Immunohistochemically Detected Expression of 3 Major Genes (CDKN2A/p16, TP53, and SMAD4/DPC4) Strongly Predicts Survival in Patients With Resectable Pancreatic Cancer

Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

ORIGINAL ARTICLE

Immunohistochemically Detected Expression of 3 Major Genes


(CDKN2A/p16, TP53, and SMAD4/DPC4) Strongly Predicts
Survival in Patients With Resectable Pancreatic Cancer
Minoru Oshima, MD,∗ Keiichi Okano, MD, PhD,∗ Shinobu Muraki, MT,‡ Reiji Haba, MD, PhD,†
Takashi Maeba, MD, PhD,§ Yasuyuki Suzuki, MD, PhD,∗ and Shinichi Yachida, MD, PhD∗ ¶

Conclusions: Genetic alterations of these 3 genes and their accumulation are


Objective: The goal of this retrospective study was to clarify the clinical
strongly associated with malignant behavior of PDAC. Their immunohisto-
implications of the status of the 3 major genes (CDKN2A/p16, TP53, and
chemical assessment at the time of diagnosis may provide a new prognostic
SMAD4/DPC4).
tool, assisting in deciding optimal therapeutic strategies for patients.
Background: Recent whole-exome sequencing had shown that the landscape
of the pancreatic ductal adenocarcinoma (PDAC) genome is notable for 4 Keywords: immunohistochemistry, pancreatic ductal adenocarcinoma, p16,
frequently mutated genes (KRAS, TP53, CDKN2A/p16, and SMAD4/DPC4). p53, Smad4/Dpc4
Methods: We determined immunohistochemically the status of TP53,
(Ann Surg 2013;258: 336–346)
CDKN2A/p16, and SMAD4/DPC4 among the 4 genes because the KRAS
gene is mutated in virtually all PDAC patients, and analyzed relationships
with clinicopathological findings, including survival and patterns of disease
progression, in 106 patients with PDAC undergoing radical surgery.
Results: Abnormal immunolabeling of p53 was detected in 81.1% of PDACs
P ancreatic ductal adenocarcinoma (PDAC) is a devastating dis-
ease and the eighth leading cause of cancer-related deaths in the
world.1 The 5-year survival of PDAC is only approximately 5%, and
and was significantly associated with tumor dedifferentiation (P = 0.022) this figure has remained largely unchanged over the past 25 years, al-
and the presence of locoregional recurrence (P = 0.020). Loss of p16 and though the incidence of PDAC has been rising worldwide.2,3 Surgery
Smad4/Dpc4 immunolabeling was identified in 67.0% and 60.4%, respec- remains the primary treatment modality and the only chance of cure,
tively. Loss of p16 immunolabeling was associated with lymphatic invasion although only 20% of patients present with localized, nonmetastatic
(P = 0.012) and postoperative widespread metastases (P < 0.001). A signifi- disease that is suitable for resection. Therefore, the aggressive resec-
cant correlation was found between Smad4/Dpc4 immunolabeling and tumor tion for locally advanced PDAC continues to be performed, especially
size (P = 0.006), lymphatic invasion (P = 0.033), and lymph node metastasis in Japan,4 but complications are not rare and prognosis after surgery
(P = 0.006). Interestingly, all of the 6 patients demonstrating 5-year survival remains poor, with 5-year survival ranging from 15% to 25%.5,6
had intact SMAD4/DPC4. Kaplan-Meier survival analysis showed that lymph Unfortunately, surgeons sometimes witness the rapid demise of pa-
node metastasis (P = 0.001), lymphatic invasion (P = 0.008), the tumor (T) tients treated with potentially curative surgery, suggesting that a large
factor (T3 vs. T1/T2, P = 0.004), loss of p16 immunolabeling (P = 0.029), proportion staged as having locoregional disease in fact have occult
and loss of Smad4/Dpc4 immunolabeling (P < 0.001) were significantly asso- distant metastases. Thus, curative intent resection for PDAC is asso-
ciated with shorter overall survival. Multivariate analysis revealed that loss of ciated with limited survival benefit in this subset of operable patients.
Smad4/Dpc4 immunolabeling was an independent and significant poor prog- Therefore, there is a need to identify prognostic subtypes of PDACs
nostic factor for overall and disease-free survival. On analysis of combinations to predict clinical and therapeutic outcomes accurately, thereby as-
of the status of these 3 genes, increasing number of alterations reflected poorer sisting in selection of optimal therapeutic strategies for patients with
survival. PDAC.
Whole-exome sequencing of 24 PDACs was reported in 2008.7
Determination of protein-determining exons of 20,661 genes indi-
cated that PDACs contain an average of 63 genomic alterations, the
majority of which are point mutations. Moreover, the genetic land-
From the ∗ Department of Gastroenterological Surgery, Faculty of Medicine,
scape of the PDAC genome is notable for 4 frequently mutated genes,
Kagawa University, Kagawa, Japan; †Department of Pathology, Faculty of designated “mountains,” corresponding to KRAS, CDKN2A/p16,
Medicine, Kagawa University, Kagawa, Japan; ‡Department of Laboratory TP53, and SMAD4/DPC4. Numerous candidate cancer genes altered
Medicine, Social Insurance Ritsurin Hospital, Takamatsu, Kagawa, Japan; §De- at low frequency, designated as “hills,” were also identified. The
partment of Surgery, Social Insurance Ritsurin Hospital, Takamatsu, Kagawa,
Japan; and ¶Division of Refractory Cancer Research, National Cancer Center
4 “mountain” genes are well recognized as contributing to pancre-
Research Institute, Tokyo, Japan. atic carcinogenesis.8 Furthermore, comparative lesion sequencing, in
Supported by a Grant-in-Aid for Scientific Research (23390326, 24659612, which genetic alterations present in 1 cancer sample are analyzed in
24592029) from the Ministry of Education, Science and Culture of Japan, additional geographically or temporally distinct samples from that
a Kagawa University High-Priority Grant, Takeda Science Foundation, the Ue-
hara Memorial Foundation, the Japanese Society of Gastroenterology, and the
same patient,9 has identified the original parental clones of cells
Pancreas Research Foundation of Japan. consisting of mutations (designated “founder” mutations) that were
Disclosure: The authors have no financial conflicts of interest related to this work. present in all samples analyzed for each patient. The parental clones
Supplemental digital content is available for this article. Direct URL citations include all known driver mutations frequently identified in the genome
appear in the printed text and are provided in the HTML and PDF versions of
this article on the journal’s Web site (www.annalsofsurgery.com).
project (KRAS, CDKN2A/p16, TP53, and SMAD4/DPC4).
Reprints: Shinichi Yachida, Laboratory Head, Division of Refractory Cancer Re- Although the parental clones accumulate additional mutations
search, National Cancer Center Research Institute, 5-1-1 Tsukiji, Chuo-ku, (designated “progressor” mutations) during tumor progression and
Tokyo 104-0045, Japan. E-mail: syachida@ncc.go.jp. metastasize to distant organs, the composition of the driver genes
Copyright C 2013 by Lippincott Williams & Wilkins
ISSN: 0003-4932/13/25802-0336
contained within the parental clones determines the basic character-
DOI: 10.1097/SLA.0b013e3182827a65 istics for that carcinoma.9,10 Although the relationships between the

