Clinical Imaging 51 (2018) 23–29

Contents lists available at ScienceDirect

Clinical Imaging
journal homepage: www.elsevier.com/locate/clinimag

Diagnosing conduit leak after esophagectomy for esophageal cancer by T

computed tomography leak protocol and standard esophagram: Is old school
still the best?
⁎
Diana Palacioa, ,1, Edith M. Maromb, Arlene Corread, Sonia L. Betancourt-Cuellarc,
Wayne L. Hofstetterd
a
Department of Medical Imaging, The University of Arizona, Banner Medical Center, Tucson, AZ 85724, United States
b
Department of Diagnostic Imaging, The Chaim Sheba Medical Center, Ramat Gan, Israel affiliated with the Tel Aviv University, Tel Aviv, Israel
c
Department of Radiology, Diagnostic Imaging Division, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, Houston, TX 77005, United States
d
Department of Surgery, Cardiothoracic Surgery Division, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, Houston, TX 77005, United States

A R T I C L E I N F O A B S T R A C T

Keywords: The imaging modalities available to evaluate anastomotic leak complicating esophagectomy include
Esophageal leak CT-Esophageal Protocol (CTEP) and esophagram. The purpose of this study was to compare the performance of
Esophagectomy these two modalities, alone or in combination, with the final diagnosis of leak established by endoscopy, surgery
Esophagram and/or the clinical course and evaluate management implications.
CT esophagram
Esophageal cancer

1. Introduction 2. Materials and methods

Esophageal conduit leak is a major complication after esopha-
gectomy and can result in serious morbidity, prolonged intensive care,
longer hospital admissions, and increased mortality [1–3]. In the past
decade, patient outcome after conduit leak has improved as a result of
prompt diagnosis and appropriate management [4–8] and effective
therapy relies heavily on early diagnosis. Standard Esophagram (SE)
has traditionally been used to evaluate conduit leaks. The use of SE in
the evaluation of a conduit leak has been validated in studies with small
numbers of patients [9–13]. Conduit leak after esophagectomy poses a
different clinical and imaging scenario than that of esophageal per-
foration. In this regard, SE early after esophagectomy can be difficult to
perform as patients have limited mobility and may be unable to
swallow the large amounts of contrast needed for evaluation. Since
2005 a CT esophageal conduit leak protocol (CTLP) has been im-
plemented at our institution to diagnose esophageal conduit leaks. The
purpose of this study was to review the role of CTLP and SE in the
diagnosis and treatment of patients with a clinical suspicion for a
conduit leak after esophagectomy for esophageal cancer.
In this retrospective study, we reviewed the charts of 382 con-
secutive patients who had undergone esophagectomy for esophageal
cancer between September 1, 2005, and August 30, 2009 [5] from the
Thoracic Surgery database. There are 6 different types of esophageal
resections performed at our institution; decision for each is based on
tumor location, patient co-morbidities, and surgeon preference. Based
on the database from our thoracic surgery department, the incidence of
leak at our institution is statistically similar at all levels of anastomosis
and all techniques (data not shown). In order to focus on the char-
acteristics and performance of the imaging studies we left the surgery
descriptions out of this paper. We searched for the patients with clinical
suspicion of conduit leak, which then formed the study group. We re-
vised the imaging studies pertinent to the suspected conduit leak and
conducted a longitudinal review of clinical and hospital discharge notes
pertaining to the first 31 days after esophagectomy.
2.1. Definition of conduit leak
Since there is no universal consensus on how to define anastomotic
leak after esophagectomy [14], the presence of conduit leak, for the
purposes of this study was defined and established retrospectively in

⁎
Corresponding author.
E-mail address: dpalacio@radiology.arizona.edu (D. Palacio).
1
Present address: 1515 N Campbell Ave., Tucson, AZ 85724, United States.

https://doi.org/10.1016/j.clinimag.2018.01.010
Received 17 November 2017; Received in revised form 17 January 2018; Accepted 19 January 2018
0899-7071/ © 2018 Elsevier Inc. All rights reserved.
D. Palacio et al. Clinical Imaging 51 (2018) 23–29

Fig. 1. a) Delayed SE view and b) CTELP axial image: Two sequential cone down axial images at the level of the pulmonary arteries. A small well sealed pocket of extravasated oral
contrast on the left side of the mediastinum, is seen in both studies, near the anastomosis, in keeping with a small contained leak (arrows).

