European Journal of Cardio-thoracic Surgery 33 (2008) 11171123

www.elsevier.com/locate/ejcts

Indications and outcome of salvage surgery for oesophageal cancer

Xavier-Benoit DJourno a, Pierre Michelet b, Laetitia Dahan c, Christophe Doddoli a,d,
Jean-Francois Seitz c, Roger Giudicelli a, Pierre A. Fuentes a, Pascal A. Thomas a,d,*
a

Department of Thoracic Surgery, Ste Marguerite University Hospital, Marseille, France

b
Intensive care Unit, Ste Marguerite University Hospital, Marseille, France
c
Department of Digestive Oncology, La Timone Hospital, Marseille, France
d
UMR 6020, IFR 48, University of the Mediterranean, Marseille, France

Received 30 July 2007; received in revised form 6 January 2008; accepted 16 January 2008; Available online 14 March 2008

Abstract
Objective: Some patients with localised oesophageal cancer are treated with denitive chemoradiotherapy (CRT) rather than surgery. A subset
of these patients experiences local failure, relapse or treatment-related complication without distant metastases, with no other curative
treatment option but salvage oesophagectomy. The aim of this study was to assess the benet/risk ratio of surgery in such context. Methods:
Review of a single institution experience with 24 patients: 18 men and 6 women, with a mean age of 59 years (9). Histology was squamous cell
carcinoma in 18 cases and adenocarcinoma in 6. Initial stages were cIIA (n = 5), cIIB (n = 1) and cIII (n = 18). CRT consisted of 26 sessions of the
association 5-uorouracil/cisplatin concomitantly with a 5075 Gy radiation therapy. Salvage oesophagectomy was considered for the following
reasons: relapse of the disease with conclusive (n = 11) or inconclusive biopsies (n = 7), intractable stenosis (n = 3), and perforation or severe
oesophagitis (n = 3), at a mean delay of 74 days (14240 days) following completion of CRT. Results: All patients underwent a transthoracic enbloc oesophagectomy with 2-eld lymphadenectomy. Thirty-day and 90-day mortality rates were 21% and 25%, respectively. Anastomotic leakage
( p = 0.05), cardiac failure ( p = 0.05), length of stay ( p = 0.03) and the number of packed red blood cells ( p = 0.02) were more frequent in patients
who received more than 55 Gy, leading to a doubled in-hospital mortality when compared to that of patients having received lower doses. A R0
resection was achieved in 21 patients (87.5%). A complete pathological response (ypT0N0) was observed in 3 patients (12.5%). Overall and
disease-free 5-year survival rates were 35% and 21%, respectively. There was no long-term survivor following R1R2 resections. Functional results
were good in more than 80% of the long-term survivors. Conclusion: Salvage surgery is a highly invasive and morbid operation after a volume dose
of radiation exceeding 55 Gy. The indication must be carefully considered, with care taken to avoid incomplete resections. Given that long-term
survival with a fair quality of life can be achieved, such high-risk surgery should be considered in selected patients at an experienced centre.
# 2008 European Association for Cardio-Thoracic Surgery. Published by Elsevier B.V. All rights reserved.
Keywords: Oesophageal neoplasms; Chemotherapy; Radiotherapy; Oesophagectomy

1. Introduction
Ongoing controversy surrounds the question of whether
locally advanced cancer of the oesophagus should be resected
or treated with non-surgical methods. The largest and most
complete meta-analysis of randomised neoadjuvant treatment trials done so far in patients with oesophageal cancer
provides evidence supporting surgery following induction
concurrent chemoradiation therapy (CRT) as the standard of
treatment for t patients with locally advanced oesophageal
cancer, especially in cases of adenocarcinoma [1]. The role of
surgery in the multimodal approach to locoregional oesophageal cancer, however, has recently been questioned. Results of

Presented at the 15th European Conference on General Thoracic Surgery,

Leuven, Belgium, June 36, 2007.
* Corresponding author. Address: Department of Thoracic Surgery, Ste Marguerite Hospital, CHU Sud, 270 Bvd Ste Marguerite, 13274 Marseille Cedex 9,
France. Tel.: +33 491 744 680; fax: +33 491 744 590.
E-mail address: Pascal-alexandre.Thomas@mail.ap-hm.fr (P.A. Thomas).