336 | www.annalsofsurgery.com Annals of Surgery r Volume 258, Number 2, August 2013

Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 258, Number 2, August 2013 p16, p53 and Smad4 in Pancreatic Cancer

genetic status of these 4 genes and clinicopathological features, in- (>10 metastases, often hundreds to thousands) in 1 or more organ
cluding survival, have been analyzed, most of the previous reports sites, as previously reported.30
had focused on the relationships between the status of individual
genes and outcome, and they have been inconsistent and frequently Immunohistochemistry
conflicting.11 Furthermore, although genetically engineered mouse Paraffin-embedded samples of the primary carcinomas from
models have been providing the information that the concomitant 106 patients were immunostained for p16, p53, and Smad4/Dpc4.
expression of these mutated genes is essential to invasion and metas- At least 3 different slides were stained for each case to evaluate het-
tasis progress in PDACs,12–14 the influence of the coexistence of erogeneity within the primary cancer. Immunohistochemical label-
these gene alterations in the same PDAC on biological behavior and ing was carried out using a Bond Max instrument (Leica Microsys-
survival outcome is also conflicting.15,16 tems, Wetzlar, Germany), as previously described.31 An antihuman
The objective of the current study was therefore to clar- p16 monoclonal antibody (clone E6H4, ready-to-use; Roche mtm
ify the clinical implications of the status of these “mountain” laboratories AG, Heidelberg, Germany), an antihuman p53 mouse
genes and their combinations in PDAC. Fortunately, it has been monoclonal antibody (clone DO-7, ready-to-use; DakoCytomation,
reported that immunohistochemical labeling of p16, p53, and Grustrup, Denmark), and an antihuman Smad4/Dpc4 mouse mono-
Smad4/Dpc4 reflects the genetic status of CDKN2A/p16,17 TP53,18-24 clonal antibody (clone B-8, diluted 1:100; Santa Cruz Biotechnol-
and SMAD4/DPC4,25,26 respectively. Therefore, we investigated ogy, Santa Cruz, CA) were used. Islet cells in each case served as
immunohistochemically the status of these 3 “mountain” genes an internal control for positive p16 immunolabeling. Immunohisto-
(CDKN2A/p16, TP53, and SMAD4/DPC4) among the total of 4 genes chemical labeling of p16 was scored as intact (positive), indicating
because the KRAS gene is mutated in virtually all PDACs,7,15,27 and the presence of an intact gene, or lost (negative), indicating a deletion
we analyzed the relationships of their status, alone and in combination, or inactivating mutation of the gene had occurred (Fig. 1).17,32 For
with clinicopathological findings in patients with PDAC undergoing p53 immunolabeling, scattered acinar and ductal cells with nuclear
surgical resection.

PATIENTS AND METHODS


This study was approved by the Kagawa University Review
Board.

Patients and Tissue Samples


This study involved patients with PDAC undergoing radical
operations between 2000 and 2011 at the Kagawa University Hospi-
tal and Social Insurance Ritsurin Hospital, Kagawa, Japan. Detailed
clinicopathological and outcome data were collected from medical
records and radiographical images in both hospitals. To check for re-
currence during follow-up, whole-body computed tomography (CT)
and the serum CEA and CA19-9 levels were examined every 3 months
in the first 3 years after the operation and then every 6 months in the
following 2 years in all patients. 18 F-fluorodeoxyglucose positron
emission tomography (FDG-PET) was also performed to confirm the
diagnosis of recurrence if necessary.28 Adenocarcinomas arising in
the presence of intraductal papillary mucinous neoplasms or muci-
nous cystic neoplasms were excluded.
This study included 106 patients with PDAC. The Whip-
ple pancreaticoduodenectomy was conducted for 70 (66.0%) pa-
tients, left-sided pancreatectomy for 32 (30.2%), and total pancre-
atectomy for 4 (3.8%). Postoperatively, 87 (82.1%) of the 106 pa-
tients received adjuvant chemotherapy: 76 were given gemcitabine
or gemcitabine-based chemotherapy, 7 received 5-fluorouracil (5-
FU)-based chemotherapy, and 4 received S-1,29 as the first-line
chemotherapy.

First Site of Progression and Patterns of Failure


The primary site of recurrence was defined as locoregional,
distant, or synchronous locoregional and distant, based on the ra-
diographical findings. Locoregional recurrence was defined as recur-
rence in the region of the pancreatic bed and the root of the mesentery
or recurrence in the soft tissues or lymph nodes beyond the pancre-
atic bed or within the peritoneal cavity (including ascites and/or the FIGURE 1. Typical immunohistochemical labeling profiles of
presence of wound recurrence). Distant recurrence was defined as p16 in pancreatic ductal adenocarcinomas. A, Positive p16 im-
recurrence in the liver, lungs, or other distant organs. munolabeling indicating an intact CDKN2A/p16 gene. Nuclear
The dominant pattern of failure was determined from the last and cytoplasmic labeling for p16 is present within the neo-
follow-up CT and classified as locoregional recurrence, widespread plastic glands. Normal islet cells serve as positive controls (N).
metastases, and a combined type indistinguishable because of rapid B, Negative p16 immunolabeling indicating alteration of the
progression, with both locoregional and widespread metastases. CDKN2A/p16 gene. In contrast, positive labeling is seen within
Widespread metastases were defined as a high metastatic burden adjacent normal islet cells (N).


C 2013 Lippincott Williams & Wilkins www.annalsofsurgery.com | 337

Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Oshima et al Annals of Surgery r Volume 258, Number 2, August 2013

