SEARCH

SEARCH BY CITATION

Keywords:

  • pancreas cancer;
  • p16;
  • cyclin-dependent kinase inhibitor;
  • tissue microarray

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. KRAS
  6. DISCUSSION
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

BACKGROUND:

Pancreatic cancer is associated with mutations in the tumor suppressor gene cyclin-dependent kinase inhibitor 2A (p16INK4A), a regulator of the cell cycle and apoptosis. This study investigates whether immunohistochemical expression of p16INK4A as well as hypoxia markers and poly adenosine diphosphate-ribose polymerase (PARP) correlates with survival in patients with resected pancreatic adenocarcinoma.

METHODS:

Seventy-three patients with pancreatic adenocarcinoma who underwent curative resection at Stanford University were included. From the surgical specimens, a tissue microarray was constructed using triplicate tissue cores from the primary tumor and used for immunohistochemical staining for the following markers: carbonic anhydrase IX, dihydrofolate reductase, p16INK4A, and PARP1/2. Staining was scored as either positive or negative and percentage positive staining. Staining score was correlated with overall survival (OS) and progression-free survival (PFS).

RESULTS:

Of the markers tested, only immunohistochemical expression of p16INK4A correlated with clinical outcome. On univariate analysis, p16INK4A expression in the tumor was associated with improved OS (P = .038) but not PFS (P = .28). The median survival for patients with positive versus negative p16INK4A staining was 28.8 months versus 18 months. On multivariate analysis, p16INK4A expression was associated with improved OS (P = .026) but not PFS (P = .25). Age (P = .0019) and number of nodes involved (P = .025) were also significant for OS. Adjuvant chemotherapy and margin status did not correlate with OS or PFS.

CONCLUSIONS:

Expression of p16INK4A is associated with improved OS in patients with resected pancreatic adenocarcinoma. Further investigation is needed for validation, given conflicting data in the published literature. Cancer 2010. © 2010 American Cancer Society.

Pancreatic adenocarcinoma is the 10th most common cancer, but the fourth leading cause of cancer death in the United States,1 indicating its high rate of mortality. The molecular pathogenesis of pancreatic adenocarcinoma involves a multistep progression of genetic alterations. Activating mutations in the K-ras gene represents 1 of the early genetic alterations seen in pancreatic adenocarcinoma. In >90% of cases, KRAS is constitutively activated by point mutations, most of which occur in codon 12.2, 3 Late-stage genetic alterations include inactivation of the p53 and DPC4 (SMAD4) genes.2, 3 The p16 tumor suppressor gene (MTS1/INK4A/CDKN2) located on the short arm of chromosome 9 (9p21) is reportedly inactivated in the vast majority (∼95%) of pancreatic adenocarcinomas and is thought to represent an intermediate genetic alteration in the progression of pancreatic adenocarcinoma.2, 3 The p16 gene encodes for the p16INK4A protein. Its principal function is to inhibit the retinoblastoma protein (pRB) phosphorylation by competitive binding to and inactivation of cyclin-dependent kinases 4 and 6, thereby controlling progression through the G1/S transition of the cell cycle. Thus, the genetic inactivation of the p16 gene in pancreatic adenocarcinoma results in loss of a critical regulator of the cell cycle, which may potentiate tumor progression. Immunohistochemical evaluation of p16INK4A expression may be a useful tool for evaluating pancreatic adenocarcinomas.

Tissue hypoxia has been demonstrated to be present in a wide range of human malignancies and is a recognized factor contributing to radiation resistance and probably chemotherapy resistance.4, 5 Mechanistically, the presence of oxygen greatly enhances the efficiency of chromosomal radiation damage by chemically stabilizing radiation-induced free radical production in DNA. In addition, hypoxia is associated with and contributes to an aggressive tumor phenotype, including a predisposition to metastasize.6 It results in the activation of numerous cellular pathways, including those regulated by hypoxia inducible factor 1, resulting in decreased apoptosis, abnormal angiogenesis, and genomic instability among multiple other malignant features.4, 6-9 Historically, tumor hypoxia has been measured invasively by insertion of a polarographic electrode (Eppendorf probe) into the tumor tissue, revealing the presence of radiobiologically significant hypoxia in a range of human malignancies.10-12 Clinical studies have demonstrated a prognostic association between hypoxia and increased risk of locoregional recurrence and distant metastases as well as decreased survival in multiple solid tumor types, notably head and neck, lung, cervical cancers, and soft tissue sarcomas.10-13 In a study by Koong et al, Eppendorf probe measurements directly into pancreas tumors demonstrated the presence of hypoxia.14 A panel of hypoxia tissue markers has been previously shown to correlate with clinical outcome in head and neck cancers, including carbonic anhydrase IX (CAIX) and dihydrofolate reductase (DHFR).15 Thus, immunohistochemical evaluation of markers of tissue hypoxia may be a useful tool for evaluating pancreatic adenocarcinomas.