Fig. 2. Delayed SE, demonstrating a leak into the right pleural space, being drained by a
chest tube, in keeping with a small contained leak.

Fig. 3. CTELP, cone down axial images at the right lung base (a–c), demonstrating the
collaboration with the Thoracic Surgery Service at our institution, as trajectory of the chest tube (arrows). There is radiodense material within the tube lumen
two or more of the following: positive imaging reports confirmed by (b–c), evidencing a small contained leak. No spillage of oral contrast was identified in the
endoscopic evaluation, surgical findings, longitudinal follow-up data pleural fluid.
consistent with the suspected development of a conduit leak or any
combination. We then classified the patients with a conduit leak ac-
cording to the type of treatment implemented: Patients with type I

24
D. Palacio et al. Clinical Imaging 51 (2018) 23–29

Fig. 4. Delayed SE. (a, b). Orthogonal views demonstrating large extravasation of oral contrast on the right side of the mediastinum (thin arrows). Note moderate distention of the gastric
conduit (short arrows) associated with some degree of stenosis distally within the pyloric region (arrowhead in a).

conduit leak required no specific treatment; patients with type II con-
duit leak received conservative treatment (such as antibiotics, oral diet
withholding, or supplemental nutrition) or had undergone minimally
invasive procedures, such as esophageal stent, mediastinal drain, or
chest tube placement; patients with type III conduit leak underwent
surgical management; patients with type IV conduit leak had surgical
management and conduit necrosis.
Describing leak with some level of agreement between clinicians has
been historically difficult. A recent multi-national Delphi process con-
ducted by a group known as ECCG (Esophagectomy Complications
Consensus Group) has published on a version of the categories defined
in this paper. We feel that this is the best method to categorize leaks
currently [15].
2.2. Imaging techniques
2.2.1. CTLP technique
All of the CTLP studies were conducted with use of 64 multidetector
light-speed VCT Scanners (GE Medical Systems Milwaukee, WI). The
CTLP was acquired with the patient in supine position, at 2.5-mm slice
thickness, and with axial images reconstructed in a standard algorithm
before and after administration of 120 cm3 of IV iohexol or iodixanol
(Omnipaque and Visipaque respectively, GE Healthcare, Cork, Ireland)
and oral iohexol (Omnipaque, GE Healthcare) at 15 cm3 diluted in
100 cm3 of water given immediate before the scan or on the table.
Coronal and sagittal reformatted images were available for interpreta-
tion.
2.2.2. SE technique
Under fluoroscopic examination, patients initially swallowed un-
diluted iohexol (Omnipaque, GE Healthcare). If no conduit leak was
identified, the patient was given liquefied full-strength barium. Antero-
posterior and bilateral anterior oblique fluoroscopic cine images in the
full upright position to various degrees of inclination depending on
patient tolerance were obtained. Bilateral posterior oblique views as
well as postero-anterior views, in the prone position complemented the
study, depending on the patient's condition.
2.3. Conduit leak classification by imaging
Based on the CTLP and SE reports, patients with positive results
were subdivided into two groups: those with small or contained conduit
leaks and those with large or uncontained conduit leaks. Small or
contained conduit leaks (Figs. 1–3) had a blind ending track or a sealed-
off collection of oral contrast within the mediastinum with or without
drainage into the pleural space but with evidence of appropriate drai-
nage by a chest tube in close proximity to the fluid collection. Large or
uncontained conduit leaks had large, diffuse, or free spillage of oral
contrast in the mediastinum and/or pleural space, with no adequate
direct drainage by a chest tube (Figs. 4,5).
2.4. Imaging data collection
We reviewed the reports of all diagnostic CTLP and SE studies
performed within the first month after esophagectomy. A diagnostic
CTLP or SE was considered the initial study performed to investigate a
clinical suspicion of conduit leak that was not previously diagnosed.
Therefore, CTLP and SE studies conducted for conduit leak surveillance
were excluded from the analysis, as were conventional CT studies re-
quested for other purposes during the immediate postoperative period.
The overall individual CTLP and SE performance for diagnosis of
conduit leak was correlated with positive imaging reports, endoscopic
evaluation, surgical findings, or longitudinal follow-up data. We also
compared performance of CTLP and SE in patients who had both studies