labeling were typically present in the adjacent normal tissue ber 2011. The median survival was estimated using the Kaplan-Meier
(Fig. 2A). Therefore, p53 immunolabeling was considered to be “ab- method, and the difference was tested using the log-rank test. The 1-,
normal” when either the neoplastic cells showed a virtual absence 3- and 5-year survival rates were estimated using life tables. P values
of immunolabeling compared with adjacent normal tissue (immuno- less than 0.05 were considered statistically significant. Variables that
labeling in <5% of neoplastic cells), suggesting the presence of an were found to be significant on univariate analysis at P < 0.1 were
intragenic deletion, nonsense or frameshift mutation (Fig. 2D), or included in multivariate analysis in a backward stepwise fashion. Cox
showed robust nuclear accumulation of immunolabeled protein in proportional hazards models were generated for multivariate analy-
≥30% of neoplastic cells compared with adjacent normal cells (Fig. sis. Statistical analysis was performed using IBM SPSS Statistics 20
2C).18-24 Normal acinar cells, ductal cells, islet cells, and stromal cells (IBM, Armonk, NY).
in each case served as internal controls for positive Smad4/Dpc4 im-
munolabeling. Immunohistochemical labeling of Smad4/Dpc4 was
scored as intact (positive), indicating the presence of an intact gene, RESULTS
or lost (negative), indicating a deletion or inactivating mutation of the
gene had occurred (Fig. 3).25 Negative controls for each of the anti- Clinicopathological Characteristics and Outcome
bodies were included using nonimmune serum instead of the primary The cohort of 106 patients (Table 1) consisted of 44 women and
antibodies. 62 men. The mean age at operation was 68.0 years, with a median age
of 69.5 years and range of 36–86 years. Thirty-six (34.0%) patients
were alive at the census date (December 2011). The 30-day mortality
Statistics rate was 0%. Sixty-six (62.3%) patients died of PDAC and 4 (3.8%) of
Frequency distributions were compared by χ 2 test. Continuous other causes. The median overall survival was 22.1 months, with 1-,
variables were compared using the Student t-test. The principal out- 3-, and 5-year survival rates of 71.9%, 28.8%, and 17.5%, respec-
come measure was length of survival as measured from the time of tively, in all 106 patients. The majority of tumors were well dif-
the original surgery. Patients alive at the time of follow-up point were ferentiated (57.5%), followed by moderately differentiated (28.3%),
censored. The last follow-up period for patients still alive was Decem- and 14.2% of tumors were poorly differentiated. Most tumors were

FIGURE 2. Typical immunohistochemical labeling profiles of p53 in pancreatic tissues. A, Normal pancreatic tissue. Scattered
ductal and acinar cells with positive nuclear labeling are present. B, Pancreatic ductal adenocarcinoma (PDAC) showing a
“normal” pattern of p53 immunohistochemical labeling. Scattered cells in the neoplastic glands with positive nuclear labeling are
present. C, Example of PDAC with diffusely positive nuclear labeling for p53. D, Example of PDAC with loss of nuclear labeling
for p53 (right side). In contrast, scattered reactive pancreatic ducts (N) with positive nuclear labeling are present in the adjacent
nonneoplastic tissue (left side).

338 | www.annalsofsurgery.com 
C 2013 Lippincott Williams & Wilkins

Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 258, Number 2, August 2013 p16, p53 and Smad4 in Pancreatic Cancer

38 (35.8%), and synchronous locoregional and distal in 17 (16.3%).


The 27 patients who had locoregional recurrence as the first sign of
progression had significantly longer disease-free survival than the 38
patients who had distant metastases as the first sign (11.0 vs. 4.0
months, P = 0.016).
The patients who had the dominant pattern of failure deter-
mined as locoregional recurrence often had no metastasis or a low
metastatic burden (1–10 metastases). On the other hand, the patients
who had the dominant pattern of failure determined as widespread
metastases often had 100–1000 metastases without locoregional re-
currence, in keeping with previous reports from a large tertiary care
medical center.30 Of the 63 patients who died of PDAC with avail-
able information, the dominant pattern of failure was locoregional
recurrence in 9 (14.1%), widespread metastases in 29 (45.3%), and
combined in 25 (39.1%).

Immunohistochemical Analysis
Immunohistochemical results are also presented in Table 1.
Loss of p16 immunolabeling was identified in 71 (67.0%) of the 106
patients (Fig. 1). Abnormal immunolabeling of p53 was detected in
86 (81.1%) of the 106 PDACs (Fig. 2). Twenty-seven (25.5%) PDACs
showed a virtual absence of p53 immunolabeling compared with ad-
jacent normal tissue (immunolabeling in <5% of neoplastic cells)
(Fig. 2D) and 58 (54.7%) showed robust nuclear accumulation of im-
munolabeled p53 protein in ≥30% of neoplastic cells compared with
adjacent normal cells (Fig. 2C). Intratumoral heterogeneity of p53 im-
munolabeling was observed in 1 (1.0%) PDAC. Loss of Smad4/Dpc4
immunolabeling was identified in 58 (54.7%) of the 106 PDACs
(Fig. 3) and intratumoral heterogeneity in 6 (5.7%). Immunohis-
tochemically detected heterogeneity of p53 and Smad4/Dpc4 was
categorized into abnormal immunolabeling of p53 and loss of im-
munolabeling of Smad4/Dpc4, respectively.

Immunohistochemical Labeling of p16, p53, and


Smad4/Dpc4 Versus Clinicopathological Factors
FIGURE 3. Typical immunohistochemical labeling profiles
Table 2 summarizes data for relationships between p16, p53,
of Smad4/Dpc4 in pancreatic ductal adenocarcinoma. A,
or Smad4/Dpc4 immunolabeling and clinicopathological parameters.
Positive Smad4/Dpc4 immunolabeling indicating an intact
There was a significant association between loss of p16 immunola-
SMAD4/DPC4 gene. Nuclear and cytoplasmic labeling for
beling and lymphatic invasion (P = 0.012). Abnormal immunolabel-
Smad4/Dpc4 is present in the neoplastic glands, as well as in
ing of p53 was significantly linked with tumor dedifferentiation (P
adjacent islet cells (N) present in the same section. B, Neg-
= 0.022) and the presence of locoregional recurrence (P = 0.020),
ative Smad4/Dpc4 immunolabeling indicating alteration of
whereas loss of Smad4/Dpc4 immunolabeling was associated with tu-
the SMAD4/DPC4 gene. Nuclear and cytoplasmic labeling for
mor size (P = 0.006), lymph node metastasis (P = 0.006), lymphatic
Smad4/Dpc4 is virtually absent within the neoplastic glands
invasion (P = 0.033), and pathological UICC stage (P = 0.018). No-
(asterisks). In contrast, strong positive labeling is apparent
tably, all of the 6 patients with more than 5-year survival at the census
within adjacent islet cells (N).
date had positive Smad4/Dpc4 immunolabeling.

Clinicopathological Characteristics Versus Survival


located in the head of the pancreas (67.9%) and were more than 20 mm The results of univariate analysis using log-rank tests of the
in maximal diameter (77.4%). The great majority (82.1%) of patients clinicopathological characteristics in relation to overall or disease-
had T3 tumors. Lymph node metastases were present in 72 (67.9%) of free survival are presented in Table 1. Lymph node metastasis (P =
the 106 patients, lymphatic invasion in 82 (77.4%), and vascular inva- 0.001), presence of lymphatic invasion (P = 0.008), and higher tumor
sion in 92 (86.8%). Four (3.8%), 5 (4.7%), 24 (22.6%), and 73 (68.9%) (T) factor (T3 vs. T1/T2, P = 0.004) were significantly associated with
patients presented with the Union for International Cancer Control shorter overall survival. The clinicopathological factors associated
(UICC) stage IA, IB, IIA, and IIB disease, respectively. No patients with a significantly shorter disease-free survival on univariate analysis
presented with UICC stage III and IV disease. Thirty-one (29.2%) were larger tumor size (>20 mm vs. ≤20 mm, P = 0.042), lymph
patients had a positive surgical margin. Eighty-seven (82.1%) of the node metastases (P < 0.001), presence of lymphatic invasion (P =
106 received adjuvant chemotherapy, but adjuvant chemotherapy in 0.009), and T3 tumors as compared with T1/T2 tumors (P = 0.030).
any form provided no significant survival benefit within this cohort
of PDAC. Immunohistochemical Labeling of p16, p53, and
During the study period, recurrent disease occurred in 82 Smad4/Dpc4 Versus Survival
(78.1%) of the 105 patients with available information. The first Loss of p16 (P = 0.029) and Smad4/Dpc4 (P < 0.001) im-
site of progression was locoregional in 27 (25.5%) patients, distant in munolabeling was associated with a significantly shorter overall


C 2013 Lippincott Williams & Wilkins www.annalsofsurgery.com | 339

Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Oshima et al Annals of Surgery r Volume 258, Number 2, August 2013

TABLE 1. Clinicopathological Parameters and Outcome (n = 106)


Overall Survival Disease-Free Survival
Variable No. of Patients (%) Median (Months) Log-Rank (P Value) Median (Months) Log-Rank (P Value)
Gender
Female 44 (41.5) 24.8 0.200 11.0 0.189
Male 62 (58.5) 20.1 8.0
Age, years
Mean 68.0 (SD, 9.8)
Median 69.5
Range 36–86
Outcome
Follow-up, months
Range 2.8–124.8
Median 17.3
Death
PDAC 65 (61.3)
Others 4 (3.8)
Alive
NED 24 (22.6)
AWD 13 (12.3)
Tumor location
Head 72 (67.9) 20.9 0.157 8.0 0.087
Body/tail 34 (32.1) 28.1 13.8
Tumor size, mm
Mean 31.0 (SD, 13.7)
Median 29.5
Range 10–77
≤ 20 mm 24 (22.6) 21.5 0.104 14.0 0.042
> 20 mm 82 (77.4) 28.1 7.4
Lymph nodes
Negative 34 (32.1) 35.9 0.001 17.0 < 0.001
Positive 72 (67.9) 20.9 7.0
Margin status
R0 75 (70.8) 25.0 0.060 9.8 0.059
R1 31 (29.2) 18.3 5.0
Differentiation
Well 61 (57.5) 24.8 0.502 10.0 0.397
Moderate 30 (28.3) 20.8 7.4
Poor 15 (14.2) 18.5 4.0
Lymphatic invasion
Negative 24 (22.6) 42.5 0.008 16.0 0.009
Positive 82 (77.4) 20.9 7.4
Vascular invasion
Negative 14 (13.2) 25.0 0.359 14.0 0.151
Positive 92 (86.8) 21.6 7.4
T factor (UICC)
T1/T2 19 (17.9) 39.5 0.004 15.0 0.030
T3 87 (82.1) 20.1 7.3
Stage (UICC)
IA 4 (3.8)
IB 5 (4.7)
IIA 24 (22.6) 20.5 0.126 9.5 0.018
IIB 73 (68.9) 21.5 7.0
p16 Immunohistochemistry
Positive (intact) 35 (33.0) 31.0 0.029 15.3 0.015
Negative (loss) 71 (67.0) 20.1 6.0
p53 Immunohistochemistry
Normal 20 (18.9) 31.0 0.060 16.0 0.051
Abnormal 86 (81.1) 20.1 7.3
Smad4/Dpc4 mmunohistochemistry
Positive (intact) 42 (39.6) 30.1 < 0.001 13.5 < 0.001
Negative (loss) 64 (60.4) 18.3 6.0
Chemotherapy
No adjuvant 18 (17.0) 9.7 0.832 5.1 0.221
Adjuvant 87 (82,1) 23.6 9.7
Unknown 1 (0.9) (Continued)

340 | www.annalsofsurgery.com 
C 2013 Lippincott Williams & Wilkins

Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 258, Number 2, August 2013 p16, p53 and Smad4 in Pancreatic Cancer

TABLE 1. (Continued)
Overall Survival Disease-Free Survival
Variable No. of Patients (%) Median (Months) Log-Rank (P Value) Median (Months) Log-Rank (P Value)
Locoregional recurrence
Present 56 (52.8)
Absent 45 (42.5)
Unknown 5 (4.7)
First sign of recurrence
No recurrence 23 (21.7)
Locoregional 27 (25.5) 24.8 0.004 11.0 0.047
Distant metastases 38 (35.8) 20.1 4.0
Synchronous 17 (16.3) 9.3 4.9
Unknown 1 (0.9)
Dominant pattern of failure
No recurrence 23 (21.7)
Locoregional 16 (15.1) 25.1 0.044 11.5 0.045
Widespread 36 (34.0) 16.7 4.0
Combined 27 (25.5) 17.5 5.9
Unknown 4 (3.8)
The bold values indicate P values less than 0.05. AWD, alive with disease; NED, no evidence of disease; PDAC, pancreatic ductal adenocarcinoma; SD, standard
deviation; UICC, Union for International Cancer Control.