two randomised trials suggest that in cases of squamous cell

cancer there is no clear survival advantage favouring surgery,
even if local tumour control is signicantly improved after
resection [2,3]. Furthermore, the risks of surgery in this
context reect a signicant effect of CRT on postoperative
mortality within 90 days, due to three main adverse events:
respiratory complications, heart failure, and anastomotic leak
[4]. As a result, the view that completion CRT is an alternative
to surgery in patients with squamous-cell carcinomas who
show a morphological response to induction CRT is growingly
shared by oncologists, because such treatment strategy seems
to produce a similar overall survival, but with less posttreatment morbidity, and last but not least, similar quality of
life [5]. In other words, full-dose CRT (denitive CRT) tends to
be preferred for responders to a half-dose of CRT as much as
oesophagectomy, whereas oesophagectomy is likely to be
preferred for non-responders.
Unfortunately, crude locoregional control rate remains
quite poor with denitive CRT, and roughly half of the

1010-7940/$ see front matter # 2008 European Association for Cardio-Thoracic Surgery. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.ejcts.2008.01.056

1118

X.-B. DJourno et al. / European Journal of Cardio-thoracic Surgery 33 (2008) 11171123

patients present with a persistent or a relapsing tumour at

the primary site within 1 year [6,7]. Accordingly, oesophagectomy stands out as a possible opportunity of cure for t
patients without distant metastases. Besides, local complications of denitive CRT such as intractable strictures, ulcer
or perforation, may lead to a rescue surgery. Finally, the
debate over denitive CRT versus neoadjuvant CRT and
surgery may be reworded in terms of salvage versus planned
oesophagectomy. Although both types of surgery are done in
the setting of previous CRT, one may anticipate that they are
different in several ways. Very few studies have addressed
this issue [812]. Preliminary data suggest that despite an
increased morbidity and mortality, a subset of patients will
be offered a second chance of cure [812]. The selection of
the winners however, remains challenging. The present
report aims to add some information on the topic.

2. Materials and methods

We conducted a retrospective review of all patients
having undergone oesophageal resection (n = 268) between
1996 and 2006 at our institution, and selected those patients
who received salvage surgery (n = 24). Patient charts were
identied by screening of a database into which data were
entered prospectively for any patient undergoing surgery for
thoracic malignancy at our department. Salvage oesophagectomy was dened as an operation performed after
denitive concurrent chemoradiation which included platinum-based chemotherapy and more than or equal to 50 Gy
radiotherapy, and selectively indicated for isolated local
failures and recurrences, or treatment-related complications. In almost all patients, the initial treatment was
planned at an outside centre. Once referred at our
institution, a multidisciplinary decision-making process was
followed. The operation was proposed to patients who were
deemed physiologically amenable to surgery, whose tumour
was thought to be resectable and who had no evidence of
distant metastases at the work-up revaluation.
Hospital records were reviewed for age, sex, body mass
index, initial clinical stage of the disease, American Society of
Anesthesiology risk classication, preoperative medical history, pulmonary function test performances, tumour location,
histology, residual pathologic stage graded according to the
TNM classication [13], and results of preoperative laboratory
and imaging studies (Table 1). All medical charts were also
reviewed for details regarding the initial CRT. There were 6
females and 18 males whose mean age was 59  9 years
(range: 3370). Tumour types included 8 adenocarcinomas
located to the lower oesophagus (classied as Siewert I and II)
and 16 squamous cell carcinomas predominantly located in the
middle (n = 9) and the lower oesophagus (n = 7). At pretreatment evaluation, 18 patients presented with a locally
advanced stage cIIb or cIII disease. Three high-risk patients
presented with a stage cIIA disease. One patient was classied
as having a stage cIVB due to the presence of a single lung
metastasis. In 2 cIIA patients, the justication of the rst-line
CRT was unclear. CRT consisted of the association of 5uorouracil and cisplatin, and concurrent radiotherapy. The
average number of cycles was 2.88 (range: 26). The average
dose of fractionated radiation delivered to the oesophagus was

Table 1
Characteristics of the patients
Variables

Sex, F/M
Adenocarcinoma/squamous cell
Mac Keown/Ivor Lewis

6/18
8/16
9/15

25/75
33/66
37/63

Initial clinical stage

IIA
IIB
III
IVA
IVB

5
1
17
0
1

21
4
70
0
4

AGE
ASA score
NYHA score
Performance status
Body mass index
FEV1 (l)
FVC (l)
FEV1/FVC (%)
Hb (g/dl)

Mean

SD

59
2.4
2.2
1.1
21
2.5
3.4
75
12.5

9
0.6
0.6
0.9
3.7
0.9
0.9
10
2

Mean and standard deviation are presented.

FEV1: forced expiratory volume in 1 s; FVC: forced vital capacity; Hb: haemoglobin.