The current study was undertaken to determine the prognostic value of immunohistochemical expression of various markers, including p16INK4A as well as other potentially important biomarkers, such as markers of tissue hypoxia (DHFR and CAIX). Expression of the DNA repair enzyme poly adenosine diphosphate-ribose polymerase (PARP), a potential therapeutic target,16 was also examined.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. KRAS
  6. DISCUSSION
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

Study Group

The clinicopathologic records of 73 patients with Whipple or radical resection of primary pancreatic adenocarcinoma accessioned at the Department of Pathology, Stanford University Hospital from July 1998 to June 2008 were reviewed. Specifically excluded from this study were patients with metastatic disease, patients with disease arising from the duodenum or ampulla of Vater, and patients who received neoadjuvant therapy. These 73 patients represent all patients who underwent curative resection for pancreatic adenocarcinoma during this time period who met the above inclusion criteria. The pathology reports and hospital charts were reviewed, and the following information was obtained: type of initial surgical procedure and subsequent therapy, margins status, the extent of tumor invasion at initial presentation, and the presence or absence of lymph node metastases. Demographic and intraoperative data were obtained from hospital and clinic charts under the guidelines of the Stanford University Institutional Review Board. Clinical follow-up data were obtained by reviewing patient records, contacting patient families and primary care physicians, and using the Social Security Death Index. Information on adjuvant chemotherapy and/or radiotherapy was also recorded. Follow-up status on living patients was collected through June 2009.

Pathologic Evaluation and Tissue Microarray Construction

For each case, the hematoxylin and eosin (H&E)-stained slides were reviewed by a single pathologist (R.K.P.) to confirm the diagnosis. The following histologic features were recorded for each pancreatic tumor: histologic subtype, tumor grade, extent of invasion, and presence or absence of lymph node metastasis. Each tumor was assigned a histologic subtype according to the World Health Organization classification of pancreatic adenocarcinomas: adenosquamous, undifferentiated (anaplastic), mucinous (colloid), signet ring, foamy gland, and adenocarcinoma not otherwise specified.17 A 3-tiered tumor grading scheme based on the degree of gland formation was also used: well differentiated (>75% gland formation), moderately differentiated (10%-75% gland formation), and poorly differentiated (≪10% gland formation).

The tissue microarrays consisted of 73 pancreatic ductal adenocarcinomas and matched non-neoplastic pancreas from the same patient (Fig. 1). Three tissue microarrays containing a total of 438 tissue cores, each measuring 1.0 mm, were created using a tissue arrayer (Beecher Instruments, Silver Spring, Md) according to a previously described method.18 Representative areas from each case were selected for the microarray from paraffin blocks based on H&E-stained sections.

thumbnail image

Figure 1. (A) A representative photomicrograph shows a tissue microarray (original magnification, ×20). (B) Each patient had 3 representative cores of tumor (Top) and adjacent benign pancreatic tissue (Bottom)(original magnification, ×40).

Download figure to PowerPoint

Immunohistochemical Staining

Immunohistochemical stains were performed on 4-μm-thick sections of paraffin-embedded tissue from each tissue microarray with the following commercially available antibodies, dilutions, and antigen retrieval conditions: p16INK4A (clone E6H4, CINtec, citrate pretreatment, 1:50 dilution), PARP-1/2 (polyclonal; Santa Cruz Biotechnology, Santa Cruz, Calif; ethylenediaminetetraacetic acid [EDTA] pretreatment, 1:150 dilution), an antibody directed at both PARP-1 and PARP-2; CAIX (clone M7519; citrate pretreatment, 1:500 dilution); and DHFR (polyclonal; Santa Cruz Biotechnology; EDTA pretreatment, 1:150 dilution).