25
D. Palacio et al. Clinical Imaging 51 (2018) 23–29

Table 1
Patient distribution according to the clinical management of the conduit
leak.

Leak severity grade No. (%) of patients

I 7 (16.6)
II 9 (21.4)
III 21 (50)
IV 5 (11.9)
Total 42 (100)

Table 2
Distribution of patients diagnosed with conduit leak by imaging and type of leak.

Leak Diagnosis Diagnosis Diagnosis Diagnosis Diagnosis Total

type per by CTELP by SE (no by both by CTELP by SE but
Imaging (no SE CTELP CTELP but not SE not CTELP
performed) performed) and SE

Small 5 3 7 3 7 25
Large 6 1 4 11
Total 11 3 8 3 11 36

Table 3
Correlation of conduit leak type per imaging and management classification of the leak.

Conduit leak diagnosis Leak type per imaging Total

Small/contained Large/uncontained

I–II 13 0 13
III 10 8 18
IV 2 3 5
Total 25 11 36

Fig. 5. CTELP axial images at the level of the main pulmonary arteries (a) and at the level
of the inferior pulmonary veins (b showing a large uncontained leak (arrows). Note high
density of the pleural fluid (*), explained by spillage of oral contrast into the pleural
space.
performed within the first month after esophagectomy. Recognizing
that a time bias may be introduced in doing so, we decided to extract a
subgroup of patients who underwent both CTLP and SE within 3 days of
one another and similarly analyzed performance for both diagnostic
modalities. A cross-match was obtained between the clinical classifi-
cation of the conduit leak according to severity and the imaging conduit
leak characteristics independent of whether the diagnosis was made
with CTLP, SE, or a combination of both.
We performed a Cox regression analysis of conduit leak-free prob-
ability as a function of time and investigated how these findings could
affect the diagnostic value of the CTLP studies. Information regarding
alternative diagnoses offered by CTLP studies was also recorded.
3. Results
Between September 1, 2005, and August 31, 2009, a total of 382
patients underwent esophagectomy for esophageal cancer. The proce-
dures were performed by the thoracic surgery team members, with a
minimum of 10 years of experience. The records of 125 patients, in-
dicated symptoms and/or signs concerning for esophageal leak during
the first month after surgery. At the cut-off time of 31 days post-pro-
cedure, 42 patients were diagnosed with true leaks 42/125 following
the previously stated criteria. The distribution of patients according to
type of conduit leak is shown in Table 1. The diagnosis of conduit leak
after surgery was made at a median of 11 days (range, 6–21 days).
Of the 125 patients demonstrating concern for a potential leak, four
patients did not have any pertinent imaging study and three of them
were finally classified as having grade III conduit leaks whereas one was
classified as a grade II conduit leak.
The remaining 121 patients were imaged to assess for conduit leak:
17 underwent CTLP, 36 had SE, and 68 underwent both studies within
the first month after esophagectomy.
3.1. Clinical information
The 125 patients who were suspected to have a conduit leak pre-
sented with a variety of somewhat specific signs and symptoms, such as
change in character or appearance of drain output, tachycardia, atrial
fibrillation but also more generic findings, such as chest pain, fever, and
signs of sepsis including unexplained confusion, fatigue, malaise,
nausea/vomiting or any combination, without a clear cause.
3.2. Conduit leak type by imaging
On imaging, 36/121 (29.7%) patients were positive for a conduit
leak and 2 patients with proven conduit leaks, one by endoscopy and
another at surgery had false negative CTLP and SE studies. Both pa-
tients were status post Ivor Lewis esophagectomy. One of the patients
demonstrated anastomotic dehiscence at subsequent surgery measuring
about 1.2 × 0.8 cm. The second patient showed a 5 mm defect at en-
doscopy, approximately 1 cm distal to the anastomosis. Independent of
the imaging method used, 69.4% (n = 25) were small/contained con-
duit leaks and 30.5% (n = 11) were large/uncontained conduit leaks.
Table 2 displays the distribution of patients by diagnostic modality. Of
the patients with a small/contained conduit leak, 12/25 (48%) required
surgical intervention and all 11 patients with a large/uncontained
conduit leak required surgical intervention. The correlation of conduit
leak type per imaging and management classification are shown in
Table 3.