56 Gy. This value served as cut-off to split the patients

population in 2 groups: 14 patients had received 5055 Gy,
while 10 had received 5675 Gy.
Preoperative disease restaging was based on the results of
barium swallow, whole-body computed tomography (CT) scan
and oesophagoscopy in all patients. Patients with a
supracarinal oesophageal tumour underwent routine beroptic bronchoscopy to rule out any invasion of the
tracheobronchial tree. Nine of the patients received positron
emission tomography (PET) with [18F]-uoro-2-deoxy-Dglucose or integrated CT-PET for initial staging or preoperative restaging. Endoscopic ultrasonography (EUS) was carried
out in 22 patients with no attempt of ne needle aspiration
(FNA), and was not feasible in 2. CT scan ndings provided
some arguments in favour of the presence of an oesophageal
tumour in 16 patients whereas EUS, when available,
displayed in all cases a high suspicion of persistent or
recurrent disease. However, preoperative conrmation of
malignancy was obtained histologically in 12 patients only
(Table 2). Finally, indications for salvage surgery were as
follows: documented or suspected residual or recurrent
disease in 18 patients, and treatment-related local complications in 6: intractable stenosis in 3, perforation in 2, and
radiation-induced oesophagitis in 1.
The average time between salvage surgery and completion of CRT was 74 days (range: 14240 days). Surgical
technique consisted of an en-bloc transthoracic oesophagectomy with two-eld lymphadenectomy in all cases.
According to the location of the tumour, 15 patients received
an Ivor Lewis procedure (intrathoracic anastomosis) and 9 a
Mac Keown operation (cervical anastomosis). In all cases, a
gastric tube reconstruction was performed in the posterior
mediastinum. Intrathoracic anastomoses were performed
with a circular stapler while cervical anastomoses were hand
fashioned. Pyloroplasty and feeding jejunostomy were
performed routinely.

X.-B. DJourno et al. / European Journal of Cardio-thoracic Surgery 33 (2008) 11171123

1119

Table 2
Preoperative work up revaluation

No evidence of malignancy
Features of malignancy

CT scan, n = 24

Oesophagoscopy, n = 24

Histological documentation
(oesophageal biopsy), n = 23

Echoendoscopy, n = 22

8 (33%)
16 (66%)

9 (37%)
15 (63%)

11 negative (48%)
12 positive (52%)

0
22 (100%)

Medical and surgical complications were recorded.

Respiratory complications were dened by all medical events
concerning the lung parenchyma (i.e. pneumonia, airway
congestion, atelectasis, acute lung injury, and acute
respiratory distress syndrome) in the absence of surgical
complications requiring reoperation. Surgical complications
included anastomotic leakage, laryngeal paralysis, chylothorax, pleural effusion, empyema and bleeding. Early
mortality was checked 30 and 90 days after surgery.
All patients were seen at the outpatient clinic at intervals
of three months during the rst two years and every six
months thereafter. Symptoms, body weight and imaging
ndings were routinely recorded. A self-rated scale from 1
(worse results) to 10 (best results) was used to assess the
patients digestive comfort. For patients lost to medical
follow-up, missing survival data were obtained by consulting
the City Hall registry. Statistical analysis included the MannWhitney test, the Pearson x2 test, and Fishers exact test
when appropriate. Overall survival was measured from the
date of operation and survivorship calculated according to
the KaplanMeier method, including the operative mortality.
Disease-free survival was counted up to the date of rst
relapse or death with cancer. Software used included Excel
(Microsoft Corporation, Redmond, Wash), and SPSS (SPSS
Inc., Chicago, Ill).

3. Results
3.1. Pathological ndings
The absence of viable cancer cells was observed on the
operative specimen in 3 patients (12.5%). Three additional
patients (12.5%) had no residual oesophageal tumour but
presented with invaded regional lymph nodes. A lung
metastatic disease was found intraoperatively in three
patients (stage yp IVB) and distant lymph node involvement
was found in two (stage yp IVA). A complete R0 resection was
achieved in 21 patients (87.5%). In all three cases of
incomplete resection, the tumour was located above the
level of the carina (Table 4).
3.2. Mortality and morbidity
Thirty-day and 90-day mortality rates were 21% (n = 5) and
25% (n = 6), respectively. Among the 6 patients who died
within 90 days, 3 were operated on for treatment-related
local complications. There was a high rate of medical
complications (45%), and respiratory events appeared as the
most common morbidity (41%). There was no signicant
difference in early mortality according to the type of surgery:
Thirty-day and 90-day mortality rates were 20% and 26%,
respectively following Ivor Lewis operations, and 22% and