All slides were reviewed by a pathologist (R.K.P.) blinded to treatment outcomes. The labeling was scored for extent of staining. Immunolabeling was interpreted as negative (no staining), 1+ (1%-25%), 2+ (26%-75%), or 3+ (>75%), or uninterpretable because of lack of lesional tissue in the tissue core. For p16INK4A, PARP, and DHFR, nuclear and/or nuclear and cytoplasmic staining was considered positive. For CAIX, strong membranous staining was considered positive. For p16INK4A, a paraffin-embedded section of invasive uterine cervical squamous cell carcinoma was included as a positive control for each run. For CAIX and DHFR, paraffin-embedded tissue sections of known-positive head and neck squamous cell carcinomas were used as positive controls for each run.15 For PARP, a paraffin-embedded section of normal lymph node served as a positive control. In addition, negative control serum was applied to all control sections as a negative control.

KRAS Mutation Analysis

Manual microdissection was performed, when required, to exclude overabundance of noncarcinomatous tissues. DNA was extracted from paraffin sections, using the Qiamp DNA mini kit (Qiagen, Chatsworth, Calif) after manual microdissection, when required. Mutant KRAS was detected using a validated KRAS mutation kit (Mutector II, TrimGen, Sparks Glencoe, Md) that identifies somatic missense mutations located in codons 12 and 13. Amplification and detection were performed on an ABI 7900 (Applied Biosytems, Foster City, Calif). The evaluation of Mutector II assay results was performed according to the manufacturer's instructions.

Statistical Analysis

Statistical analyses were performed using SAS (SAS Institute, Cary, NC). The endpoints selected for analysis included progression-free survival (PFS) and overall survival (OS). PFS was defined as the time (measured in months) from the date of initial diagnosis to the date of radiographic or clinical recurrence. OS was defined as the time (measured in months) from the date of initial diagnosis to the date of death or date of last follow-up. Survival rates were determined by the Kaplan-Meier method. Stratifications of p16INK4A staining and selected clinical factors were assessed for statistical significance by the log-rank test statistic. Multivariate analysis was performed using the Cox proportional hazard survival regression to determine the effect of individual predictors present at the time of diagnosis and determined to be of significance on univariate analysis. A 2-sided p-value of .05 in both types of analysis was considered statistically significant.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. KRAS
  6. DISCUSSION
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

The median follow-up from the date of surgery was identical for the entire patient population as well as for the living patients (1.2 years; range, 0.05-8.1). At last follow-up, 22 (30%) patients were alive without evidence of disease, 7 (9.6%) patients were alive with active disease, and 44 (60%) patients had died. As a first site of failure, 7 (9.6%) were locoregional failures, of which only 2 were isolated. Distant metastases were documented in 27 (37%) patients. Of 17 patients who had died, 3 patients died of intercurrent disease, and the cause of death in the remaining 14 patients was unknown.

p16INK4A

Overall, 16 (22%) of 73 cases showed positive p16INK4A staining. Half of these cases (8 of 16, 50%) demonstrated 3+ (>75%) staining within the tumor, with fewer cases showing 1+ (6 of 16, 37.5%) and 2+ (2 of 16, 12.5%) reactivity (Fig. 2). Nine (56%) of the 16 cases showing positive p16INK4A staining showed no staining in the adjacent benign pancreas, and 8 (14%) of 57 cases showing negative p16INK4A staining had positive staining in the benign adjacent pancreas. There was no correlation between p16INK4A staining in the tumor versus the adjacent normal pancreas (P = .81). The clinicopathologic and p16INK4A immunohistochemistry results are shown in Table 1.

thumbnail image

Figure 2. Immunohistochemical expression of p16INK4A in representative core samples of pancreatic adenocarcinoma demonstrates (A) 0, no staining (original magnification, ×400); (B) 1+, 1% to 25% staining (original magnification, ×200); (C) 2+, >25% and ≤75% staining (original magnification, ×200); and (D) 3+, >75% staining (original magnification, ×400).

Download figure to PowerPoint

Table 1. Clinical and Pathologic Features of Pancreatic Adenocarcinomas Stratified by p16INK4A Immunoreactivity
Featurep16INK4A Positive, n=16, No. (%)p16INK4A Negative, n=57, No. (%)P
  • NOS indicates not otherwise specified.

  • a

    These tumors comprised 1 adenosquamous carcinoma and 1 colloid carcinoma.

  • b

    Includes 2 patients who only received intraoperative radiation therapy.

