SURROGATE MARKERS OF RESECTABILITY IN PATIENTS UNDERGOING EXPLORATION OF POTENTIALLY RESECTABLE PANC

1
SURROGATE MARKERS OF RESECTABILITY IN PATIENTS
UNDERGOING EXPLORATION OF POTENTIALLY RESECTABLE
PANCREATIC ADENOCARCINOMA
Ong SL, Garcea G, Thomasset SC, Mann CD, Neal CP,
Abu Amara M, Dennison AR, Berry DP
Department of Hepatobiliary and Pancreatic
Surgery, The Leicester General Hospital,
Leicester, UK
BACKGROUND
RESULTS
Statistical Analysis Students t-test, Fishers
exact test and Chi-squared test were applied to
compare the groups on all factors. Receiver
Operator Characteristics (ROC) curve and
logistical regression were utilised for further
analysis of factors that showed a significant
difference between the groups. All statistical
analyses were carried out using statistical
software MedCalcTM version 9.3.0.0. A value of
Plt0.05 was considered statistically significant.
Adenocarcinoma of the head of the pancreas has a
poor prognosis. Surgery remains the only curative
treatment for this devastating disease.
Unfortunately, the majority of patients will have
metastatic or locally advanced disease at
presentation, with only 20 of cases eligible for
resection at the time of diagnosis. Furthermore,
pancreatico-duodenectomy is associated with 3-5
peri-operative mortality and significant
morbidity. The overall 5-year survival after
resection is 15-20, although this may approach
40-50 in patients with favourable prognostic
factors. Despite extensive pre-operative
staging, a significant proportion of pancreatic
cancers are found to be unresectable at surgical
exploration. A thorough selection process is,
therefore, critical in ensuring that only
patients with potentially curable disease proceed
to resection. This study aimed to identify
surrogate markers of resectability in patients
undergoing surgical exploration of the pancreas
with curative intent.
The bypass group had a significantly higher
CA19.9 level, lower platelet count, higher
neutrophillymphocyte ratio and higher
CA19.9bilirubin ratio with P values of 0.003,
0.013, 0.026 and 0.022 respectively. Other
serological studies, including full blood count,
liver and renal function tests were not notably
different between the two groups. There was no
difference in the proportion of patients
requiring biliary stenting to relieve jaundice
pre-operatively (P0.221). Tumour
characteristics between the two groups were
compared. The bypass group was noted to have a
significantly larger tumour size than the
resection group when assessed intra-operatively
by ultrasonography during staging laparoscopy
(median diameter of 40mm vs. 30mm, P0.047).
However, there was no significant difference
between the groups in the tumour size recorded on
staging CT scan. The ROC curve was utilised to
analyse the value of neutrophillymphocyte ratio
(gt1.93), CA19.9bilirubin ratio (gt71.18), age
(gt65), platelet count (297x109/L) and CA19.9
(gt473Ku/L) in predicting which patients were not
suitable for curative resection. The areas under
the ROC curves plotted were 0.60, 0.60, 0.62,
0.62 and 0.69 respectively (Table2). NLR and
CA19.9bilirubin ratio had high specificity for
unresectability (92.9 and 97.0 respectively)
while CA19.9 on its own was the most sensitive
predictor of unresectability (81.8). When the
above factors were analysed with logistic
regression a raised CA19.9 was identified as an
independent factor for predicting unresectable
disease (P0.031).
RESULTS
Between 1999 and 2006, Leicester General Hospital
received 1202 new pancreatic cancer referrals.
113 of these patients underwent pancreatic
exploration with curative intent, comprising only
9 of the referrals. Of this number only 5
proceeded to a resection with curative
intent. The resectable group included the 55
patients who underwent a pancreaticoduodenectomy
and the bypass group included the remaining 58
patients who underwent a triple bypass. There was
no significant difference in the time from
diagnosis to laparotomy between the two groups.
There were also no significant differences noted
between patient demographics, including gender,
body mass index and socioeconomic status (Table
1). The bypass group was found to be
significantly older than the resection group
(P0.049).
METHODS
Patients This study examined all patients with
adenocarcinoma of the head of the pancreas who
were selected for pancreaticoduodenectomy from
1999 to 2006 at the Department of Hepatobiliary
and Pancreatic Surgery, Leicester General
Hospital. Within this population of patients,
the study compared the resectable group of
patients (who were selected for curative surgery
and underwent a pancreatico-duodenectomy) with
the bypass group of patients (who were selected
for curative surgery but were found to have
extra-pancreatic tumour spread intra-operatively).
The latter group of patients proceeded to have a
triple bypass in the form of hepaticojejunostomy-e
n-Y and gastrojejunostomy. Each patient
underwent a rigorous staging process to determine
tumour resectability prior to being selected for
potentially curative surgery. Patients with
resectable disease on CT imaging underwent
staging laparoscopy and intra-operative
ultrasound scan to identify extra-pancreatic
disease below the resolution level of
conventional CT imaging. All suspicious
peritoneal or hepatic deposits were biopsied.
Patients with tumour size favourable for
resection (diameter lt4.0 cm) and no evidence of
extra-pancreatic spread of disease were
subsequently consented for pancreatic exploration
with a view to proceed to pancreaticoduodenectomy.
Comparison factors Comparisons between the two
groups were made in relation to the patients
demographics, socioeconomic status, radiological
staging and serology results. Tumour size and
suggestion of lymph node involvement on
pre-operative CT imaging were recorded.
Serological data included full blood count, urea
and electrolytes, liver function tests and CA19.9
level. Time from diagnosis to surgery was also
noted.
Table 1
Comparison between the resection group and the
bypass group
ROC curve analysis for pre-operative markers
Table 2
CONCLUSIONS
Pre-operative serology including platelet count,
neutrophillymphocyte ratio, CA19.9bilirubin
ratio and in particular CA19.9 may be used as
additional predictors of resectability in
combination with conventional staging. In
collaboration with other markers of inflammatory
response, such as CRP, it is suggested that these
factors could be used as indicators for further
staging, such as EUS and/or combined CT-PET
resulting in a highly selective approach to
pancreatic exploration for adenocarcinoma of the
head of the pancreas and thereby avoiding
unnecessary laparotomy. This management
protocol may be particularly valuable in patients
who are at high operative risk secondary to
existing co-morbidities. With further
investigation and validation from other studies,
these pre-operative serology markers of
inflammatory response could become established in
the staging process for pancreatic cancer as well
as in determining the outcome following curative
pancreatic resection.