22%, respectively following Mac Keown operations. Mortality

and morbidity were related to the radiation dose (Table 3).
Anastomotic leakage ( p = 0.05) and cardiac failure ( p = 0.05)
were more common in patients who received more than
55 Gy. In turn, median duration of stay in the intensive care
unit (5 days vs 18 days, p = 0.005), length of hospital stay (22
days vs 32 days, p = 0.03) and number of packed red blood
cells (1 unit vs 6 units, p = 0.02) were signicantly higher in
this subset of patients. Thirty-day mortality rates were twice
as high in patients who received more than a 55 Gy radiation
dose when compared to that of patients who received lower
doses, but the difference did not reach statistical signicance. Causes of early death were directly linked to surgery
in two patients (leakage), to respiratory complications in
three, and to cardiac failure in one.
3.3. Survival, recurrence and quality of life
Overall 5-year survival rate was 35%, with 4 patients alive
more than 3 years after the operation and 1 patient alive
more than 5 years after surgery (Fig. 1). Five year diseasefree survival rate was 21%. With a median follow-up of 17
months, 2 of the 18 patients who survived the operation died
from non-cancer-related causes. Eight patients experienced
cancer recurrences: one died from locoregional recurrence
and two from distant metastasis whereas the ve remaining
patients were alive and concurrently treated for locoregional
(n = 1) or distant relapse (n = 4). At last follow-up, eight
Table 3
Complications after salvage oesophagectomy
55 Gy,
n = 14

>56 Gy,
n = 10

Hospital mortality
Thirty-day mortality
Ninety-day mortality

2
2
3

14
14
21

3
3
3

30
30
30

0.61
0.61
0.66

Medical complication
Respiratory complication
Pneumonia
ARDS
Tracheotomy
Cardiac failure

6
4
4
2
2
0

42
28
28
14
14
0

6
6
3
2
5
3

60
60
30
20
50
30

0.68
0.21
1
1
0.08
0.05

Surgical complication
Pleural effusion
Anastomotic leakage
Laryngeal paralysis
Chylothorax

4
1
0
1
2

28
7
0
7
14

6
4
3
1
0

60
40
30
10
0

0.21
0.12
0.05
1
0.49

Length of hospital stay (days)

Length of USI stay (days)
Packed red blood cells (units)

Median

Range

Median

Range

22.5
5
1

174
174
024

32.5
18.5
6

2165
497
019

0.03
0.005
0.02

Statistical analysis included the Mann-Whitney test and Fishers exact test as
appropriate. Median and range are presented.

X.-B. DJourno et al. / European Journal of Cardio-thoracic Surgery 33 (2008) 11171123

1120

were observed in case of complete R0 resections when

compared to that of R1R2 resections (36% vs 0%; p = 0.66)
corresponding to median survival times of 27 months and 11
months, respectively (Table 4).
3.4. Functional assessment

Fig. 1. Overall and disease-free survival curves, including operative mortality

(KaplanMeier method).

We looked specically at the 13 long-term survivors (8 who

were free of disease, and 5 with disease) to assess their quality
of life at last follow-up. Two patients required repeated
endoscopic dilations. Eleven patients (84.6%) had a stable
(variation within 10% of the preoperative value) or improved
(>10%) body weight; whereas 2 patients lost more than 10% of
their body weight. Eleven patients self-rated their digestive
comfort among whom 9 had a score exceeding 5/10 (82%).
There was a clear although not signicant difference between
those patients who were free of disease (n = 7; 7 patients with
a score higher than 5/10) and those who were not (n = 4; 2
patients with a score higher than 5/10).

Table 4
Pathologic ndings on resected specimen

4. Discussion

Variable

Resection R0/R1R2

21/3

87/12

yp Stage
0
I
IIA
IIB
III
IVA
IVB

3
1
10
2
3
2
3

12
4
42
8
12
8
16

patients were still alive and well. At univariate analysis, the

lymph node status did not affect overall survival: median
survival time and 5-year survival rates were 21 months and
32%, versus 27 months and 28% in ypN0 and ypN1 patients,
respectively ( p = 0.43). Accordingly, there was no difference
according to the disease stage when comparing stages yp I
and yp II to stages yp III and yp IV: median survival times and
5-year survival rates were 29 months and 28% vs 27 months
and 34%, respectively ( p = 0.72). Best 5-year survival rates

Our results, combined to those of the available literature

(Table 5), clearly show that salvage oesophagectomy is a
highly morbid operation, providing an early mortality ranging
from 15% to 25% at 3 months. Two types of complications
dominate the spectrum of postoperative adverse events:
anastomotic stulas and pulmonary complications.
The very high incidence of anastomotic leakage, exceeding
basically 25% in almost all series, is likely to be the
consequence of a fragile irradiated stomach and oesophagus
and impaired blood supply. It seemed that the technique of the
anastomosis by itself, stapled or hand-fashioned, did not really
inuence the healing in this setting. Conversely, our team
recently demonstrated that thoracic epidural analgesia
improved the microcirculation of the gastric tube in the early
postoesophagectomy period [14], and was associated with a
decrease in occurrence of anastomotic leakage [15]. In the
present study, the incidence of anastomotic failure was closely
linked to the overall dose of radiation received, with no stula