Median age, y (range)67.8 (43.9-80.3)64.4 (38.8-85.6).87
Sex (men:women)8:835:22.40
Histologic type  .61
 Adenocarcinoma NOS14 (88)53 (93) 
 Pure foamy gland2 (12)2 (3) 
 Other02a (3) 
Grade  .27
 Well differentiated4 (25)9 (15) 
 Moderately differentiated6 (38)26 (46) 
 Poorly differentiated6 (38)22 (39) 
Lymph node metastasis  .99
 05 (31)18 (32) 
 14 (25)11 (19) 
 ≥27 (44)28 (49) 
Pathologic T classification  .089
 T12 (12.5)0 (0) 
 T24 (25)11 (19) 
 T310 (62.5)46 (81) 
Margins  .15
 Negative11 (69)23 (40) 
 Close/positive5 (31)34 (60) 
Tumor location  .44
 Head15 (94)47 (82) 
 Body06 (11) 
 Tail1 (6)4 (7) 
Chemotherapy given14 (87.5)40 (70).21
Radiotherapy given12 (75)30b (53).15

On univariate analysis, p16INK4A expression in the tumor was associated with improved OS (P = .038) but not PFS (P = .28). The median OS for patients with positive versus negative p16INK4A staining was 28.8 months versus 18 months. Figure 3A shows the actuarial OS of patients with p16INK4A-positive staining versus p16INK4A-negative staining. The 2-year OS was 86% for patients with p16INK4A-positive tumors versus 35% for patients with p16INK4A-negative tumors (P = .014). The 2-year OS for patients with p16INK4A staining of 2+ or 3+ was 100% versus 39% for those with p16INK4A staining of 0 or 1+ (P = .025)(Fig. 3B). There was no correlation between p16INK4A staining of benign pancreas and OS (P = .81).

thumbnail image

Figure 3. Kaplan-Meier curves show overall survival by (A) positive versus negative p16INK4A staining and (B) degree of staining.

Download figure to PowerPoint

On multivariate analysis, p16INK4A expression again was associated with improved OS (P = .026) but not PFS (P = .25). Other factors that had a statistically significant association with OS were age (P = .0019) and number of nodes involved (P = .025)(Table 2). Adjuvant chemotherapy and margin status did not correlate with OS or PFS. When p16INK4A staining was categorized as 0 or 1+ versus 2+ or 3+, age and number of involved nodes correlated with OS (Table 2). However, p16INK4A staining became borderline significant (P = .08).

Table 2. Multivariate Analysis of Clinicopathologic Features of Pancreatic Adenocarcinoma Stratified by p16INK4A Immunohistochemistry
Clinicopathologic FeatureProgression-Free Survival, POverall Survival, P
p16INK4A staining classified as 0 vs 1-3
 p16INK4A immunohistochemistry (positive vs negative).25.026
 Age (≤65 years vs >65 years).073.0019
 No. of involved lymph nodes (0 vs ≥1).14.025
 Adjuvant chemotherapy.45.54
 Margin status (negative vs close/positive).0001.26
P16INK4A staining classified as 0-1 vs 2-3
 P16INK4A status (0/1 vs 2/3).48.08
 Age (≤65 years vs >65 years).073.005
 No. of involved lymph nodes (0 vs ≥1).14.02
 Adjuvant chemotherapy.48.52
 Margin status (negative vs close/positive).0001.3

Staining of p16INK4A in the normal adjacent pancreas did not correlate with OS or PFS on univariate or multivariate analysis (data not shown).

KRAS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. KRAS
  6. DISCUSSION
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

Thirteen of the 16 cases with p16INK4A positivity had sufficient material available to test for activating missense mutations of KRAS. Ten of 13 (77%) demonstrated missense mutations in codon 12 of the KRAS gene. No cases displayed missense mutations in codon 13 of the KRAS gene. Three cases of pancreatic adenocarcinoma demonstrated wild-type KRAS.

PARP

Thirty-six (49%) patients had positive PARP1/2 staining, of whom 20 (56%) of 36 were 3+, 5 (14%) of 36 were 2+, and 11 (31%) of 36 were 1+. On univariate analysis, positive staining did not correlate with PFS or OS (Table 3).

Table 3. Univariate Analysis of Expression of Tumor Hypoxia Markers and PARP Correlated With Survival
MarkerProgression-Free Survival, POverall Survival, P
  1. PARP indicates poly adenosine diphosphate-ribose polymerase; CAIX, carbonic anhydrase IX.

