (p < 0.001 in the trauma group, and p = 0.001 in the nontrauma group). The respective sensitivity and negative predictive value (NPV) of CT versus fluoroscopic esophagography in the trauma group were 100% versus 66.7% and 100% versus 87.9%; in the nontrauma group, the sensitivity and NPV were 100% for CT and fluoroscopic esophagography. Thus, the sensitivity and NPV of CT were either superior or equal to those of fluoroscopic esophagography.
CONCLUSION. The results of our study suggest that performing fluoroscopic esophagography in patients with pneumomediastinum is unnecessary when CT is negative for esophageal perforation.
Pneumomediastinum is a common event confronted by clinicians in an emergency department setting that infrequently results from esophageal perforation [1]. Esophageal perforation is a life-threatening condition with a mortality of approximately 19.7% [2, 3]. Early diagnosis and management of esophageal perforation positively influence patient outcome [4]. When esophageal perforation is suspected, fluoroscopic esophagography is the method of choice for diagnosis [5]. A radiologist usually performs fluoroscopic esophagography emergently at the request of the surgeon to exclude esophageal perforation. In clinical practice, the incidence of esophageal perforation observed by fluoroscopic esophagography is low. However, some drawbacks of fluoroscopic esophagography have been identified including the following: Patients must be transported to the fluoroscopy suite, additional medical personnel are required, an optimal fluoroscopic esophagography examination is difficult to perform in seriously ill patients, and false-negative results may occur [6]. Besides these drawbacks, contrast medium aspiration during fluoroscopic esophagography can cause pulmonary edema if a water-soluble agent is used or severe pulmonary inflammation if a large amount of barium is aspirated.
CT is considered to be the preferred imaging modality in esophageal emergencies [7, 8]. Esophageal wall thickening and defect and periesophageal fluid and air on CT are signs that suggest esophageal perforation [1, 8–10]. Because the esophagus is wrapped in the mediastinum, esophagal fluid content leakage might be trapped near the perforator. In this study we hypothesized that leakage from the esophagus—the diagnostic basis of the fluoroscopic esophagography examination—could be depicted on CT as esophageal injury or periesophageal infiltration in combination with periesophageal air. CT is widely used as the primary survey for critically ill patients in the emergency department. However, an optimal CT protocol for esophageal perforation is technically difficult in ill patients. For esophageal perforation, pneumomediastinum is believed to have high sensitivity but poor specificity. The purpose of our study was to determine the necessity of fluoroscopic esophagography in patients with pneumomediastinum on CT but without CT findings of esophageal perforation.
Materials and Methods
Research Design
Our hospital is a 3000-bed tertiary care, university-affiliated hospital. The institutional review board of our hospital approved this study. From January 1, 2006, through December 31, 2010, there were 4305 fluoroscopic esophagography examinations including 533 with CT identified from a search of our PACS. Patients with pneumomediastinum on CT who were subsequently referred for emergent fluoroscopic esophagography to exclude esophageal perforation were enrolled. Patients were excluded if they had a history of esophageal disease, such as esophageal cancer; had undergone surgery or radiation; had a corrosive injury; or had undergone an endoscopic intervention, such as esophageal stenting, sclerotherapy, balloon dilatation, or biopsy. Fluoroscopic esophagography examinations performed within 3 days of CT were included. As a result, 103 patients met the study inclusion criteria (Fig. 1).
CT Protocol
This study was retrospective; thus, a uniform protocol could not be achieved. Additionally, some patients were referred from other hospitals with accompanying CT images. Patients whose CT images did not cover the entire esophagus or whose CT images were of poor quality were excluded. Most CT scans were obtained in our emergency department using an MDCT scanner (LightSpeed QX/i, GE Healthcare). Images were obtained in a high-speed mode with a table speed of 11.25 mm per rotation, 1.25-mm acquisition thickness, and 5-mm slice thickness with a 5-mm interval. IV contrast medium was not routinely used. If contrast medium was administered, a uniphasic IV injection protocol of 100 mL of iohexol (350 mg I/mL [Omnipaque 350, GE Healthcare]) was given in accordance with the departmental protocol. No oral contrast medium was given.