Table 5
Summary of the literature
Author

Year

Nb
Patients

Histology

Chemotherapy

Radiotherapy
(mean) (Gy)

Delay (range
in months)

Complete pathological response

and R0 resections (%)

Meunier
Swisher
Nakamura
Tomimaru
Oki
Present series

1998
2002
2004
2006
2007
2008

6
13
27
24
14
24

SCC
SCC and ADK
SCC
SCC
SCC
SCC and ADK

5FU5FUNS
5FU5FU5FU-

60
56.7
60
62
64.6
56

317
456
115
125
134
0.58

0NA
077.2
1166.7
45.866.7
050
12.587.5

Platinum
Platinum
Platinum
Platinum
Platinum

Thirty-day
mortality (%)

Ninety-day
mortality (%)

Anastomotic
leak (%)

Respiratory
complications (%)

Five-year
survival (%)

Good functional
results (%)

0
15
3.7
4.2
0
21

16.7
NA
NA
12.5
14.3
25

33.3
38
22.2
20.8
28.6
12.5

33.3
62
11.1
20.8
21.4
41

0
25
30
33
32
35

66.7
NA
NA
NA
100
>80

SCC: squamous cell carcinoma; ADK: adenocarcinoma; Gy: gray; NA: not available.

X.-B. DJourno et al. / European Journal of Cardio-thoracic Surgery 33 (2008) 11171123

below 55 Gy. Even if this information is weakened by the

retrospective nature of the study and the post hoc determination of this cut-off value, it suggests at least that a promising
way to reduce this kind of complication is probably to better
target the tumour and the involved lymph nodes to decrease
the radiation dose administered to surrounding normal tissues.
Modern radiotherapy delivery nowadays relies on tridimensional, conformal techniques. Gold standard imaging modality
remains computed-tomography scanner. However, the intrinsic lack of contrast between soft tissues leads to high
variabilities in target denition. The fusion of the different
imaging modalities, including positron emission tomography
could theoretically achieve this goal [16].
Severe pulmonary complications exceed 3040% in incidence commonly in this setting. Respiratory complications
remain the major concern after oesophagectomy, with or
without previous chemoradiation. Reasons for this pulmonary
morbidity are multifaceted, and those due specically to the
neoadjuvant treatment are probably very difcult to segregate from those due to the surgical procedure, to the
perioperative anaesthetic management, to the patient
himself and to the toxicity of the preoperative treatment.
However, concurrent CRT was shown to be associated with
signicant worsening of the diffusion capacity of the lung for
carbon monoxide (DLCO) [17]. In a recent retrospective study,
dosimetric factors but not clinical factors were found to be
strongly associated with the incidence of postoperative
pulmonary complications. The volume of the lung spared
from doses of 5 Gy and higher was the only independent factor
in multivariate analysis [18]. As hypothesised for anastomotic
complications, this suggests that restraining the radiation
elds thus ensuring an adequate volume of lung unexposed to
radiation might reduce the incidence of postoperative
pulmonary complications.
Given the high risks associated with surgery in this setting,
the question arises of which categories of patients can
benet from such a hazardous operation. Our results suggest
that oesophagectomy should probably be avoided whenever
possible in case of treatment-related local complications
since 90-day mortality reached 50% in this patient group.
Aside from these particular circumstances, one selection
approach would be to avoid operating on patients without a
proven residual or recurrent disease. Unfortunately, diagnosis of complete pathological response by imaging is difcult
and often possible merely by oesophageal resection. The only
easily reproducible modality for determination of response is
endoscopic visualisation with biopsies of suspicious areas.
Obviously, endoscopy alone cannot detect a viable disease
conned to the regional lymph nodes, a frequent event even
in the absence of any residual oesophageal tumour as
demonstrated by the present series. Endoscopy also failed to
provide conclusive tissue biopsies in 7 of the 18 patients (39%)
in whom the analysis of the operative specimen found viable
cells inside the oesophageal wall. The assessment of
locoregional tumour extension by EUS-FNA is thought
currently the most reliable method. FDG PET guided EUS
FNA is advocated in PET-positive nodes, particularly at the
coeliac region [19]. A recent study, however, demonstrated
that a complete absence of PETsignal cannot be equated with
a complete pathological response: the accuracy of the 100%
reduction in maximum standardised uptake value after

1121

neoadjuvant treatment as a predictor of a complete

pathological response was only 15% [20]. When combining
imaging modalities with FDG PET, CT, and EUS it is not
possible to conrm the absence of residual viable disease in
the primary site in 2540% of the cases [21].
In most series, long-term survival reaches roughly 3035%
at 5 years, a non-negligible rate in such a disastrous disease.
Functional aspects of the surgical results are seldom
addressed. Even if our functional evaluation method was
approximate, our data suggest that the quality of oral intake
was fair in more than 80% of the patients. Health-related
quality of life seemed to be predominantly impaired by
progression of the disease. We found that the residual TNM
was not an accurate prognosticator although the small
number of patients precluded a comprehensive analysis of
survival. Basically, patients who may benet most from
surgery are those in whom a complete R0 resection could be
performed. Indeed, R0 resection can serve as an immediate
surrogate for outcome since patients who are left with gross
or microscopic residual tumour will almost always die
promptly from progressive disease, as in our experience.
The link between local recurrence and margin of normal
tissue surrounding a resected cancer is well established. In
contrast with longitudinal clearance which is easily predictable on the basis of both endoscopic inspection with
Lugols stain screening for a multifocal disease, and EUS
examination of the proximal oesophagus looking at submucosal spreading, circumferential clearance is hard to
anticipate. In the absence of serosa, there is no specic
fascial boundary to circumferential spread of oesophageal
cancer. In a prospective study, the nding of a tumour within
1 mm of the circumferential margin of the xed resection
specimen of patients undergoing what would have been
regarded as a potentially curative resection was found as a
highly signicant predictor of both local recurrence and
survival [22]. The role of the surgeon should therefore be to
resect the oesophagus with as wide a margin of uninterrupted
normal tissue as possible around it. This goal is amenable in
most cancers located below the level of the carina, even in
cases of bulky tumours. In contrast, the upper and middle
oesophagus is surrounded closely by vital structures that
cannot be resected en-bloc. EUS does not add to the
estimation of locoregional respectability after RCT because
of disorganising brotic sequelae at the level of the tumour
and its surroundings [23]. We set up our decision to operate or
not on the basis of CT scan and barium swallow ndings
mainly, with a high suspicion index of unresectability in cases
of lumen deviation, tumour height >5 cm, aortic contact
>908, loss of the fat plane between tumour and neighbouring
organ, or tumour indenting neighbouring organ at CT scan, as
thoroughly described after neoadjuvant CRT by Piessen et al.
[24]. As a result, our 87.5% R0 resection rate compares
favourably with those of the literature.
Nevertheless, our rm belief is that the indication for
salvage surgery should be limited to patients with an initially
resectable tumour. Current treatment strategies for locally
advanced cancers currently favour neoadjuvant chemotherapy or chemoradiotherapy followed by surgery for adenocarcinomas, but chemoradiotherapy alone in patients with
SCC who have shown a morphological response after
induction treatment [25]. We want to add a word of caution

1122

X.-B. DJourno et al. / European Journal of Cardio-thoracic Surgery 33 (2008) 11171123

concerning salvage oesophagectomy that should not be

regarded as a routine rescue procedure in case of failure of
denitive CRT. In turn, indications for non-surgical treatment
strategies should be decided on solid grounds, and reserved
to those patients thoroughly investigated with EUS FNA and
integrated PET-scan, and presenting with a supracarinal
oesophageal tumour deemed consensually to be nonresectable. We, as oesophageal surgeons inside a multidisciplinary team, should be pivotal partners of the primary
decision to keep surgery or not in the treatment plan, as we
are when a salvage surgery is evoked.
In conclusion, this study conrmed increased morbidity
and mortality after salvage oesophagectomy performed after
denitive chemoradiation therapy. The increased risks
seemed to be predominantly related to radiotherapy delivery
modalities, and the management of local treatment-related
complications is indubitably the worse situation in which such
risky operation may be performed. Nevertheless, some
patients were cured, and long-term survival appeared to
be associated primarily with R0 resection. These data suggest
that salvage oesophagectomy is an elective therapeutic
option for carefully selected patients at experienced referral
centres.

References
[1] Gebski V, Burmeister B, Smithers BM, Foo K, Zalcberg J, Simes J,
Australasian Gastro-Intestinal Trials Group. Survival benets from neoadjuvant chemoradiotherapy or chemotherapy in oesophageal carcinoma: a
meta-analysis. Lancet Oncol 2007;8:22634.
[2] Stahl M, Stuschke M, Lehmann N, Meyer HJ, Walz MK, Seeber S, Klump B,
Budach W, Teichmann R, Schmitt M, Schmitt G, Franke C, Wilke H.
Chemoradiation with and without surgery in patients with locally
advanced squamous cell carcinoma of the esophagus. J Clin Oncol
2005;23:23107.
[3] Bedenne L, Michel P, Bouche O, Milan C, Mariette C, Conroy T, Pezet D,
Roullet B, Seitz JF, Herr JP, Paillot B, Arveux P, Bonnetain F, Binquet C.
Chemoradiation followed by surgery compared with chemoradiation
alone in squamous cancer of the esophagus: FFCD 9102. J Clin Oncol
2007;25:11608.
[4] Fiorica F, Di Bona D, Schepis F, Licata A, Shahied L, Venturi A, Falchi
AM, Craxi A, Camma C. Preoperative chemoradiotherapy for oesophageal cancer: a systematic review and meta-analysis. Gut 2004;53:
92530.
[5] Bonnetain F, Bouche O, Michel P, Mariette C, Conroy T, Pezet D, Roullet B,
Seitz JF, Paillot B, Arveux P, Milan C, Bedenne L. A comparative longitudinal quality of life study using the Spitzer quality of life index in a
randomized multicenter phase III trial (FFCD 9102): chemoradiation
followed by surgery compared with chemoradiation alone in locally
advanced squamous resectable thoracic esophageal cancer. Ann Oncol
2006;17:82734.
[6] Herskovic A, Martz K, al-Sarraf M, Leichman L, Brindle J, Vaitkevicius V,
Cooper J, Byhardt R, Davis L, Emami B. Combined chemotherapy and
radiotherapy compared with radiotherapy alone in patients with cancer
of the esophagus. N Engl J Med 1992;326:15938.
[7] Cooper JS, Guo MD, Herskovic A, Macdonald JS, Martenson Jr JA, AlSarraf M, Byhardt R, Russell AH, Beitler JJ, Spencer S, Asbell SO,
Graham MV, Leichman LL. Chemoradiotherapy of locally advanced
esophageal cancer: long-term follow-up of a prospective randomized
trial (RTOG 85-01). Radiation Therapy Oncology Group. JAMA 1999;
281:16237.
[8] Meunier B, Raoul J, Le Prise E, Lakehal M, Launois B. Salvage esophagectomy after unsuccessful curative chemoradiotherapy for squamous
cell cancer of the esophagus. Dig Surg 1998;15:2246.
[9] Swisher SG, Wynn P, Putnam JB, Mosheim MB, Correa AM, Komaki RR,
Ajani JA, Smythe WR, Vaporciyan AA, Roth JA, Walsh GL. Salvage esophagectomy for recurrent tumors after denitive chemotherapy and
radiotherapy. J Thorac Cardiovasc Surg 2002;123:17583.

[10] Nakamura T, Hayashi K, Ota M, Eguchi R, Ide H, Takasaki K, Mitsuhashi N.

Salvage esophagectomy after denitive chemotherapy and radiotherapy
for advanced esophageal cancer. Am J Surg 2004;188:2616.
[11] Tomimaru Y, Yano M, Takachi K, Miyashiro I, Ishihara R, Nishiyama K,
Sasaki Y, Ishikawa O, Doki Y, Imaoka S. Factors affecting the prognosis of
patients with esophageal cancer undergoing salvage surgery after denitive chemoradiotherapy. J Surg Oncol 2006;93:4228.
[12] Oki E, Morita M, Kakeji Y, Ikebe M, Sadanaga N, Egasira A, Nishida K, Koga
T, Ohata M, Honboh T, Yamamoto M, Baba H, Maehara Y. Salvage esophagectomy after denitive chemoradiotherapy for esophageal cancer.
Dis Esophagus 2007;20:3014.
[13] Sobin LH, Wittekid C, editors. TNM Classication of Malignant Tumors.
International Union Against Cancer. 5th ed., New York, NY: Wiley;
1997.
[14] Michelet P, Roch A, DJourno XB, Blayac D, Barrau K, Papazian L, Thomas
P, Auffray JP. Effect of thoracic epidural analgesia on gastric blood ow
after oesophagectomy. Acta Anaesthesiol Scand 2007;51:58794.
[15] Michelet P, DJourno XB, Roch A, Papazian L, Ragni J, Thomas P, Auffray
JP. Perioperative risk factors for anastomotic leakage after esophagectomy: inuence of thoracic epidural analgesia. Chest 2005;128:
34616.
[16] Moureau-Zabotto L, Touboul E, Lerouge D, Deniaud-Alexandre E, Grahek
D, Foulquier JN, Petegnief Y, Gres B, El Balaa H, Kerrou K, Montravers F,
Keraudy K, Tiret E, Gendre JP, Grange JD, Houry S, Talbot JN. Impact of CT
and 18F-deoxyglucose positron emission tomography image fusion for
conformal radiotherapy in esophageal carcinoma. Int J Radiat Oncol Biol
Phys 2005;63:3405.
[17] Abou-Jawde RM, Mekhail T, Adelstein DJ, Rybicki LA, Mazzone PJ, Caroll
MA, Rice TW. Impact of induction concurrent chemoradiotherapy on
pulmonary function and postoperative acute respiratory complications
in esophageal cancer. Chest 2005;128:2505.
[18] Wang SL, Liao Z, Vaporciyan AA, Tucker SL, Liu H, Wei X, Swisher S, Ajani
JA, Cox JD, Komaki R. Investigation of clinical and dosimetric factors
associated with postoperative pulmonary complications in esophageal
cancer patients treated with concurrent chemoradiotherapy followed by
surgery. Int J Radiat Oncol Biol Phys 2006;64:6929.
[19] Plukker JT, van Westreenen HL. Staging in oesophageal cancer. Best Pract
Res Clin Gastroenterol 2006;20:87791.
[20] Port JL, Lee PC, Korst RJ, Liss Y, Meherally D, Christos P, Mazumdar M,
Altorki NK. Positron emission tomographic scanning predicts survival after
induction chemotherapy for esophageal carcinoma. Ann Thorac Surg
2007;84:393400.
[21] Swisher SG, Maish M, Erasmus JJ, Correa AM, Ajani JA, Bresalier R, Komaki
R, Macapinlac H, Munden RF, Putnam JB, Rice D, Smythe WR, Vaporciyan
AA, Walsh GL, Wu TT, Roth JA. Utility of PET, CT, and EUS to identify
pathologic responders in esophageal cancer. Ann Thorac Surg 2004;
78:115260.
[22] Dexter SP, Sue-Ling H, McMahon MJ, Quirke P, Mapstone N, Martin IG.
Circumferential resection margin involvement: an independent predictor
of survival following surgery for oesophageal cancer. Gut 2001;48:
66770.
[23] Laterza E, de Manzoni G, Guglielmi A, Rodella L, Tedesco P, Cordiano C.
Endoscopic ultrasonography in the staging of esophageal carcinoma after
preoperative radiotherapy and chemotherapy. Ann Thorac Surg 1999;
67:14669.
[24] Piessen G, Briez N, Triboulet JP, Mariette C. Patients with locally
advanced esophageal carcinoma nonresponder to radiochemotherapy:
who will benet from surgery? Ann Surg Oncol 2007;14:203644.
[25] Mariette C, Piessen G, Triboulet JP. Therapeutic strategies in oesophageal
carcinoma: role of surgery and other modalities. Lancet Oncol 2007;8:
54553.

Appendix A. Conference discussion

Dr R. Berrisford (Exeter, UK): I would just like to ask the audience to put
your hand up when you consider a patient for oesophagectomy you use this kind
of concept of salvage oesophagectomy (very few members of the audience
raised their hand). That is reassuring because in our MDTs we dont usually use
that concept. I was very interested to see that in your denition of these you
include patients who had chemotherapy who are node positive. We include
patients who are node positive after chemotherapy quite routinely but we
dont think that they are salvage patients, maybe we should do. So what do you
think your denition of salvage oesophagectomy should really include?

X.-B. DJourno et al. / European Journal of Cardio-thoracic Surgery 33 (2008) 11171123

Dr DJourno: Thank you very much for your question. It is a problem of
denition. Firstly it is important to note that all these patients were initially
treated at an outside institution so we didnt participate in the initial treatment
strategies. We included patients with a persistence or relapse of tumour after
denitive chemoradiotherapy. We believe that in this very selected subgroups of
patients, surgery provides the unique alternative option to rescue them from a
fatal issue. In fact there is no other possibility of treatment such as a palliative
chemotherapy. The unique curative option is just surgery.
The main result of our study is probably that morbidity and mortality were
related to volume of radiation. So when you propose a patient for a salvage
oesophagectomy, maybe you have to look on the volume of radiation. For a
volume dose of radiation up to 50 Gy, the operative risk is probably prohibitive.
Dr J. Duffy (Nottingham, UK): Just looking at the group you operated on, 2
of them with an oesophageal perforation. Did they survive?

1123

Dr DJourno: No.
Dr Duffy: So you are quite hard on yourself in your results. I am sure many
other series would have excluded the patients with oesophageal perforation.
The second question is, why did these patients have chemoradiotherapy as
opposed to surgery in the rst place and why wasnt surgery part of that plan of
treatment?
Dr DJourno: I dont know because the treatment strategy was given at
another institution. We didnt participate in the initial discussion. The patients
were referred to our hospital, maybe 3 or 6 months after completion of the
denitive chemoradiotherapy.
Dr Duffy: In your 5-year survival are you including surgical mortality?
Dr DJourno: Yes, we included the operative mortality. But it is difcult to
draw some conclusions on long-term survival because its a very small series.