26
D. Palacio et al. Clinical Imaging 51 (2018) 23–29

Table 4 Table 6
Total of patients who underwent SE only. Comparison of report results with the final Patients who underwent CTELP and SE within the first month after esophagectomy.
diagnosis of conduit leak. Report result correlation with the final diagnosis of leak and management.

SE leak report Conduit leak diagnosis Total Conduit leak Conduit leak diagnosis Total

No Yes No Yes

No 31 0 31 CTELP leak report No 43 13 56

Yes 2 3 5 Yes 3 9 12
Total 33 3 36 Total 46 22 68
SE report No 45 4 49
Yes 1 18 19
Total 46 22 68
Table 5
Total of patients who underwent CTELP only. Result report correlation with the final
diagnosis of conduit leak.
Table 7
CTELP leak report Conduit leak diagnosis Total
Total of patients who underwent CTELP and SE within 3 days from one another in the first
month after esophagectomy. Correlation with the final diagnosis of leak and management
No Yes
classification.
No 4 1 5
Conduit leak report Conduit leak diagnosis Total
Yes 1 11 12
Total 17
No Yes

CTELP No 19 3 22
Yes 1 7 8
3.3. SE report analysis Total 20 10 30
SE No 19 2 21
Yes 1 8 9
Of the 125 patients, 36 (28.8%) with a clinical suspicion of conduit
Total 20 10 30
leak underwent SE. The median day for SE was postoperative day 9
(range, 4–27 days). The SE report distribution compared with that of
the management classification is shown in Table 4. The performance accuracy of 86.6%. SE studies performance resulted in a sensitivity of
screening of SE for this group resulted in a sensitivity of 84%, specificity 80%, specificity of 95%, PPV of 88.8%, NPV of 90.4%, and accuracy of
of 97.4%, positive predictive value (PPV) of 87.5%, negative predictive 90%. There was concordance between CTLP and SE in 76.7% of the 30
value (VPN) of 96.2%, and accuracy of 93.2%. cases (p = 0.032).

3.4. CTLP report analysis

3.6. CTLP accuracy as a function of time

Of the 125 patients, 17 (13.6%) with a clinical suspicion of conduit

The Cox regression test depicted the cumulative conduit leak free
leak underwent CTLP. CTLP studies were performed at a median of
probability as a function of time performance of the CTELP studies
9 days postoperatively (range, 1–31 days). The patient distribution ac-
(measured in weeks from the date of the esophagectomy). The results
cording to the examination reports and correlation with the manage-
showed that when the CTELP reports within the first 14 days are ne-
ment classification are listed in Table 5. The screening values in this
gative, the reliability of these results is low; in other words, a sub-
group for CTLP resulted in a sensitivity of 58.8%, specificity of 87%,
stantial number of patients were proven to have a conduit leak in spite
PPV of 83.3%, NPV of 87%, and accuracy of 78.8%.
of the negative results, whereas the negative results delivered after
14 days were more likely to represent true absence of conduit leak
3.5. CTLP and SE report performance in patients who underwent both
(Table 8).
studies for clinical suspicion of conduit leak