Patient Grouping
The 103 patients were divided into groups on the basis of whether the clinical history indicated potential damage to the esophagus. The trauma group included patients with clinical history of potential direct damage to the esophagus such as identifiable blunt trauma, insults from endoscopic procedures, or ingestion of rigid foreign bodies. All other patients were included in the nontrauma group. All patients' medical charts and images were retrospectively reviewed.
Reference Standards
The diagnosis of esophageal perforation was established by positive findings of surgery, endoscopy, or fluoroscopic esophagography. Otherwise a negative diagnosis was considered in conditions supported by discharge notes and clinical follow-up for 1 month. Patients with a positive result on CT but a negative result on fluoroscopic esophagography were monitored clinically, were followed up with further investigation, or proceeded to surgery if clinical symptoms progressed.
Image Interpretation
All CT and fluoroscopic esophagography examinations were retrospectively reviewed by two board-certified radiologists blinded to the official reports and clinical diagnosis. The two readers, who are interested in emergency department image interpretation, had 8 and 5 years of experience, respectively, at the time of the study. CT findings were defined as follows: Esophageal injury was defined as esophageal wall thickening (thickness > 5 mm) [11] or esophageal wall defect (discontinuity of the wall); periesophageal infiltration, as increased focal stranding in periesophageal fat or a focal periesophageal fluid collection; and periesophageal air, as extraluminal air abutting the esophagus. We defined a positive CT result for esophageal perforation as esophageal injury or periesophageal infiltration coexisting with periesophageal air (Figs. 2 and 3) and a positive fluoroscopic esophagography result as oral contrast medium leakage. The images were reviewed from the hospital's PACS; for cases with interobserver disagreement, a consensus was reached after open discussion.
Fig. 2A —78-year-old woman with swallowed fish bone complicated by esophageal perforation.
A, CT image shows esophageal injury caused by fish bone (white arrow) with periesophageal infiltration as well as periesophageal air collection (black arrow).
Fig. 2B —78-year-old woman with swallowed fish bone complicated by esophageal perforation.
B, No discernible oral contrast medium leak can be identified on follow-up fluoroscopic esophagography image.
Fig. 3A —55-year-old woman with clinical diagnosis of Boerhaave syndrome.
A, CT image shows periesophageal infiltration (arrowhead) and esophageal wall defect (arrow) coexisting with conspicuous periesophageal air.
Fig. 3B —55-year-old woman with clinical diagnosis of Boerhaave syndrome.
B, Follow-up fluoroscopic esophagography image shows perforation at distal esophagus (arrow) on left wall corresponding to CT findings shown in A.
Statistical Analyses
The data were analyzed using the chi-square test (Pearson or Fisher, as appropriate) to determine the relation of the categoric data (sex, trauma, IV contrast use for CT, results of CT and fluoroscopic esophagography) with a final diagnosis of esophageal perforation. The Fisher exact test was used to assess interobserver agreement on CT readings with regard to the issue of IV contrast use. The Mann-Whitney U test was used to examine the relationship between the continuous variables of clinical parameters (age, time lag between CT and fluoroscopic esophagography) and the final diagnosis. The kappa statistic was used to assess interobserver agreement for interpretation of a single modality (CT, fluoroscopic esophagography) and interrater agreement between the diagnosis by these two modalities (CT, fluoroscopic esophagography). The McNemar test was used to compare the accuracy of fluoroscopic esophagography and CT in diagnosing esophageal perforation with reference to the final diagnosis. A p value of ≤ 0.05 was deemed statistically significant.