survival (Table 1 and Fig. 4). Similarly, loss of p16 (P = 0.015) loss of Smad4/Dpc4 immunolabeling was an independent predictor
and Smad4/Dpc4 (P < 0.001) immunolabeling was associated with (hazard ratio 1.888, P = 0.015) (Supplemental Digital Content 2,
a significantly shorter disease-free survival (Table 1 and Supplemen- available at http://links.lww.com/SLA/A348).
tal Digital Content 1, available at http://links.lww.com/SLA/A349).
There were borderline significant differences in abnormal labeling Clinicopathological Characteristics Versus Pattern
of p53 with regard to overall survival (P = 0.060) and disease- of Disease Progression
free survival (P = 0.051). There was no difference of survival be- We determined the relationships between the clinicopatholog-
tween the types of abnormal p53 immunolabeling (virtual absence ical factors (tumor size, lymph node metastasis, margin status, dif-
of p53 immunolabeling vs. robust nuclear accumulation of p53 im- ferentiation, lymphatic invasion, and vascular invasion), and the first
munolabeling: overall survival, P = 0.528; disease-free survival, site of recurrence and the dominant pattern of disease progression.
P = 0.643). Among them, only lymphatic invasion was significantly associated
Next, based on the number of altered genes, we classified with locoregional recurrence as the primary sign of recurrence (P
the patients into 4 groups: zero gene (n = 4), 1 gene (n = 18), 2 = 0.011). No clinicopathological factors significantly impacted upon
genes (n = 49), and 3 genes (n = 35). Statistical analysis compar- the pattern of recurrence. The disease-free survival duration in 38
ing the zero gene and other groups was not performed because of patients who presented with distant metastases as the first sign of
the small number of patients. Kaplan-Meier survival analysis showed recurrence was very short (median 4.0 months) and this rapid pro-
that there was a significant difference between 1-gene and 3-genes gression strongly suggests that occult metastatic disease was present
group (P < 0.001) and between 2-genes and 3-genes group (P = at the time of surgery.
0.002) in overall survival (Fig. 5). The patients with 2 altered genes
had shorter overall survival than those with 1 altered gene but without Immunohistochemical Labeling of p16, p53, and
statistical significance (P = 0.076). Similarly, Kaplan-Meier survival
analysis for disease-free survival showed that there was a signif-
Smad4/Dpc4 Versus Pattern of Disease Progression
icant difference between 1-gene and 2-genes group (P = 0.015), We also compared the immunohistochemical findings of p16,
between 1-gene and 3-genes group (P < 0.001) and between 2-genes p53, and Smad4/Dpc4 with the first sign of recurrence (Table 2) and
and 3-genes group (P = 0.007) (Fig. 5). The increasing number found that loss of p16 immunolabeling was significantly associated
of the altered genes robustly reflected major differences in survival with distant metastases (locoregional recurrence vs. distant metas-
outcome. tases, P = 0.005). Similarly, comparing the immunohistochemical
findings for p16, p53, and Smad4/Dpc4 with regard to the domi-
nant pattern of disease progression, loss of p16 immunolabeling was
Relationship Between Prognostic Factors and significantly associated with widespread metastases (locoregional re-
Survival by Multivariate Analysis currence vs. widespread metastases, P < 0.001).
Multivariate models using Cox proportional hazards analysis
were conducted with the parameters that were significant at the P < DISCUSSION
0.1 level on univariate analysis using log-rank tests. Multivariate anal- It is accepted that various pathological factors, including tumor
ysis demonstrated loss of Smad4/Dpc4 immunolabeling (hazard ratio differentiation, tumor size, lymph node status, lymphatic/vascular in-
2.045, P = 0.014) to be an independent prognostic factor for overall vasion, and resection margin involvement, influence outcome after
survival (Table 3). There were borderline significant relationships of PDAC resection.33 However, “preoperative” identification of patients
overall survival with the T factor (hazard ratio 2.167, P = 0.056) with a poor prognosis is desirable to aid appropriate clinical decision-
and abnormal p53 immunolabeling (hazard ratio 1.978, P = 0.051). making. In the present study, univariate analyses revealed that tradi-
Multivariate analysis for disease-free survival duration showed that tional clinicopathological factors, including T factor (T3 vs. T1/T2),


C 2013 Lippincott Williams & Wilkins www.annalsofsurgery.com | 341

Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Oshima et al Annals of Surgery r Volume 258, Number 2, August 2013

TABLE 2. Clinicopathological Parameters and Immunohistochemical labeling of p53, p16 and Smad4/Dpc4 (n = 106)
p16 Immunohistochemistry p53 Immunohistochemistry Smad4/Dpc4 Immunohistochemistry
Positive Negative Abnormal Positive Negative
Variable (Intact) (%) (Loss) (%) P Value Normal (%) (%) P Value (Intact) (%) (Loss) (%) P Value
Tumor size
Mean (mm) 32.1 30.4 0.539 28.1 31.6 0.300 26.5 33.9 0.006
≤ 20 mm 6 (17.1) 18 (25.4) 0.342 7 (35.0) 17 (19.8) 0.143 13 (31.0) 11 (17.2) 0.098
> 20 mm 29 (82.9) 53 (74.6) 13 (65.0) 69 (80.2) 29 (79.0) 53 (82.8)
Lymph nodes
Negative 15 (42.9) 19 (26.8) 0.095 8 (40.0) 26 (30.2) 0.393 20 (47.6) 14 (21.9) 0.006
Positive 20 (57.1) 52 (73.2) 12 (60.0) 60 (69.8) 22 (52.4) 50 (78.1)
Margin status
R0 29 (82.9) 46 (64.8) 0.055 17 (85.0) 58 (67.4) 0.120 31 (73.6) 44 (68.8) 0.575
R1 6 (17.1) 25 (35.2) 3 (15.0) 28 (32.6) 11 (26.2) 20 (31.2)
Differentiation
Well 22 (62.9) 39 (54.9) 0.669 17 (85.0) 44 (51.2) 0.022 26 (61.9) 35 (54.7) 0.695
Moderate 8 (22.9) 22 (31.0) 2 (10.0) 28 (32.6) 10 (23.8) 20 (31.2)
Poor 5 (14.3) 10 (14.1) 1 (5.0) 14 (16.3) 6 (14.3) 9 (14.1)
Lymphatic invasion
Negative 13 (37.1) 11 (15.5) 0.012 6 (30.0) 18 (20.9) 0.383 14 (33.3) 10 (15.6) 0.033
Positive 22 (62.9) 14 (70.0) 68 (79.1) 60 (84.5) 28 (66.7) 54 (84.4)
Vascular invasion
Negative 7 (20.0) 7 (9.9) 0.147 5 (25.0) 9 (10.5) 0.084 7 (16.7) 7 (10.9) 0.394
Positive 28 (80.0) 64 (90.1) 15 (75.0) 77 (89.5) 35 (83.3) 57 (89.1)
T-factor (UICC)
T1 3 (8.6) 4 (5.6) 0.648 3 (15.0) 4 (4.7) 0.244 5 (11.9) 2 (3.1) 0.195
T2 5 (14.3) 7 (9.9) 2 (10.0) 10 (11.6) 5 (11.9) 7 (10.9)
T3 27 (77.1) 60 (84.5) 15 (75.0) 72 (83.7) 32 (76.2) 55 (85.9)
T4 0 0 0 0 0 0
Stage (UICC)
IA 3 (8.6) 1 (1.4) 0.107 2 (10.0) 2 (2.3) 0.226 4 (9.5) 0 0.018
IB 3 (8.6) 2 (2.8) 2 (10.0) 3 (3.5) 3 (7.1) 2 (3.1)
IIA 9 (25.7) 15 (21.1) 4 (20.0) 20 (23.3) 12 (28.6) 12 (18.8)
IIB 20 (57.1) 53 (74.6) 12 (60.0) 61 (70.9) 23 (54.8) 50 (78.1)
Locoregional recurrence
Present 18 (52.9) 38 (56.7) 0.718 6 (31.6) 50 (61.0) 0.020 20 (48.8) 36 (60.0) 0.265
Absent 16 (47.1) 29 (43.3) 13 (68.4) 32 (39.0) 21 (51.2) 24 (40.0)
Unknown 1 4 1 4 1 4
First sign of recurrence
No recurrence 12 11 7 16 16 7
Locoregional 13 (56.5) 14 (23.7) 0.015 5 (38.5) 22 (31.9) 0.450 8 (30.8) 19 (33.9) 0.640
Distant metastases 6 (26.1) 32 (54.2) 7 (53.8) 31 (44.9) 11 (42.3) 27 (48.2)
Synchronous 4 (17.4) 13 (22.0) 1 (7.7) 16 (23.2) 7 (26.9) 10 (17.9)
Unknown 0 1 1 1
Dominant pattern of failure
No recurrence 12 11 7 16 16 7
Locoregional 10 (45.5) 6 (10.5) 0.001 2 (15.4) 14 (21.2) 0.791 4 (16.0) 12 (22.2) 0.074
Widespread 5 (22.7) 31 (54.4) 7 (53.8) 29 (43.9) 8 (32.0) 28 (51.9)
Combined 7 (25.9) 20 (35.1) 4 (30.8) 23 (34.8) 13 (52.0) 14 (25.9)
Unknown 1 3 0 4 1 3
The bold values indicate P values less than 0.05. UICC, Union for International Cancer Control.

lymphatic invasion, and lymph node metastasis, predict outcome as encodes a critical transcription factor involved in the TGF-β signal
expected, and furthermore that p16 and Smad4/Dpc4 immunolabeling pathway, whose dysregulation promotes the epithelial mesenchy-
correlated significantly with the prognosis. Multivariate analyses in- mal transition, an event that occurs normally during embryonic
dicated that loss of Smad4/Dpc4 immunolabeling was an independent development and is thought to be a critical contributor to tumor
predictor of shorter overall and disease-free survival. In addition, loss invasiveness and metastasis.39–41 Furthermore, it has been recently
of p16 immunolabeling was significantly associated with widespread documented that mutations of SMAD4/DPC4 predict survival15 and
metastases. loss of Smad4/Dpc4 immunolabeling correlates with patterns of
The clinical implications of the genetic changes linked tumor progression.30,42 We here clarified clinical significance of
to prognosis may still be conflicting and controversial.15,16,34–36 Smad4/Dpc4 immunolabeling loss as a predictive marker of lymph
SMAD4/DPC4 was here found to be inactivated in 64 (60.4%) of node metastasis and postoperative survival outcome in a population
the 106 PDACs, similar to the value (55%) reported by Hahn et al of patients independent from those initially described,15,30 although
(homozygous deletion in 35% of cases and loss of 1 allele coupled it should be stressed that it was not associated with the pattern of
with an intragenic mutation in the second allele in 20%).37,38 It tumor progression.

342 | www.annalsofsurgery.com 
C 2013 Lippincott Williams & Wilkins

Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 258, Number 2, August 2013 p16, p53 and Smad4 in Pancreatic Cancer

suggesting wild-type TP53; (b) a virtual absence of p53 immunola-


beling (Fig. 2D) suggesting the presence of an intragenic deletion,
nonsense, or frameshift mutation; and (c) robust nuclear accumula-
tion of immunolabeled p53 protein (Fig. 2C) suggesting the presence
of a missense mutation.18-24 Such methodological variation in assess-
ment of p53 expression might explain the inconsistence between the
present study and previous reports.22
The CDKN2A/p16 gene is reported to be inactivated by ho-
mozygous deletion in 40% of PDACs, by loss of 1 allele coupled with
an intragenic mutation in the second allele in a further 40%, and by
promoter methylation in an additional 15%,38 the total proportion be-
ing higher than that (67.0%) immunohistochemically detected in the
present study. Immunohistochemically detected abnormal expression
of p16 is considered to indicate homozygous deletions or null mu-
tations of the gene, which are potentially harmful genetic alterations
in the Rb/p16 tumor-suppressive pathway.17 Although CDKN2A/p16
alteration is considered to occur in the relatively early phase of pan-
creatic carcinogenesis, it is interesting to note that activated KRAS
and CDKN2A/p16 deficiency cooperate to produce metastatic PDAC
in a genetically engineered mouse model, whereas sole activation of
KRAS is linked primarily to the development of focal premalignant
ductal lesions.46,47 This might be related to our finding that loss of
p16 immunolabeling was associated with widespread metastases.
Considering our data on the 3 genes together, assessment of
combinations of genetic alterations is presumably more predictive of
survival outcome than that of single changes. From comparative se-
quencing studies9,10,48 and genetically engineered mouse models,12-14
there is no doubt that accumulation of genetic alterations has a great
influence on the biology of PDACs by contributing to clonal evolution.
“Mountain” genes constituting the “parental” clone in each PDAC
might influence the biological characteristics, and further studies on
their cooperative interaction appear warranted.
In the last decade, the introduction of endoscopic ultrasound-
fine needle aspiration (EUS-FNA) has changed the diagnostic strat-
egy for PDAC.49 The advantage of this approach is the ability to
sample sufficient cancer tissue. Most important, genetic alterations
can be preoperatively estimated immunohistochemically using EUS-
FNA materials. Immunohistochemical analysis is used widely for
evaluating molecular markers in clinical tissue specimens, and im-
munolabeling patterns of p16, p53, and Smad4/Dpc4 are functional
markers of their genetic status, regardless of the mechanism of al-
terations. Although several more sophisticated techniques, such as
the cDNA microarray, fluorescence in-situ hybridization, and quan-
titative reverse transcriptase-polymerase chain reaction, are being
translated into clinical practice, they are still impractical in routine
clinical settings. By assessing the expression status of molecules
in EUS-FNA specimens in a prospective manner, accurate classifi-
FIGURE 4. Kaplan-Meier survival curves for patients after cation of PDAC patients should be achievable. For example, more
surgery for pancreatic ductal adenocarcinoma demonstrating aggressive local therapeutic approaches, including vascular resec-
prognostic influence for p16 immunolabeling (A), p53 im- tion/reconstruction, may be more beneficial in patients having a small
munolabeling (B), and Smad4/Dpc4 immunolabeling (C) with number of the altered “mountain” genes. In contrast, patients with
reference to overall survival. borderline resectable PDAC and alteration of all 3 genes might be
spared the risk of surgery because the postoperative survival du-
ration might be expected to be short. Moreover, because cancer is
That TP53 mutations occur in pancreatic carcinogenesis is fundamentally a “genetic” disease, stratification of patients based on
well established, but it remains uncertain whether the pattern of p53 the genetic status for entry into future clinical trials is reasonable
immunolabeling has any prognostic implications.11 In the present and may provide sufficient power to accurately define the role of
study, there were borderline significant differences in abnormal la- currently available treatment modalities in the routine management
beling of p53 with overall survival (P = 0.060) and disease-free of PDAC.
survival (P = 0.051). In previous studies, p53 immunolabeling was We acknowledge that the present study has several limitations.
simply classified into “positive (or high)” or “negative (or low)”.43–45 First, the study design is retrospective and the number of patients is not
However, when observed carefully, the pattern of p53 immunola- large. Furthermore, the patients did not all receive the same treatment
beling in PDACs could be classified into 3 categories, as compared (ie, type of operation and adjuvant therapy), although the differences
with adjacent normal tissue: (a) normal immunolabeling (Fig. 2B) did not appear to influence survival outcome in the present study.


C 2013 Lippincott Williams & Wilkins www.annalsofsurgery.com | 343

Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Oshima et al Annals of Surgery r Volume 258, Number 2, August 2013

FIGURE 5. Kaplan-Meier survival curves for patients after surgery for pancreatic ductal adenocarcinoma demonstrating relation-
ships of the number of altered genes with postoperative overall survival (A) and disease-free survival (B).

Prospective validation of the expression of p16, p53, Smad4/Dpc4 (14.1%) died as a result of complications of locoregional tumor pro-
and their combinations as prognostic biomarkers in a large series of gression, a lower proportion than in the rapid autopsy series from
patients is clearly warranted. The present study also focused on the the Johns Hopkins Medical Institutions,30 probably because of differ-
first sign of recurrence and patterns of failure after surgery. Among ences in study populations. In the present study, however, one-third of
the 63 patients who died of PDAC with available information, 9 patients were classified into the combined pattern (both locoregional

344 | www.annalsofsurgery.com 
C 2013 Lippincott Williams & Wilkins

Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 258, Number 2, August 2013 p16, p53 and Smad4 in Pancreatic Cancer

TABLE 3. Multivariate Analysis of Prognostic Factors for Overall Survival


(n = 106)
Hazard 95% Confidence
Variable Ratio Limit P Value
Lymph node metastasis
Positive (n = 72) 1.644 0.872–3.101 0.125
Negative (n = 34) 1
Margin status
Positive (n = 31) 1.182 0.689–2.029 0.543
Negative (n = 75) 1
T-factor (UICC)
T3 (n = 87) 2.167 0.981–4.787 0.056
T1/T2 (n = 19) 1
Lymphatic invasion
Positive (n = 82) 1.813 0.825–3.983 0.139
Negative (n = 24) 1
p16 Immunohistochemistry
Negative (loss) (n = 71) 1.508 0.823–2.765 0.184
Positive (intact) (n = 35) 1
p53 Immunohistochemistry
Abnormal (n = 86) 1.978 0.997–3.922 0.051
Normal (n = 20) 1
Smad4/Dpc4 immunohistochemistry
Negative (loss) (n = 64) 2.045 1.154–3.624 0.014
Positive (intact) (n = 42) 1
The bold value indicates P value less than 0.05.

recurrence and widespread metastases), which indicates that it might 9. Yachida S, Jones S, Bozic I, et al. Distant metastasis occurs late during the
be practically difficult to classify patterns of progression only with genetic evolution of pancreatic cancer. Nature. 2010;467:1114–1117.
radiographical images, in contrast to the autopsy series. 10. Campbell PJ, Shinichi Y, Mudie LJ, et al. The patterns and dynamics of
genomic instability in metastatic pancreatic cancer. Nature. 2010;467:1109–
In conclusion, genetic alterations of 3 major genes and their 1113.
accumulation are strongly associated with malignant behavior of 11. Ansari D, Rosendahl A, Elebro J, et al. Systematic review of immunohisto-
PDACs. Furthermore, the present study suggests the possibility that chemical biomarkers to identify prognostic subgroups of patients with pancre-
prognosis may be preoperatively predictable on the basis of im- atic cancer. Br J Surg. 2011;98:1041–1055.
munohistochemically detected expression of 3 genes using EUS-FNA 12. Izeradjene K, Combs C, Best M, et al. Kras(G12D) and Smad4/Dpc4 hap-
materials. loinsufficiency cooperate to induce mucinous cystic neoplasms and invasive
adenocarcinoma of the pancreas. Cancer Cell. 2007;11:229–243.
13. Frese KK, Tuveson DA, Maximizing mouse cancer models. Nat Rev Cancer.
ACKNOWLEDGMENT 2007;7:645–658.
14. Tuveson D, Hanahan D, Translational medicine: Cancer lessons from mice to
The authors thank Professor Christine A. Iacobuzio-Donahue humans. Nature. 2011;471:316–317.
for critical reading of the manuscript. 15. Blackford A, Serrano OK, Wolfgang CL, et al. SMAD4 gene mutations
are associated with poor prognosis in pancreatic cancer. Clin Cancer Res.
2009;15:4674–4679.
REFERENCES 16. Salek C, Minarikova P, Benesova L, et al. Mutation status of K-ras, p53
1. Ferlay J, Shin H-R, Bray F, et al. Estimates of worldwide burden of cancer in and allelic losses at 9p and 18q are not prognostic markers in patients with
2008: GLOBOCAN 2008. Int J Cancer. 2010;127:2893–2917. pancreatic cancer. Anticancer Res. 2009;29:1803–1810.
2. Siegel R, Ward E, Brawley O, et al. Cancer statistics, 2011: the impact of 17. Geradts J, Hruban RH, Schutte M, et al. Immunohistochemical p16INK4a
eliminating socioeconomic and racial disparities on premature cancer deaths. analysis of archival tumors with deletion, hypermethylation, or mutation of
CA Cancer J Clin. 2011;61:212–236. the CDKN2/MTS1 gene. A comparison of four commercial antibodies. Appl
3. Malvezzi M, Arfé A, Bertuccio P, et al. European cancer mortality predictions Immunohistochem Mol Morphol. 2000;8:71–79.
for the year 2011. Ann Oncol. 2011;22:947–956. 18. Baas IO, Mulder J-WR, Offerhaus GJA, et al. An evaluation of six antibodies
4. Yamamoto Y, Sakamoto Y, Ban D, et al. Is celiac axis resection justified for for immunohistochemistry of mutant p53 gene product in archival colorectal
T4 pancreatic body cancer? Surgery. 2012;151:61–69. neoplasms. J Pathol. 1994;172:5–12.
5. Ueno H, Kosuge T, Matsuyama Y, et al. A randomised phase III trial comparing 19. Melhem MF, Law JC, el-Ashmawy L, et al. Assessment of sensitivity and
gemcitabine with surgery-only in patients with resected pancreatic cancer: specificity of immunohistochemical staining of p53 in lung and head and neck
Japanese Study Group of Adjuvant Therapy for Pancreatic Cancer. Br J Cancer. cancers. Am J Pathol. 1995;146:1170–1177.
2009;101:908–915. 20. Sjogren S, Inganäs M, Norberg T, et al. The p53 gene in breast cancer: prognos-
6. Herman JM, Swartz MJ, Hsu CC, et al. Analysis of fluorouracil-based adjuvant tic value of complementary DNA sequencing versus immunohistochemistry. J
chemotherapy and radiation after pancreaticoduodenectomy for ductal adeno- Natl Cancer Inst. 1996;88:173–182.
carcinoma of the pancreas: results of a large, prospectively collected database 21. Obata A, Eura M, Sasaki J, et al. Clinical significance of p53 functional loss in
at the Johns Hopkins Hospital. J Clin Oncol. 2008;26:3503–3510. squamous cell carcinoma of the oropharynx. Int J Cancer. 2000;89:187–193.
7. Jones S, Zhang X, Parsons DW, et al. Core signaling pathways in human pan- 22. Nenutil R, Smardova J, Pavlova S, et al. Discriminating functional and non-
creatic cancers revealed by global genomic analyses. Science. 2008;321:1801– functional p53 in human tumours by p53 and MDM2 immunohistochemistry.
1806. J Pathol. 2005;207:251–259.
8. Hruban RH, Adsay NV, Albores-Saavedra J, et al. Pancreatic intraepithelial 23. Langeveld D, van Hattem WA, de Leng WWJ, et al. SMAD4 immunohisto-
neoplasia: a new nomenclature and classification system for pancreatic duct chemistry reflects genetic status in juvenile polyposis syndrome. Clin Cancer
lesions. Am J Surg Pathol. 2001;25:579–586. Res. 2010;16:4126–4134.


C 2013 Lippincott Williams & Wilkins www.annalsofsurgery.com | 345

Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Oshima et al Annals of Surgery r Volume 258, Number 2, August 2013

24. Yachida S, Vakiani E, White CM, et al. Small cell and large cell neu- 37. Hahn SA, Schutte M, Hoque AT, et al. DPC4, a candidate tumor suppressor
roendocrine carcinomas of the pancreas are genetically similar and distinct gene at human chromosome 18q21.1. Science. 1996;271:350–353.
from well-differentiated pancreatic neuroendocrine tumors. Am J Surg Pathol. 38. Maitra A, Hruban RH. Pancreatic cancer. Annu Rev Pathol. 2008;3:157–188.
2012;36:173–184.
39. Bardeesy N, Cheng KH, Berger JH, et al. Smad4 is dispensable for normal
25. Wilentz RE, Su GH, Dai JL, et al. Immunohistochemical labeling for dpc4 pancreas development yet critical in progression and tumor biology of pancreas
mirrors genetic status in pancreatic adenocarcinomas: a new marker of DPC4 cancer. Genes Dev. 2006;20:3130–3146.
inactivation. Am J Pathol. 2000;156:37–43.
40. Giampieri S, Manning C, Hooper S, et al. Localized and reversible TGFbeta
26. Iacobuzio-Donahue CA, Song J, Parmiagiani G, et al. Missense mutations of signalling switches breast cancer cells from cohesive to single cell motility.
MADH4: characterization of the mutational hot spot and functional conse- Nat Cell Biol. 2009;11:1287–1296.
quences in human tumors. Clin Cancer Res. 2004;10:1597–1604.
41. Ding Z, Wu C-J, Chu GC, et al. SMAD4-dependent barrier constrains prostate
27. Kanda M, Matthaei H, Wu J, et al. Presence of somatic mutations in most early- cancer growth and metastatic progression. Nature. 2011;470:269–273.
stage pancreatic intraepithelial neoplasia. Gastroenterology. 2012;142:730 e9–
42. Crane CH, Varadhachary GR, Yordy JS, et al. Phase II trial of cetuximab, gemc-
733 e9.
itabine, and oxaliplatin followed by chemoradiation with cetuximab for locally
28. Okano K, Kakinoki K, Akamoto S, et al. 18F-fluorodeoxyglucose positron advanced (T4) pancreatic adenocarcinoma: correlation of Smad4(Dpc4) im-
emission tomography in the diagnosis of small pancreatic cancer. World J munostaining with pattern of disease progression. J Clin Oncol. 2011;29:3037–
Gastroenterol. 2011;17:231–235. 3043.
29. Murakami Y, Uemura K, Sudo T, et al. Long-term results of adjuvant gemc- 43. Lundin J, Nordling S, von Boguslawsky K, et al. Prognostic value of immuno-
itabine plus S-1 chemotherapy after surgical resection for pancreatic carcinoma. histochemical expression of p53 in patients with pancreatic cancer. Oncology.
J Surg Oncol. 2012;106:174–180. 1996;53:104–111.
30. Iacobuzio-Donahue CA, Fu B, Yachida S, et al. DPC4 gene status of the primary 44. Sinicrope FA, Evans DB, Leach SD, et al. bcl-2 and p53 expression in resectable
carcinoma correlates with patterns of failure in patients with pancreatic cancer. pancreatic adenocarcinomas: association with clinical outcome. Clin Cancer
J Clin Oncol. 2009;27:1806–1813. Res. 1996;2:2015–2022.
31. Yachida S, Sakamoto M, Imaida K, et al. p27(Kip1) is overexpressed in very 45. Hermanova M, Karasek P, Nenutil R, et al. Clinicopathological corre-
early stages of hepatocarcinogenesis. Cancer Sci. 2008;99:2152–2159. lations of cyclooxygenase-2, MDM2, and p53 expressions in surgically
32. Wilentz RE, Geradts J, Maynard R, et al. Inactivation of the p16 (INK4A) resectable pancreatic invasive ductal adenocarcinoma. Pancreas. 2009;38:
tumor-suppressor gene in pancreatic duct lesions: loss of intranuclear expres- 565–571.
sion. Cancer Res. 1998;58:4740–4744. 46. Aguirre AJ, Bardeesy N, Sinha M, et al. Activated Kras and Ink4a/Arf de-
33. Winter JM, Cameron JL, Campbell KA, et al. 1423 pancreaticoduodenec- ficiency cooperate to produce metastatic pancreatic ductal adenocarcinoma.
tomies for pancreatic cancer: a single-institution experience. J Gastrointest Genes Dev. 2003;17:3112–3126.
Surg. 2006;10:1199–1210; discussion 1210-1211. 47. Qiu W, Sahin F, Iacobuzio-Donahue CA, et al. Disruption of p16 and activation
34. Tascilar M, Skinner HG, Rosty C, et al. The SMAD4 protein and prognosis of of Kras in pancreas increase ductal adenocarcinoma formation and metastasis
pancreatic ductal adenocarcinoma. Clin Cancer Res. 2001;7:4115–4121. in vivo. Oncotarget. 2011;2:862–873.
35. Gerdes B, Ramaswamy A, Ziegler A, et al. p16INK4a is a prognostic marker 48. Iacobuzio-Donahue CA, Genetic evolution of pancreatic cancer: lessons learnt
in resected ductal pancreatic cancer: an analysis of p16INK4a, p53, MDM2, from the pancreatic cancer genome sequencing project. Gut. 2011;61:1085–
an Rb. Ann Surg. 2002;235:51–59. 1094.
36. Biankin AV, Morey AL, Lee C-S, et al. DPC4/Smad4 expression and outcome 49. Yamao K, Bhatia V, Mizuno N, et al. Interventional endoscopic ultrasonogra-
in pancreatic ductal adenocarcinoma. J Clin Oncol. 2002;20:4531–4542. phy. J Gastroenterol Hepatol. 2009;24:509–519.

346 | www.annalsofsurgery.com 
C 2013 Lippincott Williams & Wilkins

Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

You might also like