A total of 68 patients underwent both CTLP and SE during the first
month after surgery. The median timeframe between studies was 4 days
(range, 0–23 days). A total of 43 patients (63.2%) underwent CTLP
before SE with a median timeframe of 6 days (range, 1–23 days); 19
patients (27.9%) underwent CTLP after SE, with a median timeframe of
2 days (range, 1–12 days); and 6 patients (8.8%) underwent the studies
within the same day. The CTLP and SE report result distribution for this
patient group is shown in Table 6. In this group, the performance values
for CTLP resulted in a sensitivity of 40.95%, specificity of 93.4%, PPV
of 75%, NVP of 76.7%, and accuracy of 76.4%. Performance values for
SE revealed a sensitivity of 81.8%, specificity of 97.8%, PPV of 94.7%,
NVP of 91.8%, and accuracy of 92.6%. Recognizing that there would be
a significant time bias regarding results comparing CTLP and SE that
are separated by many days, we elected to evaluate a subgroup of 30
patients (100%) who underwent both CTLP and SE within 3 days of one
another: Nine patients had CTLP before SE, 15 had CTLP after SE, and 6
had both studies within the same day. The distribution of patients
compared with the management classification for CTLP and SE is dis-
played in Table 7. In this group, CTLP performance resulted in a sen-
sitivity of 70%, specificity of 95%, PPV of 87.5%, NPV of 86.3%, and
3.7. Other findings on CTLP
Eight patients were reported to have complex pleural effusions, four
of whom required tube drainage and one who underwent decortication;
the remaining three had chylothorax, one requiring open drainage, a
second requiring a redo thoracotomy for thoracic duct ligation, and a
third requiring pleurodesis. Four other patients were diagnosed as
having multifocal pneumonia. One additional patient had pulmonary
embolism, another had an omental infarction, and one had an infected
mediastinal seroma.
4. Discussion
In clinical practice, the occurrence of “leak” after esophagectomy,
can take several forms, including anastomotic dehiscense, focal defects
distal to the anastomosis and conduit necrosis. On imaging, the loca-
lization of a leak is challenging and often, this differentiation can be
difficult to discern even on endoscopy. Since management does not per
se depend as much on location and final diagnosis of anastomotic
versus conduit leak is unobtainable data in many situations, we elected