Results
Of the 103 patients, 68 were male and 35 were female; the mean age was 35.1 years (range, 4–83 years). Esophageal perforation was diagnosed in 15 patients (Table 1). Of those 15 patients, 12 underwent surgery and three received medical treatment. The surgical procedures of primary repair, esophagectomy, or mediastinotomy with or without drainage were chosen at the surgeon's discretion. Patient 12 diagnosed with esophageal perforation died of sepsis with multiple comorbidities such as hypertension and diabetes mellitus that were presumed to have contributed to the fatality. Inconsistent results of CT and fluoroscopic esophagography occurred in 13 patients. Nine patients were in the trauma group and four were in the nontrauma group. Of these nine patients, four were diagnosed as having an esophageal perforation caused by a foreign body.
| Patient No. | Cause of Perforation | Location of Perforation and Findings of Esophagus | Management |
|---|---|---|---|
| 1 | Swallowed FB | Cervical segment; endoscopic retrieval of FB failed, but perforation noted | Focal repair |
| 2 | Swallowed FB days earlier | Cervical segment; no FB was noted at fluoroscopic esophagography; a periesophageal abscess was detected on CT and a small perforation was seen at subsequent endoscopic examination | Abscess drainage |
| 3 | Swallowed FB days earlier | Thoracic segment; no FB was noted at esophagography; periesophageal abscess was found at thoracotomy | Esophagectomy |
| 4 | Esophagoscopy | Cervical segment at posterior wall | Nasogastric tubing bypass; healing established by endoscopy |
| 5 | Chest trauma | Thoracic segment | Mediastinotomy and drainage |
| 6 | Motorcycle accident | Cervical segment | Nasogastric tubing bypass; no leak detected on fluoroscopic esophagography |
| 7 | Corrosive injury | Thoracic segment | Esophagectomy |
| 8 | Boerhaave syndrome | Abdominal segment | Esophagectomy |
| 9 | Swallowed FB | Thoracic segment; endoscopic retrieval of FB failed, but perforation noted | Mediastinotomy and drainage |
| 10 | Boerhaave syndrome | Abdominal segment | Esophagectomy |
| 11 | Motorcycle accident | Cervical segment | Thoracotomy and staged surgery |
| 12 | Swallowed FB days earlier | Cervical segment near thoracic inlet | Drainage of neck and mediastinuma |
| 13 | Swallowed FB | Cervical segment near thoracic inlet; FB was lodged at cervical esophagus near thoracic inlet; endoscopic retrieval of FB was successful | Focal repair |
| 14 | Focal trauma at neck | Cervical segment | Mediastinotomy with drainage |
| 15 | Swallowed FB | Upper thoracic segment; complicated small periesophageal abscess at upper thoracic segment was noted at fluoroscopic esophagography | Nasogastric tubing bypass |
Note—FB = foreign body.
a
Patient died of sepsis and had multiple comorbidities.
Regarding the presence or absence of esophageal perforation, no significant difference in sex was observed; however, a significant difference was noted concerning age (p < 0.001) and history of trauma (p = 0.001). The mean age of the patients with esophageal perforation was older than of those without perforation (57.6 vs 31.3 years, respectively).
IV contrast-enhanced CT was performed in 86 patients and unenhanced CT was performed in 17 patients. CT with or without IV enhancement did not statistically correlate with the diagnosis of esophageal perforation (p = 0.455) or the interobserver agreement of CT readings (p = 0.835). The mean time lag between CT and fluoroscopic esophagography was 5.2 hours (range, 10 minutes–44 hours). Except in an extreme case of infectious pneumomediastinitis presumably related to a liver abscess in which CT and fluoroscopic esophagography were performed 44 hours apart, other studies were performed within hours of one another. In the current study, the mean time lag between examinations in patients with esophageal perforation and in those without esophageal perforation was not significantly different (p = 0.095).
Esophageal perforation was primarily located in the cervical segment (8/15, 53.3%), followed by the thoracic (5/15, 33.3%) and abdominal (2/15, 13.3%) segments. Among those patients with cervical perforation, the perforation was located primarily in the lower part around the thoracic inlet (6/8, 75%). Of these six cases, four were caused by foreign bodies and two by blunt trauma. The cause of esophageal perforation was a foreign body (7/15, 46.7%), blunt trauma (4/15, 26.7%), clinical suspicion of Boerhaave syndrome (2/15, 13.3%), endoscopic procedures (1/15, 6.7%), and corrosive injury (1/15, 6.7%). These clinical data are shown in Tables 2 and 3.
| Characteristic | Esophageal Perforation | p | |
|---|---|---|---|
| Yes (n = 15) | No (n = 88) | ||
| Age | < 0.001a | ||
| Mean ± SD | 57.6 ± 18.36 | 31.3 ± 17.16 | |
| Median | 58 | 23 | |
| Sex, no. (%) of patients | 0.138 | ||
| Male (n = 68) | 7 (46.7) | 61 (69.3) | |
| Female (n = 35) | 8 (53.3) | 27 (30.7) | |
| CT, no. (%) of patients | 0.455 | ||
| With IV contrast medium (n = 86) | 14 (93.3) | 72 (81.8) | |
| Without IV contrast medium (n = 17) | 1 (6.7) | 16 (18.2) | |
| Time lag between CT and fluoroscopic esophagography, h | 0.095 | ||
| Mean ± SD | 3.1 ± 2.05 | 5.5 ± 5.80 | |
| Median | 3 | 4 | |
| Clinical history, no. (%) of patients | 0.001b | ||
| Trauma ( n = 41) | 12 (80) | 29 (33) | |
| Fish bone (n = 12) | 7 | 5 | |
| Blunt (n = 28) | 4 | 24 | |
| Endoscopy (n = 1) | 1 | 0 | |
| Nontrauma (n = 62) | 3 (20) | 59 (67) | |
| Idiopathic cause (n = 53) | 0 | 53 | |
| Boerhaave syndromec (n = 4) | 2 | 2 | |
| Corrosive injury (n = 2) | 1 | 1 | |
| Mediastinitisd (n = 3) | 0 | 3 | |
a
Mann-Whitney U test.
b
Pearson chi-square test.
c
Clinical suspicion of Boerhaave syndrome.
d
The patients with mediastinitis had complications due to direct invasion of a deep neck infection (two patients) and liver abscess (one patient).
| Cause | No. (%) of Cases | Total | ||
|---|---|---|---|---|
| Cervical Segment | Thoracic Segment | Abdominal Segment | ||
| Foreign body | 4 | 3 | 0 | 7 (46.7) |
| Trauma | 3 | 1 | 0 | 4 (26.7) |
| Endoscope | 1 | 0 | 0 | 1 (6.7) |
| Boerhaave syndrome | 0 | 0 | 2 | 2 (13.3) |
| Corrosive injury | 0 | 1 | 0 | 1 (6.7) |
| Total | 8 (5 3.3) | 5 (33.3) | 2 (13.3) | 15 |
The correlation between esophageal perforation and CT and fluoroscopic esophagography results was significant (p < 0.001, Tables 4 and 5). None of the esophageal perforations was missed on CT. The percentage of positive CT results was 41.5% (17/41) in the trauma group and 11.3% (7/62) in the nontrauma group. The percentage of positive fluoroscopic esophagography results was 19.5% (8/41) in the trauma group and 4.8% (3/62) in the nontrauma group. In general, the sensitivity and negative predictive value (NPV) of CT were superior to fluoroscopic esophagography and reached 100%. In contrast, the specificity and positive predictive value (PPV) of fluoroscopic esophagography were superior. The interrater reliability between CT and fluoroscopic esophagography results was fair (trauma group: κ = 0.510, p < 0.001; and nontrauma group: κ = 0.571, p < 0.001). No statistical difference was found between the accuracy of CT and fluoroscopic esophagography (p = 0.267).
| Imaging Findings | Reference Standarda | |||||
|---|---|---|---|---|---|---|
| All Patients (n = 103) | Trauma Group (n = 41) | Nontrauma Group (n = 62) | ||||
| + | − | + | − | + | − | |
| CT | ||||||
| + | 15 | 9 | 12 | 5 | 3 | 4 |
| − | 0 | 79 | 0 | 24 | 0 | 55 |
| Fluoroscopic esophagography | ||||||
| + | 11 | 0 | 8 | 0 | 3 | 0 |
| − | 4 | 88 | 4 | 29 | 0 | 59 |
Note—Plus sign (+) indicates positive findings for esophageal perforation, and negative sign (–) indicates negative findings for esophageal perforation.
a
Surgery, endoscopy, or fluoroscopic esophagography.
| Performance Value | All Patients (n = 103) | Trauma Group (n = 41) | Nontrauma Group (n = 62) |
|---|---|---|---|
| CT | |||
| Sensitivity, % (95% CI) | 100 (100–100) | 100 (100–100) | 100 (100–100) |
| Specificity, % (95% CI) | 89.8 (83.4–96.1) | 82.8 (69.0–96.5) | 93.2 (86.8–99.6) |
| PPV (%) | 62.5 | 70.6 | 42.9 |
| NPV (%) | 100 | 100 | 100 |
| Accuracy, % (95% CI) | 91.3 (85.8–96.7) | 8 7.8 (7 7.8 – 9 7. 8) | 93.5 (87.4–99.7) |
| Positive likelihood ratio | 9.778 | 5.800 | 14.750 |
| p | < 0.001a | < 0.001a | 0.001a |
| Kappa value for interobserver agreement (p) | 0.944 (< 0.001) | 1 (< 0.001) | 0.816 (< 0.001) |
| Fluoroscopic esophagography | |||
| Sensitivity, % (95% CI) | 73.3 (51.0–95.7) | 66.7 (40.0–93.3) | 100 (100–100) |
| Specificity, % (95% CI) | 100 (100–100) | 100 (100–100) | 100 (100–100) |
| PPV (%) | 100 | 100 | 100 |
| NPV (%) | 95.7 | 8 7.9 | 100 |
| Accuracy, % (95% CI) | 96.1 (92.4–99.8) | 90.2 (81.2–99.3) | 100 (100–100) |
| Positive likelihood ratio | NA | NA | NA |
| p | < 0.001a | < 0.001a | < 0.001a |
| Kappa value for interobserver agreement (p) | 0.898 (< 0.001) | 0.845 (< 0.001) | 1 (< 0.001) |
| Kappa value for agreement between CT and fluoroscopic esophagography | 0.565 (< 0.001) | 0.510 (< 0.001) | 0.571 (< 0.001) |
Note—PPV = positive predictive value, NPV = negative predictive value, NA = not applicable.
a
p < 0.05 by Fisher test.
Interobserver agreement regarding the single-modality image readings on CT and fluoroscopic esophagography was very good to excellent. Concerning image readings, the observers disagreed about three examinations: two fluoroscopic esophagography examinations and one CT examination. In the two cases of interobserver disagreement about fluoroscopic esophagography, one case was finally diagnosed as confined esophageal perforation (Fig. 4) and the other as silent choking, which had been misread as esophageal perforation with little contrast medium leakage. Because of the trace amount of runoff contrast medium on those suboptimal images, this result was misleading. One patient who was the subject of interobserver disagreement on CT was diagnosed as having infected pneumomediastinitis associated with a liver abscess. Misreading was caused by overlooking a small periesophageal air bubble. A consensus was reached after open discussion.
Fig. 4A —70-year-old woman with diagnosis of esophageal perforation caused by swallowed fish bone.
A, CT image shows periesophageal fluid collection with air in thoracic segment (arrow).
Fig. 4B —70-year-old woman with diagnosis of esophageal perforation caused by swallowed fish bone.
B, Fluoroscopic esophagography image shows barely noticeable trace amount of contrast medium (arrow) has leaked from esophagus. Leak was missed by one radiologist but was seen by second radiologist; interpretation was corrected in open discussion.
Discussion
The main question addressed by the current study was whether CT is adequate compared with fluoroscopic esophagography in excluding esophageal perforation for patients with pneumomediastinum. This postulate is supported by the higher sensitivity and NPV of CT (up to 100%). In addition, a highly significant correlation between CT results and esophageal perforation was observed (p < 0.001 in the trauma group and p = 0.001 in the nontrauma group). Both the CT and fluoroscopic esophagography results were significantly correlated with esophageal perforation, but the interrater agreement between imaging modalities (CT and fluoroscopic esophagography) was fair. No statistical difference was found between the accuracy of CT and fluoroscopic esophagography. The sensitivity and NPV of CT were superior or equal to those of fluoroscopic esophagography and reached 100%. These results indicate the high value of CT compared with fluoroscopic esophagography in screening for esophageal perforation in patients with pneumomediastinum.
Thoracic CT can precisely assess periesophageal inflammation [3]; CT has been reported to have high sensitivity (100%) and specificity (85%) in detecting major aerodigestive tract injuries in patients with pneumomediastinum [12]. Other investigators have also asserted that CT can play both primary and complementary roles in the diagnosis and evaluation of esophageal emergencies [7, 8]. Investigators have reported that CT esophagography with ingested contrast medium can serve as a substitute for fluoroscopic esophagography in the detection of esophageal perforation [13] and is more sensitive than fluoroscopic esophagography [14]. At our institution, oral contrast medium is not routinely used in the emergency department to balance the risk of aspiration, choking, and emergent disease. Additionally, some patients with major trauma present with unclear consciousness and ingested contrast medium might mask luminal foreign bodies. With regard to CT without oral contrast medium, none of these studies had meticulously compared unenhanced CT with fluoroscopic esophagography.
In the current study, we noted that esophageal perforation was significantly correlated with trauma (p = 0.01) and patient age (p < 0.01). Accidental ingestion of foreign bodies was a major cause of esophageal perforation, particularly in the cervical esophagus. Investigators previously reported that foreign bodies usually affect the cervical esophagus and account for 80% of cervical perforations [15]. In cases of radiographically imperceptible foreign body impaction, CT is suggested to be the first line of diagnosis [16]. The use of barium for esophagography carries a risk of aspiration that may impede subsequent esophagoscopy [16]. We believe that the use of IV contrast medium increased the contrast of the studies to better depict esophageal injuries. However, in this limited study, no significant difference in interobserver agreement was found.
Blunt trauma with resultant esophageal perforation is an unusual event primarily related to motor vehicle crashes [17]. In the current study, blunt trauma–related esophageal perforation was not rare. Motorcycle crashes are prevalent in our country and our hospital is located adjacent to a highway, meaning that severely injured patients are frequently transferred to our institution. Beal et al. [18] reported that perforation occurs at the cervical and upper thoracic segments in 82% of cases with traumatic esophageal perforation. Dissanaike et al. [12] reported that pneumomediastinum occurred in up to 10% of patients with blunt thoracic and cervical trauma but that only 1.5% of cases were caused by esophageal perforation and that 1% of patients had an esophageal injury requiring surgical intervention. Thoracic esophageal injury is less common than cervical esophageal injury [19].
The incidence of esophageal perforation related to endoscopic procedures has increased because of the wide application of endoscopy. Endoscopic procedures are reported to be responsible for 60% of esophageal perforation cases [15]. In daily practice, both CT and fluoroscopic esophagography studies are ordered because this iatrogenic complication is infrequent, explaining its low percentage in our study.
When the diagnosis of esophageal perforation is in doubt in patients with spontaneous pneumomediastinum, chest CT or fluoroscopic esophagography should be performed [20]. We observed that Boerhaave syndrome was a cause of spontaneous esophageal rupture and CT revealed decisive criteria for the diagnosis [21]. It is an accurate technique to delineate esophageal rupture [22].
Caustic agents frequently burn the esophagus at the anatomic narrowing because of the relative delay in transit [17]. The esophagus is predisposed to damage by ingestion of alkaline agents rather than acidic agents. Alkalis have been shown to cause deep burns by liquefactive necrosis, whereas acids produce coagulative necrosis [17].
In our study, the prevalence of pneumomediastinum associated with esophageal perforation was 29.3% (12/41) in the trauma group and 4.8% (3/62) in the nontrauma group. The mortality rate of the patients with esophageal perforation was 6.7% (1/15), which is similar to a previous report [3].
Some drawbacks exist for CT diagnosis of esophageal perforation. First, the false-positives occurred in patients with chest trauma or infective mediastinitis presenting with periesophageal infiltration (Fig. 5). The paucity of fatty adventitia around the esophagus could make interpretation about the presence of periesophageal infiltration difficult. A confined microperforation would be imperceptible and could be successfully treated with medical conservative treatment [2, 17], resulting in misclassification as a false-positive CT result. These factors could decrease the PPV of CT. To define and classify perforation, fluoroscopic esophagography is more direct and specific for surgeons. In some scenarios of esophageal injury, endoscopic study is considered to be the first choice for diagnosis and treatment. Endoscopy would have precluded further study with fluoroscopic esophagography and prolonged the time lag between CT and fluoroscopic esophagography in our study.
Fig. 5A —54-year-old woman with motor vehicle crash–related vertebral body fracture and pneumomediastinum.
A, CT image shows periesophageal infiltration with air (arrow).
Fig. 5B —54-year-old woman with motor vehicle crash–related vertebral body fracture and pneumomediastinum.
B, Sagittal CT image with bone window depicts vertebral body fracture.
Fig. 5C —54-year-old woman with motor vehicle crash–related vertebral body fracture and pneumomediastinum.
C, Because spinal fracture precluded positioning for oblique view, fluoroscopic esophagography study was suboptimal and no esophageal perforation was identified.
Our study has recognized limitations. Because of the retrospective nature of our study, data about the radiation dose of CT and fluoroscopic esophagography could not be collected. Most patients with overt pneumomediastinum did not undergo both CT and fluoroscopic esophagography. In clinical practice, pneumomediastinum in most patients was noted on radiography and examined using fluoroscopic esophagography only to exclude esophageal perforation in our institution. In severely injured patients, such as those with active bleeding, cardiovascular injury, multiple trauma, or unstable vital signs, additional emergent fluoroscopic esophagography is not recommended. Furthermore, esophageal perforation is not a frequent source of pneumomediastinum, which led to the small positive sample size.
In conclusion, the results of CT significantly correlated with the reference standard for the diagnosis of esophageal perforation. A combination of CT findings that reflect esophageal injury, periesophageal fat infiltration, or periesophageal air, were used in this study as imaging criteria reflecting perforation. The sensitivity and NPV of CT for esophageal perforation were either superior or equal to those of fluoroscopic esophagography. The leakage of contrast medium from the esophagus, the diagnostic basis of the fluoroscopic esophagography examination, could be substituted by CT findings. For patients with pneumomediastinum but a negative CT result for esophageal perforation, additional fluoroscopic esophagography to exclude esophageal perforation is not recommended. However, follow-up of a positive CT result for esophageal perforation with fluoroscopic esophagography would confirm leakage and provide further information necessary for surgery.
References
1.
de Lutio di Castelguidone E, Merola S, Pinto A, Raissaki M, Gagliardi N, Romano L. Esophageal injuries: spectrum of multidetector row CT findings. Eur J Radiol 2006; 59:344–348
2.
Eroglu A, Turkyilmaz A, Aydin Y, Yekeler E, Karaoglanoglu N. Current management of esophageal perforation: 20 years experience. Dis Esophagus 2009; 22:374–380
3.
Huber-Lang M, Henne-Bruns D, Schmitz B, Wuerl P. Esophageal perforation: principles of diagnosis and surgical management. Surg Today 2006; 36:332–340
4.
Shaker H, Elsayed H, Whittle I, Hussein S, Shackcloth M. The influence of the ‘golden 24-h rule’ on the prognosis of oesophageal perforation in the modern era. Eur J Cardiothorac Surg 2010; 38:216–222
5.
Giménez A, Franquet T, Erasmus JJ, Martínez S, Estrada P. Thoracic complications of esophageal disorders. Radio-Graphics 2002; 22(Spec No):S247–S258
6.
Swanson JO, Levine MS, Redfern RO, Rubesin SE. Usefulness of high-density barium for detection of leaks after esophagogastrectomy, total gastrectomy, and total laryngectomy. AJR 2003; 181:415–420
7.
White CS, Templeton PA, Attar S. Esophageal perforation: CT findings. AJR 1993; 160:767–770
8.
Young CA, Menias CO, Bhalla S, Prasad SR. CT features of esophageal emergencies. Radio-Graphics 2008; 28:1541–1553
9.
Lee S, Mergo PJ, Ros PR. The leaking esophagus: CT patterns of esophageal rupture, perforation, and fistulization. Crit Rev Diagn Imaging 1996; 37:461–490
10.
De Lutio di Castelguidone E, Pinto A, Merola S, Stavolo C, Romano L. Role of spiral and multislice computed tomography in the evaluation of traumatic and spontaneous oesophageal perforation: our experience. Radiol Med (Torino) 2005; 109:252–259
11.
Berkovich GY, Levine MS, Miller WT Jr. CT findings in patients with esophagitis. AJR 2000; 175:1431–1434
12.
Dissanaike S, Shalhub S, Jurkovich GJ. The evaluation of pneumomediastinum in blunt trauma patients. J Trauma 2008; 65:1340–1345
13.
Fadoo F, Ruiz DE, Dawn SK, Webb WR, Gotway MB. Helical CT esophagography for the evaluation of suspected esophageal perforation or rupture. AJR 2004; 182:1177–1179
14.
Chirica M, Champault A, Dray X, et al. Esophageal perforations. J Visc Surg 2010; 147:e117–e128
15.
Brinster CJ, Singhal S, Lee L, Marshall MB, Kaiser LR, Kucharczuk JC. Evolving options in the management of esophageal perforation. Ann Thorac Surg 2004; 77:1475–1483
16.
Marco De Lucas E, Sádaba P, Lastra García-Barón P, et al. Value of helical computed tomography in the management of upper esophageal foreign bodies. Acta Radiol 2004; 45:369–374
17.
Wu JT, Mattox KL, Wall MJ Jr. Esophageal perforations: new perspectives and treatment paradigms. J Trauma 2007; 63:1173–1184
18.
Beal SL, Pottmeyer EW, Spisso JM. Esophageal perfor1ation following external blunt trauma. J Trauma 1988; 28:1425–1432
19.
Monzon JR, Ryan B. Thoracic esophageal perforation secondary to blunt trauma. J Trauma 2000; 49:1129–1131
20.
Jougon JB, Ballester M, Delcambre F, Mac Bride T, Dromer CE, Velly JF. Assessment of spontaneous pneumomediastinum: experience with 12 patients. Ann Thorac Surg 2003; 75:1711–1714
21.
Ghanem N, Altehoefer C, Springer O, et al. Radiological findings in Boerhaave's syndrome. Emerg Radiol 2003; 10:8–13
22.
Koffel C, Dangelse G, Italiano C, Hennia F, Potelon P, Haegy JM. Spontaneous rupture of the oesophagus: two cases recently seen in an intensive care unit (in French). Presse Med 2004; 33:250–252
Information & Authors
Information
Published In
Copyright
© American Roentgen Ray Society.
History
Submitted: November 20, 2012
Accepted: February 20, 2013
First published: October 22, 2013
Keywords
Authors
Metrics & Citations
Metrics
Citations
Export Citations
To download the citation to this article, select your reference manager software.