CAIX.48.83
Dihydrofolate reductase.27.75
PARP1/2.39.93

CAIX and DHFR

Forty-three (59%) patients had positive CAIX staining, of whom 26 (60%) of 43 were 3+, 11 (26%) of 43 were 2+, and 6 (14%) of 43 were 1+. Fifty (68%) patients had positive DHFR staining, of whom 4 (8%) of 50 were 3+, 20 (40%) of 50 were 2+, and 26 (52%) of 50 were 1+. On univariate analysis, positive staining for either hypoxia marker did not correlate with PFS or OS (Table 3).

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. KRAS
  6. DISCUSSION
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

In this study, expression of p16INK4A correlated with improved survival in patients who underwent curative resection for pancreatic adenocarcinoma. This finding was true even after adjusting for other potential confounding factors, including the use of chemotherapy, age, tumor size, margin status, tumor grade, and lymph node status. We believe that this study should prompt validation of these results as well as further investigation into the role of p16INK4A in the development and behavior of pancreatic cancer.

The implications of these results are that p16INK4A could potentially serve as an additional prognostic factor when risk stratifying, either in the setting of a clinical trial or as a tool for selecting therapy (eg, chemoradiation vs chemotherapy). In addition, mutant p16 or compensatory signaling events could represent potential targets for new therapeutics. Given the unfavorable natural history and poor prognosis of this disease, the molecular biology will be crucial in determining areas of future therapeutic research and improvement in outcome.

Other molecular studies have been done to search for candidate biomarkers for diagnostic and prognostic value in pancreatic adenocarcinoma. Proteomic profiling has revealed that different protein expression patterns can differentiate between malignant and benign pancreatic lesions20 as well as predict lymph node spread.21 The molecular biology of a tumor also has other implications on therapy. What has been shown in other cancer sites is that certain genotypes impact response to systemic agents, including targeted agents. For example, colorectal cancers showing mutations in KRAS do not respond to cetuximab therapy.22 Also, tumors showing microsatellite instability may not benefit from 5-fluorouracil–based adjuvant chemotherapy.23 Understanding the mechanism of how molecular pathways determine response to various treatments is a vital step in developing better systemic agents and just as importantly, selecting the appropriate patients for these therapies.

p16 inactivation has been found to be frequent in pancreatic cancer,24, 25 and is thought to play an integral role in the sequence of events in tumorigenesis.26 In the progression model of pancreatic adenocarcinoma, activating KRAS mutations give rise to premalignant ductal lesions known as pancreatic intraepithelial neoplasia (PanIn). Intact p16INK4A is thought to restrain the oncogenic potential of these activating KRAS mutations. In mouse models, activating K-ras mutations associated with inactivation of the p16 gene give rise to an earlier appearance of PanIn lesions within mice and rapid progression to highly invasive carcinomas with metastatic disease.27, 28 In addition, Wilentz and colleagues found that loss of p16INK4A expression by immunohistochemistry correlated with higher grade ductal lesions.29 p16INK4A inactivation in pancreatic adenocarcinoma can occur by 3 different mechanisms: 1) homozygous deletion (40%), 2) loss of heterozygosity of 1 allele followed by intragenic mutation of the second allele (40%), and 3) hypermethylation of the gene promoter resulting in complete gene silencing (15%).24, 30-32 The scenarios of homozygous p16 gene deletion and p16 gene promoter methylation should result in complete absence of the p16INK4A protein. In contrast, loss of heterozygosity of 1 allele followed by intragenic mutation of the second allele may result in expression of a truncated functionally deficient p16INK4A protein.24, 33 A familial syndrome associated with pancreatic cancer, familial atypical multiple melanoma, is caused by a germline mutation in the p16INK4A gene.34

Our results indicate that those pancreatic adenocarcinomas that retain p16INK4A immunohistochemical expression exhibit improved OS. On the basis of what is known of its mechanism of action, it can be hypothesized that the improved OS for tumors expressing p16INK4A by immunohistochemistry may be because of the preserved tumor suppressor function of this protein that slows cell cycling and allows for more indolent behavior. Indeed, overexpression of cyclin E, another cell cycle regulator that promotes transition from G1 to S, has been shown to correlate with poor prognosis in both breast cancer35 and pancreatic cancer.36 Further studies to validate this hypothesis are warranted.