27
D. Palacio et al.
Table 8
Leak free probability by CTELP report results: CTELP+ (positive for conduit leak) and CTELP− (negative for conduit leak). (a) studies performed < 7 days (b) between 7 and 14 days and
(c) later than 14 days after esophagectomy. After 7 days following esophagectomy, the hazard of leak was lower when the reports were negative than when they were positive (a,b).
to report “conduit leaks” rather than any specific type of leak. Our study
was specifically focused on the SE and CTELP imaging performance, in
the setting of clinical suspicion for a conduit leak.
The results showed that the absolute values in the screening analysis
for SE reports were slightly better than were results for CTLP, and this
was statistically significant (p < 0.04). In the group of 68 patients in
whom both CTLP and SE studies were performed, we noted a significant
number of false-negative results for the CTLP studies (13 of 68). This
analysis did not consider the exact time at which the diagnosis of
conduit leak was made. Because of that, this large number of false-
negative results is probably not inherent to limitations of the diagnostic
modality itself, but simply related to the dynamic nature of conduit
leaks. Many of the studies may have been performed when the conduit
leak was too early to become evident, as patients may be symptomatic
secondary to an area of infection or early ischemia that has not com-
pletely opened up as a leak. In fact, in up to 64% of patients, the CTLP
studies were performed before the SE examinations were conducted,
with a median timeframe of 6 days (range, 1–23 days) after surgery.
The CTLP values significantly improved in the subgroup of 30 patients
who had both the CTLP and SE studies performed within a short period
of time of one another, which we arbitrarily selected to be 3 days.
Identifying a conduit leak after esophagectomy can be difficult
when using any imaging modality, taking into account that the CT signs
of a conduit leak differ from the typical appearance of a perforated
esophagus. In this regard, a small amount of fluid and/or air in the
mediastinum is suggestive of an esophageal perforation but can be a
normal finding in the early post-operative period after esophagectomy
[16,17]. Accordingly, it can be difficult to differentiate between an
early sign of conduit leak and a result of surgical resection. In a recent
study by Lantos [18], 40% of conduit leaks on CT diagnosed on the
28
Clinical Imaging 51 (2018) 23–29
basis of extraluminal air, fluid, and/or contrast material in the med-
iastinum were falsely positive and a result of surgical resection. Our
results showed were 4/85 (4.7%) false positive conduit leaks diagnosed
on CTLP. In a study by Strauss et al. [19], in which conduit leaks were
suspected on CT performed between 3 and 7 days after esophagectomy
on the basis of mediastinal air alone, 15 of 16 patients had no conduit
leak by endoscopy and/or surgery. To complicate matters, the post-
operative finding of oral contrast within chest tube drains, an obvious
sign of a conduit leak on CTLP, is easily overlooked in patients with a
conduit leak. The addition of prone images to the CTLP studies could
potentially improve sensitivity. There were 14/85 (16.4%) patients
with a false negative CTLP. Our study, similar to that of Strauss et al.
[19], shows that when a conduit leak is suspected clinically in the first
two weeks after esophagectomy, a negative CT is not reliable in ex-
cluding a conduit leak and further imaging and/or endoscopy or even
surgical intervention may be considered depending on the level of
concern. Furthermore, serial surveillance imaging may be a good al-
ternative in relatively stable patients and may increase the diagnostic
value of the imaging modality selected. As stated earlier, given the
design of our study it was not possible to address the possibility of silent
leaks, since we excluded all patients that were asymptomatic. However,
in accordance to previous studies, predictably these would typically be
leaks without significant clinical consequence [13,20]. Therefore rou-
tine imaging studies in all patients at some point after esophagectomy
may not be necessary or best practice. Several complications unrelated
to conduit leak were diagnosed with CTLP and is potentially a benefit
over SE in the evaluation of patients who have undergone an esopha-
gectomy.
Although our study shows that SE is reliable for the detection of
conduit leaks, the postoperative appearance of the conduit poses
D. Palacio et al.
interpretative difficulties. The irregular anatomy of the anastomotic site
accounted for 6 false-positive studies in the SE group and this pitfall for
conduit leak diagnosis has also been described in other studies [18].
This is particularly true when side-to-side anastomosis is performed,
where the appearance is similar to a small-contained conduit leak.
Additionally, redundancy of the anastomosis may be confused with a
small contained conduit leak, especially if asymmetric.
An important observation is that both CTLP and SE were able to
diagnose all of the conduit leaks requiring intervention. The correlation
between small/contained leaks on imaging and the severity of the
conduit leak according to our management classification were often not
congruent. Although general practice suggests conservative treatment
for small/contained intrathoracic conduit leaks [13,17,20], the clinical
status or projected postsurgical therapeutic plan for the malignancy
itself (for example, planned adjuvant radiotherapy) may require a more
aggressive treatment. Even though most of the small/contained conduit
leaks were classified as grade I or II, a substantial number of these
conduit leaks (12/25) actually required interventional treatment (10
grade III conduit leaks and 2 grade IV conduit leaks). As one would
expect, all of the large/uncontained conduit leaks (grade III and IV
leaks) required further management including mediastinal drain or
stent placement and or surgical intervention. This reinforces the con-
cept that the imaging characteristics of a conduit leak do not themselves
dictate treatment. This has been emphasized directly or indirectly in
prior studies [5,10,12,21].
Our retrospective study has the inherent limitations of an un-
controlled patient population: Considering the dynamic nature of con-
duit leaks, studies were performed as clinically indicated leading to a
broad variability in the study timing following surgery, which may have
influenced both SE and CTELP accuracy. Likewise, the order in which
the studies were performed could have influenced the study inter-
pretation to some degree, particularly in the group of patients who had
both studies separated by 3 days or more. In our institution, the SE
studies are interpreted by experienced gastrointestinal radiologists and
the CTELP studies by thoracic radiologists, thereby arguably decreasing
bias. Also, the manner in which the reports were issued, suggested the
findings were interpreted for each study independently and the similar
accuracy of the two tests when done within a few days from each other,
may indicate there is probably not a large co-dependency of outcome
based on order.
We confined our study to the first month after esophagectomy and
evaluation after this period may potentially have resulted in higher
accuracy of the imaging studies in diagnosing a conduit leak due to
resolution of post-operative air and fluid in the mediastinum. However,
the immediate postoperative period is when conduit leaks statistically
occur more often [13,19,20]. Because there is no unifying definition for
conduit leak after esophagectomy [14] and since symptomatic patients
can have normal imaging studies [5] [12], we combined multiple cri-
teria to create a management classification for the presence of a conduit
leak, which included clinical information with or without endoscopy,
imaging, and or surgical supporting findings. This approach has not
been validated in a prospective study. Larger-scale prospective studies
with patients undergoing simultaneous CTELP and SE evaluation would
be ideal to better define the best practice.
In summary, our results indicate that SE is a useful imaging mod-
ality in the evaluation of a clinically suspected conduit leak and is more
sensitive in detecting conduit leaks than CTLP. Although the accuracy
29
Clinical Imaging 51 (2018) 23–29
of CTLP was lower in all of the groups analyzed, CTLP was as efficient
in diagnosing clinically significant conduit leaks and has the ability to
diagnose other co-existing findings related on unrelated to leak.
Because of the limitations of SE and CTLP in the evaluation of suspected
conduit leaks, the clinical status of the patient is a major factor de-
termining further investigation of a suspected conduit leak after eso-
phagectomy.
References
[1] Rizk NP, Bach PB, Schrag D, Bains MS, Turnbull AD, Karpeh M, et al. The impact of
complications on outcomes after resection for esophageal and gastroesophageal
junction carcinoma. J Am Coll Surg 2004;198(1):42–50.
[2] Karl RC, Schreiber R, Boulware D, Baker S, Coppola D. Factors affecting morbidity,
mortality, and survival in patients undergoing Ivor Lewis esophagogastrectomy.
Ann Surg 2000;231(5):635–43.
[3] Larburu Etxaniz S, Gonzales Reyna J, Elorza Orue JL, Asensio Gallego JI, Diez Del
Val I, Eizaguirre Letamendia E, et al. Cervical anastomotic leak after esopha-
gectomy: diagnosis and management. Cir Esp 2013;91(1):31–7.
[4] Sepesi B, Swisher SG, Walsh GL, Correa A, Mehran RJ, Rice D, et al. Omental re-
inforcement of the thoracic esophagogastric anastomosis: an analysis of leak and
reintervention rates in patients undergoing planned and salvage esophagectomy. J
Thorac Cardiovasc Surg 2012;144(5):1146–50.
[5] Alanezi K, Urschel JD. Mortality secondary to esophageal anastomotic leak. Ann
Thorac Cardiovasc Surg 2004;10(2):71–5.
[6] Martin LW, Hofstetter W, Swisher SG, Roth JA. Management of intrathoracic leaks
following esophagectomy. Adv Surg 2006;40:173–90.
[7] Low DE. Diagnosis and management of anastomotic leaks after esophagectomy. J
Gastrointest Surg 2011;15(8):1319–22.
[8] Martin LW, Swisher SG, Hofstetter W, Correa AM, Mehran RJ, Rice DC, et al.
Intrathoracic leaks following esophagectomy are no longer associated with in-
creased mortality. Ann Surg 2005;242(3):392–9. [discussion 9–402].
[9] Gollub MJ, Bains MS. Barium sulfate: a new (old) contrast agent for diagnosis of
postoperative esophageal leaks. Radiology 1997;202(2):360–2.
[10] Swanson JO, Levine MS, Redfern RO, Rubesin SE. Usefulness of high-density barium
for detection of leaks after esophagogastrectomy, total gastrectomy, and total lar-
yngectomy. AJR Am J Roentgenol 2003;181(2):415–20.
[11] Cooke DT, Lin GC, Lau CL, Zhang L, Si MS, Lee J, et al. Analysis of cervical eso-
phagogastric anastomotic leaks after transhiatal esophagectomy: risk factors, pre-
sentation, and detection. Ann Thorac Surg 2009;88(1):177–84. [discussion 84–5].
[12] Sauvanet A, Baltar J, Le Mee J, Belghiti J. Diagnosis and conservative management
of intrathoracic leakage after oesophagectomy. Br J Surg 1998;85(10):1446–9.
[13] Griffin SM, Lamb PJ, Dresner SM, Richardson DL, Hayes N. Diagnosis and man-
agement of a mediastinal leak following radical oesophagectomy. Br J Surg
2001;88(10):1346–51.
[14] Bruce J, Krukowski ZH, Al-Khairy G, Russell EM, Park KG. Systematic review of the
definition and measurement of anastomotic leak after gastrointestinal surgery. Br J
Surg 2001;88(9):1157–68.
[15] Low DE, Alderson D, Cecconello I, Chang AC, Darling GE, D'Journo XB, et al.
International consensus on standardization of data collection for complications
associated with esophagectomy: Esophagectomy Complications Consensus Group
(ECCG). Ann Surg 2015;262(2):286–94.
[16] Backer CL, LoCicero 3rd J, Hartz RS, Donaldson JS, Shields T. Computed tomo-
graphy in patients with esophageal perforation. Chest 1990;98(5):1078–80.
[17] Urschel JD. Esophagogastrostomy anastomotic leaks complicating esophagectomy:
a review. Am J Surg 1995;169(6):634–40.
[18] Lantos JE, Levine MS, Rubesin SE, Lau CT, Torigian DA. Comparison between
esophagography and chest computed tomography for evaluation of leaks after
esophagectomy and gastric pull-through. J Thorac Imaging 2013;28(2):121–8.
[19] Strauss C, Mal F, Perniceni T, Bouzar N, Lenoir S, Gayet B, et al. Computed to-
mography versus water-soluble contrast swallow in the detection of intrathoracic
anastomotic leak complicating esophagogastrectomy (Ivor Lewis): a prospective
study in 97 patients. Ann Surg 2010;251(4):647–51.
[20] Junemann-Ramirez M, Awan MY, Khan ZM, Rahamim JS. Anastomotic leakage
post-esophagogastrectomy for esophageal carcinoma: retrospective analysis of
predictive factors, management and influence on longterm survival in a high vo-
lume centre. Eur J Cardiothorac Surg 2005;27(1):3–7.
[21] Upponi S, Ganeshan A, D'Costa H, Betts M, Maynard N, Bungay H, et al.
Radiological detection of post-oesophagectomy anastomotic leak - a comparison
between multidetector CT and fluoroscopy. Br J Radiol 2008;81(967):545–8.