Our rate of p16INK4A positivity by immunohistochemistry in pancreatic adenocarcinomas correlates well with published literature reports.37, 38 Kawesha et al found that 21 (13%) of 157 pancreatic adenocarcinomas expressed p16INK4A by immunohistochemistry, and Biankin et al found expression of p16INK4A in >5% of tumor cells in 30 (31%) of 98 pancreatic adenocarcinomas.37, 38 In both studies, the pattern of p16INK4A reactivity did not correlate with survival. Both studies also showed that abnormal p53, p21WAF1/CIP1, or cyclin D1 expression did not correlate with survival. However, Hu et al showed that loss of p16INK4A correlated with worse survival and higher rate of distant metastases.39 Similarly, Gerdes et al showed that altered p16INK4A was associated with worse survival.40 Naka et al showed that patients with loss of p16INK4A had worse survival and more advanced stage.16 The latter finding was also shown in a study reported by Jeong et al.41 The difference between these studies is not readily apparent, although it may be related to different p16INK4A immunohistochemical antibody clones used in each study. Certainly the mixed results indicate that further investigation into the exact role of p16INK4A is needed, and the current study can generate newer hypotheses. They also serve as an important cautionary note about relying on any single marker to predict clinical outcome.

It is unclear based on our results whether the p16INK4A-positive pancreatic adenocarcinomas in this series express truncated p16INK4A protein as a result of intragenic mutation of the p16 gene. Importantly, it is known that some p16INK4A monoclonal antibodies may not be able to reliably distinguish between normal and mutant p16INK4A protein and display cross-reactivity with mutant p16INK4A.33 Thus, nuclear staining with anti-p16INK4A antibody by immunohistochemistry cannot necessarily be interpreted as the presence of normal p16INK4A. This point is an important weakness of this study. Further genetic analysis of our subset of p16INK4A-positive pancreatic adenocarcinomas is necessary to determine whether these carcinomas harbor germline or mutant p16 genes. Salek et al found that loss of heterozygosity of the 9p loci, which harbors the p16INK4A gene, did not correlate with survival in patients with unresectable disease.42 In addition, Blackford et al recently showed that inactivation of CDKN2A did not correlate with survival.43

In our study, mutant KRAS was found in 77% of tumors that were p16INK4A positive, which is consistent with the 75% found by Kawesha et al and other data showing the high rate of KRAS mutations in this disease.26, 37 Therefore, incorporation of KRAS into the multivariate model was not done. The prognostic value of KRAS is uncertain, with a recent study suggesting no correlation with survival,42 whereas other investigators have found that the type of KRAS mutation was prognostic.37

An additional weakness of this study is the relatively small patient numbers. The differences in clinical factors between p16INK4A-positive and -negative patients, although not statistically significant, do favor the p16INK4A-positive group, which could confound these results. We believe that these results should be independently validated with a larger dataset to ensure well-balanced patient groups.

Immunohistochemical analysis with CAIX and DHFR, markers of tumor hypoxia, did not correlate with survival. Hypoxia has been demonstrated to be present in pancreatic cancer.14 It was, therefore, hypothesized that hypoxic markers could yield prognostic information, given the known worse outcomes associated with tumor hypoxia in other disease sites.8, 44 Further investigation is needed to determine its importance in pancreatic cancer and whether other markers may have prognostic significance.

Finally, PARP1/2 was investigated in this study. PARP is an enzyme that functions in DNA repair and is thought to counteract the effects of standard cytotoxic agents.45 Research is currently directed toward developing inhibitors of PARP that could potentiate the effects of therapy. Currently, several agents are being developed and investigated in solid tumors, notably breast cancer.46 Although this study failed to show correlation of PARP and expression and survival, further investigation is also needed to determine the role of PARP in pancreatic cancer and whether it represents a potential target for therapy.

Conclusions

In conclusion, expression of p16INK4A by immunohistochemistry was associated with improved survival in patients with resected pancreatic adenocarcinoma. Further investigation is needed to validate these findings in a larger cohort of patients and to understand the role of p16INK4A in determining the phenotypic behavior of these tumors for both natural history and response to systemic and/or radiation therapy.

CONFLICT OF INTEREST DISCLOSURES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. KRAS
  6. DISCUSSION
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

Supported by Liu Bie Ju Cha and Family Fellowship in Cancer.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. KRAS
  6. DISCUSSION
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES