Immunohistochemical analysis of IMP3 and p53 expression in endoscopic ultrasound-guided fine needle aspiration and resected specimens of pancreatic diseases

Pancreatology xxx (2017) 1e8

Contents lists available at ScienceDirect

Pancreatology journal homepage: www.elsevier.com/locate/pan

Immunohistochemical analysis of IMP3 and p53 expression in endoscopic ultrasound-guided fine needle aspiration and resected specimens of pancreatic diseases Junichi Senoo a, Rintaro Mikata a, *, Takashi Kishimoto b, Masahiro Hayashi a, Yuko Kusakabe a, Shin Yasui a, Mutsumi Yamato a, Hiroshi Ohyama a, Harutoshi Sugiyama a, Toshio Tsuyuguchi a, Hideyuki Yoshitomi c, Masayuki Ohtsuka c, Juri Maeda d, Satoshi Ota d, Yukio Nakatani d, Naoya Kato a a

Department of Gastroenterology, Chiba University Graduate School of Medicine, Japan Department of Molecular Pathology, Chiba University Graduate School of Medicine, Japan c Department of General Surgery, Chiba University Graduate School of Medicine, Japan d Department of Diagnostic Pathology, Chiba University Graduate School of Medicine, Japan b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 2 July 2017 Received in revised form 21 November 2017 Accepted 20 December 2017 Available online xxx

Background: Insulin-like growth factor II messenger ribonucleic acid-binding protein 3 (IMP3) is a valuable marker that distinguishes malignant from benign lesions and predicts prognosis. Methods: First, we evaluated IMP3 expression in 77 resected specimens of pancreatic ductal adenocarcinoma (PDAC), intraductal papillary mucinous neoplasm (IPMN), and chronic pancreatitis (CP). Eleven PDAC patients preoperatively underwent endoscopic ultrasound-guided fine needle aspiration (EUSFNA). Survival analysis of IMP3 and clinicopathological factors was performed. IMP3 and p53 expression was evaluated in another 127 EUS-FNA samples of solid pancreatic masses to compare the diagnostic value of routine and immunohistochemical staining. Results: IMP3 expression was detected in 72.3%, 50%, 20%, and 0% of PDAC, malignant IPMN, benign IPMN, and CP, respectively. Evaluation of IMP3 expression in EUS-FNA specimens coincided with that in resected specimens in 10 of 11. IMP3 expression correlated with tumor differentiation in PDAC samples (p ¼ .006) and with poor prognosis through univariate analysis (p ¼ .045). Tumor differentiation and lymph node metastasis were significantly associated with poor prognosis through multivariate analysis. In EUS-FNA specimens, the sensitivity, specificity, and accuracy of cytohistological analysis were 80.8%, 100%, and 85.0%, respectively. IMP3 and p53 expression were detected in 80.8% and 44.9% of malignant and 0% and 5% of benign lesions. Combined with IMP3 immunostaining, the sensitivity, specificity and accuracy of cytohistological analysis significantly increased to 87.9%, 100%, and 90.8% (p ¼ .016), respectively. Meanwhile, p53 staining had no impact on the results. Conclusions: IMP3 immunohistochemical staining can improve the diagnostic accuracy of EUS-FNA for malignant pancreatic tumors. © 2017 Published by Elsevier B.V. on behalf of IAP and EPC.

Keywords: Pancreatic cancer IMP3 p53 EUS-FNA Immunohistochemical analysis

Introduction Pancreatic ductal adenocarcinoma (PDAC) is one of the most aggressive malignant tumors with a 5-year survival rate of only 4%

* Corresponding author. Department of Gastroenterology, Graduate School of Medicine, Chiba University, 1-8-1 Inohana, Chuo-ku, Chiba 260-8670, Japan. E-mail address: [email protected] (R. Mikata).

[1]. Therefore, pathological diagnosis of malignancy before treatment is important in suspected cases. It is reported that endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) is a safe and effective technique for diagnosis of pancreatic tumors with a sensitivity of 78%e95% [2e5]. However, specimens obtained by EUS-FNA are very small, thus differential diagnosis of malignant disease from benign disease is sometimes challenging. To improve the sensitivity and accuracy of EUS-FNA for pancreatic tumors, various type of puncture needles, puncture techniques, and

https://doi.org/10.1016/j.pan.2017.12.010 1424-3903/© 2017 Published by Elsevier B.V. on behalf of IAP and EPC.

Please cite this article in press as: Senoo J, et al., Immunohistochemical analysis of IMP3 and p53 expression in endoscopic ultrasound-guided fine needle aspiration and resected specimens of pancreatic diseases, Pancreatology (2017), https://doi.org/10.1016/j.pan.2017.12.010

2

J. Senoo et al. / Pancreatology xxx (2017) 1e8

molecular biological approaches, including immunohistochemical staining and detection of gene mutations, have been investigated [6e11]. Immunohistochemical markers to differentiate between malignant and benign tissue would be beneficial. Insulin-like growth factor II messenger ribonucleic acid (mRNA) binding protein 3 (IMP3) is an oncofetal protein that is expressed during embryogenesis and is almost silenced in normal mature tissues. IMP3 may play an important role in mRNA trafficking and stabilization, localization, cell growth, and cell migration [12]. IMP3 is expressed in many malignant neoplasms, including renal cell carcinoma [13], ovarian carcinoma [14], urothelial carcinoma [15], lung carcinoma [16], colorectal cancer [17], and PDAC [18e24]. It was recently reported that IMP3 overexpression in PDAC was significantly associated with poor prognosis and IMP3 may be potential therapeutic target for PDAC. A few studies have shown that IMP3 expression in PDAC is a valuable marker to distinguish this entity from a benign lesion using histological samples obtained with a core biopsy needle or cytological samples obtained with a FNA needle [11,18]. However, it remains unclear whether cytohistological analysis combined with IMP3 expression significantly improved the diagnostic value, and there has been no report on the diagnostic usefulness of IMP3 expression in histological samples obtained with conventional FNA needles. The p53 tumor suppressor gene is significantly mutated in many tumors, including human pancreatic cancer (38.2e81.1%). Hence, immunohistochemical staining of p53 might improve the diagnostic accuracy of PDAC [25e30]. The aims of the present study were to evaluate IMP3 expression in surgical resected specimens by immunohistochemical staining and analyze the relationship between clinicopathological factors and prognosis of PDAC. In addition, IMP3 expression was examined in paired EUS-FNA specimens and 11 resected PDAC specimens. Finally, the diagnostic values of IMP3 and p53 in EUS-FNA specimens of solid pancreatic masses were evaluated. Materials and methods Case selection Large tissue sections were randomly collected from patients with pancreatic lesions, including 47 with PDAC, 10 with malignant intraductal papillary mucinous carcinoma (IPMN), 10 with benign IPMN, and 10 with chronic pancreatitis (CP) who underwent pancreatic resection from June 2010 to July 2013 at Chiba University Hospital (Chiba, Japan). In this study, IPMN was classified into four groups: IPMN with low-grade dysplasia (LGD), intermediatedgrade dysplasia (IGD), high-grade dysplasia (HGD), and an associated invasive carcinoma, according to the World Health Organization classification [31]. Malignant IPMN was defined as HGD and invasive carcinoma. Benign IPMN was defined as LGD and IGD. Among the resected samples, 11 PDAC samples were collected by EUS-FNA before surgery. First, to evaluate IMP3 expression in large tissue sections, IMP3 staining was performed for all 77 resected specimens. Next, IMP3 expression was evaluated in 11 EUS-FNA specimens and correlated to IMP3 expression of PDAC resected specimens. Moreover, an additional 127 EUS-FNA samples of solid pancreatic masses were collected from 122 patients (83 men, 39 women; mean age, 65.8 ± 10.4 years; age range, 36e84 years) between June 2013 and June 2015 in our institution. The study protocol was approved by the institutional ethics review board of Chiba University Hospital (approval number: #242, June 29th, 2016).

Resection specimens Seventy-seven resected specimens were histologically examined according to the 7th TNM classification system [32]. In PDAC, we evaluated the relationship between IMP3 expression and clinicopathological factors, including sex, age, tumor location, tumor size, residual tumor classification, lymph node metastasis, tumor differentiation, and expression levels of carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA19-9). Also, the relationships between IMP3 expression and clinicopathological factors and prognosis were evaluated. Immunohistochemical staining of resected specimens Paraffin blocks were sliced into 4 mm-thick sections. After deparaffinization and rehydration, heat-induced antigen retrieval was performed by immersing the sections in high pH buffer solution (pH 9.0) at 97  C for 20 min. Then, endogenous peroxidase activity was blocked by incubation with blocking agent for 5 min at room temperature. Afterward, the sections were incubated for 20 min at room temperature with a mouse monoclonal antibody against IMP3 (clone 69.1; Dako, Glostrup, Denmark) at a dilution of 1:100. The primary antibodies were detected by incubating the sections with polymer reagent for 20 min at room temperature. The sections were subjected to color development with 3,30 -diaminobenzidine and counterstained with hematoxylin. Normal placental tissue was used as a positive control for IMP3 immunostaining. All slides of the resected specimens stained with hematoxylin and eosin (HE) and immunohistochemical slides were reviewed by the same experienced pathologist. Specimens obtained by EUS-FNA Most patients underwent EUS-FNA for pathological diagnosis of pancreatic tumors before surgery or chemotherapy. EUS-FNA was performed with a 19, 22, or 25 gauge needle and an average of 3.1 ± 1.0 passes per session (Table 3). The aspirate biopsy materials were pushed onto filter paper by re-inserting the stylet and into formalin-filled container for histological analysis. The residual material was smeared onto a glass slide by air pressure and fixed with 95% ethanol for cytological analysis. Biopsy specimens fixed in formalin were embedded in paraffin and stained with HE for histological analysis. By histological analysis, the specimens were classified into one of five types: malignancy, suspected malignancy, atypical cells, no evidence of malignancy, or insufficient material. In this study, specimens classified as malignancy and suspected malignancy by HE staining were defined as positive for malignancy, and those classified as atypical cell, no evidence of malignancy, and insufficient materials were defined as negative for malignancy. The final diagnosis was based on the following: (i) histological analysis of surgically resected specimens; (ii) cytopathological detection of cancer cells from other organs coupled with clinical and/or radiological evidence of progressive disease, and benign lesion was diagnosed by surgically resected specimen or clinical follow-up for at least 6 months, with no evidence of progressive disease. Immunohistochemical staining of EUS-FNA specimens Of 127 consecutive cases, 75 sections were stained with HE and immunohistochemically stained with anti-IMP3 and anti-p53 antibodies at the same time, while 52 sections immunohistochemically stained against IMP3 and p53 were prepared later. The primary antibodies used for IMP3 immunohistochemistry were mouse monoclonal anti-human IMP3 (clone 69.1; Dako) at a

Please cite this article in press as: Senoo J, et al., Immunohistochemical analysis of IMP3 and p53 expression in endoscopic ultrasound-guided fine needle aspiration and resected specimens of pancreatic diseases, Pancreatology (2017), https://doi.org/10.1016/j.pan.2017.12.010

J. Senoo et al. / Pancreatology xxx (2017) 1e8

dilution 1:100 and mouse monoclonal anti-human p53 protein (Clone DO7; Dako) at a dilution 1:50. Evaluation of immunohistochemical staining IMP3 staining was evaluated using a 4-tiered system for staining intensity as described by Riener et al. [33]. Score 0 was assigned if cells were the staining was completely negative, score 1 þ was assigned if at least 10% of the cells stained weakly, score 2 þ was assigned if at least 10% of the cells stained moderately, and score 3 þ for strong cytoplasmic staining in 10% of cells (Fig. 1A). Scores of 0 or 1 þ were considered negative for IMP3 and scores of 2 þ or 3 þ were considered positive. Abnormal p53 labeling was defined as positive nuclear staining of at least 60% of tumor cells (Fig. 1B) [7,25]. Quantification of EUS-FNA samples The amount of samples obtained by EUS-FNA was assessed using the modified scoring system described by Gerke et al. [34]. The scoring system was as follows: 0, an insufficient amount of material; 1, sufficient material for limited cytological interpretation: not representative; 2, sufficient material for adequate cytological interpretation; 3, sufficient material for limited histological interpretation; and 4, sufficient material for adequate histological interpretation. The relationship between immunohistochemical expression and tissue quantity was evaluated. Statistical analysis Statistical analysis was conducted using SPSS ver. 22 software (IBM-SPSS, Inc. Chicago, IL, USA). Quantitative date are presented as the mean ± standard deviation (SD). In surgically resected specimens, immunohistochemical IMP3 expression and clinicopathological factors were compared using the chi-square test. Survival analysis was performed based on clinicopathological factors and IMP3 expression. The duration of overall survival (OS) was defined as time from the day of surgery to the date of death. Survival distributions were estimated using the KaplaneMeier method and the significance of differences between survival rates was calculated using the log-rank test. OS was also compared between groups

3

using Cox proportional-hazards regression from which hazard ratios (HRs) were estimated. For analysis of the EUS-FNA specimens, the McNemar test was applied to assess the statistical significance of sensitivity, specificity, positive predict value (PPV), negative predict value (NPV), and accuracy comparing cytology and histology alone and the combination of cytology, histology and immunohistochemical staining. A probability (p) value of <0.05 was considered statistically significant. Results Resection specimens IMP3 expression was evaluated in 77 resected sections of pancreatic tumors and CP. IMP3 expression was detected in 34 (72.3%) of 47 PDAC specimens, 5 (50%) of 10 malignant IPMN specimens, and 2 (20%) of 10 benign IPMN specimens. None of the 10 CP cases was positive for IMP3 expression. IMP3 expression in PDAC was significantly greater than that in benign IPMN (p < .01), and CP (p < .01), and tended to be greater when compared with malignant IPMN (p ¼ .11). The characteristics of the 47 PDAC patients are shown in Table 1. The mean age of these patients was 65.5 ± 10.8 (range, 32e83) years. The tumor location was the pancreatic head in 33 patients and body or tail in 14 patients. The mean tumor size was 31.3 ± 11.2 (range, 12.0e59.0) mm. Adjuvant chemotherapy was administered in 39 patients with PDAC, 21 patients were administered gemcitabine chemotherapy, seven received S-1 chemotherapy, and 11 received a combination of gemcitabine and S-1 chemotherapy. Correlation between IMP3 expression and clinicopathological factors IMP3 staining was positive in 34 (72.3%) of 47 PDAC specimens. Of these 47 specimens, three were weakly positive for IMP3 expression, four were moderately positive, and 30 were strongly positive. The proportion of positive tumor cells was >80% in 23 (67.6%) specimens, 50%e80% in 8 (23.5%) specimens, and 10%e50% in 3 (8.8%) specimens of these 34 PDAC specimens. Analysis of IMP3 expression and the clinicopathological parameters showed a

Fig. 1. A Representative images of immunohistochemical IMP3 expression. (A). Negative. (B). Weakly stained (score 1þ). (C). Moderately stained (score 2þ). (D). Strongly stained (score 3þ). B. Representative images of immunohistochemical p53 expression. (A). Negative p53 expression. (B). Overexpression of p53 and positive staining of 60% of tumor cell nuclei.

Please cite this article in press as: Senoo J, et al., Immunohistochemical analysis of IMP3 and p53 expression in endoscopic ultrasound-guided fine needle aspiration and resected specimens of pancreatic diseases, Pancreatology (2017), https://doi.org/10.1016/j.pan.2017.12.010

4

J. Senoo et al. / Pancreatology xxx (2017) 1e8

Table 1 Correlation between IMP3 expression and clinicopathological factors. Clinicopathological factor Sex Male Female Age (years) <60 >60 Location of mass Head body or tail Size of lesion (mm) <20 >20 Residual tumor classification R0 R1 Lymph node metastasis Yes No Tumor differentiation Well Moderately Poorly CEA 5.2 ng/ml >5.2 ng/ml CA19-9 36.8 ng/ml >36.8 ng/ml

PDAC (N ¼ 47)

IMP3 positive (N ¼ 34)

IMP3 positive rate (%)

22 25 65.5 ± 10.8 (32e83) 14 33

15 19

68.9 76.9

11 23

78.9 69.7

33 14 31.3 ± 11.2 (12.0e59.0) 5 42

24 10

72.7 71.4

3 31

60.0 73.8

31 16

20 14

64.5 87.5

33 14

25 9

75.8 64.3

16 26 5

7 22 5

43.8 84.6 100.0

38 9

28 6

73.7 66.7

11 36

9 25

81.8 69.4

p .560

.544

.929

.524

.099

.432

.01※

.893

.774

CEA: carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9. CA19-9: carbohydrate antigen 19-9.

significant positive correlation between IMP3 expression and tumor differentiation (p ¼ .006). IMP3 expression was more frequently detected in moderately or poorly differentiated tumors, as compared to well-differentiated tumors. There were no significant correlations between IMP3 expression and sex, age, tumor size, tumor location, residual tumor classification, lymph node metastasis, CEA values, and CA19-9 values (Table 1). The median postoperative follow-up time was 27.0 ± 18.4 months (range, 2.9e73.1). At the time of evaluation in September 2016, 31 (66.0%) of the 47 patients included in this study had died and 5 (10.6%) were lost to follow-up. Univariate analysis showed significant relationships between OS and IMP3 expression (p ¼ .045), residual tumor classification (p ¼ .020), tumor differentiation (p ¼ .25), and lymph node metastasis (p ¼ .01) (Table 2). Prognosis of patients with IMP3-positive, moderately or poorly differentiated tumors, and R1 or positive lymph node metastasis

was significantly poor (Fig. 2). There was no significant relationship between OS and sex, age, tumor location, tumor size, CEA values, and CA19-9 values. In multivariate analysis, lymph node metastasis (HR ¼ 6.215; 95% confidence interval [CI] ¼ 2.104e18.359) and tumor differentiation (HR ¼ 2.970; 95% CI ¼ 1.247e7.069) were significantly associated with survival (Table 2). Concordant EUS-FNA specimens and resected specimens Analysis of 11 EUS-FNA samples from patients with PDAC before surgery showed that IMP3 expression was concordant between resected specimens and EUS-FNA specimens in 10 (90.9%) samples. Results of EUS-FNA routine staining Of 127 EUS-FNA specimens of solid pancreatic masses, 99

Table 2 Univariate and multivariate analyses of poor prognostic factors in patients with pancreatic cancer. Variables

Category

Univariate analysis

Multivariate analysis

HR

95% CI

P

Sex Age (years) Tumor Location Tumor Size (mm) Residual tumor classification Lymph node metastasis Tumor differentiation IMP3 CEA (ng/ml) CA19-9 (U/ml)

man/female >60 vs. 60 Head/body or tail >20 vs. &20 R1 vs. R0 Yes vs. No well vs. moderately/poorly positive vs. negative >5.2 vs. 5.2 >36.8 vs. 36.8

1.192 0.787 0.635 3.392 2.343 5.643 2.641 0.375 2.049 1.721

0.587e2.422 0.333e1.651 0.291e1.388 0.802e14.344 1.141e4.812 1.939e16.422 1.129e6.177 0.143e0.980 0.931e4.510 0.740e3.999

.626 .527 .255 .97 .020* .001* .025* .045* .075 .207

HR

95% CI

p

0.02 4.27

2.104e18.359 1.247e7.069

.001* .014*

* P < .05. IMP3: Insulin-like growth factor II messenger ribonucleic acid binding protein 3. CEA: carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9. CA19-9: carbohydrate antigen 19-9.

Please cite this article in press as: Senoo J, et al., Immunohistochemical analysis of IMP3 and p53 expression in endoscopic ultrasound-guided fine needle aspiration and resected specimens of pancreatic diseases, Pancreatology (2017), https://doi.org/10.1016/j.pan.2017.12.010

J. Senoo et al. / Pancreatology xxx (2017) 1e8

5

Table 3 Patient Characteristics for EUS-FNA

Male/Female (n ¼ 122) Mean age (years) Location of mass (head/body or tail) Size of lesion (mm) Puncture route (stomach/duodenum/other) Mean number of needle passes Puncture needle (19G/22G/25G)

83/39 65.8 ± 10.4 (36e84) 46/81 26.4 ± 12.3 (7.7e80) 94/36/1 3.1 ± 1.0 6/79/55

Fig. 2. KaplaneMeier curves of overall survival (OS) in patients with pancreatic cancer. (A). The median OS duration of the IMP3-positive and -negative groups was 696.0 and 986.6 days, respectively (log-rank test; p ¼ .030). (B). The median OS duration of the R0 and R1 groups was 888.7 and 558.8 days, respectively (log-rank test; p ¼ .011). (C). The median OS duration of the well differentiation and mod/por differentiation groups was 996.8 and 678.1 days, respectively (log-rank test; p ¼ .020). (D). The median OS duration of N0 and N1groups was 1195.7 and 598.5 days, respectively (log-rank test; p ¼ .027).

(78.0%) were malignant and 28 (22.0%) were benign. The malignant lesions included 83 PDACs, five neuroendocrine tumors (NETs) including one G1 (Ki  2%), three G2 (Ki-67 6.3%, 6.4%, and 8.3% respectively), and one G3(Ki 25.2%), four solid-pseudopapillary neoplasms (SPNs), three malignant IPMNs, three adenosquamous cell carcinomas, and one metastatic tumor of esophageal cancer. The benign lesions included 20 of autoimmune pancreatitis, 4 CPs, two suspicious of serous cystic neoplasm, one serous cystic neoplasm diagnosed by surgery, one suspicious of epidermoid cyst. Cytological analysis was performed for 127 specimens. The sensitivity, specificity, and accuracy cytological analysis alone were 64.6%, 100%, and 72.4%, respectively. Histological analysis was performed for 125 specimens. The sensitivity, specificity, and accuracy were 69.1%, 100%, and 76.0%, respectively. The sensitivity, specificity, and accuracy for combined cytological and histological analyses were 80.8%, 100% and 85.8%, respectively (Table 4A).

IMP3 and p53 immunohistochemical staining of EUS-FNA specimens Samples from 125 of 127 cases were subjected to histological examination, but 10 samples could not be histologically examined owing to insufficient material. In addition, a further 10 samples could be histologically examined, but their additional sliced sections did not contain sufficient material to perform immunohistochemistry of IMP3 staining. Therefore, immunohistochemical analysis was performed in 105 of 125 specimens, which included 85 malignant and 20 benign tumors. Of the 85 specimens, 43 patients underwent surgery, 41 were diagnosed by cytological detection of malignant cell with clinical follow up. Malignant tumor included 70 PDAC, four NET, three SPN, three adenosquamous carcinoma, three malignant IPMN, one metastatic pancreatic tumor of esophageal cancer, and one case diagnosed as malignancy by clinical follow up. Of the 20 benign tumors, only one patient who was diagnosed with serous cystic neoplasm underwent surgery.

Please cite this article in press as: Senoo J, et al., Immunohistochemical analysis of IMP3 and p53 expression in endoscopic ultrasound-guided fine needle aspiration and resected specimens of pancreatic diseases, Pancreatology (2017), https://doi.org/10.1016/j.pan.2017.12.010

6

J. Senoo et al. / Pancreatology xxx (2017) 1e8

Table 4A Results of EUS-FNA (All case).

Histology Cytology Histology Histology Histology Histology Histology

þ þ þ þ þ

Cytology IMP3 p53 Cytology þ IMP3 Cytology þ p53

Sensitivity (%)

Specificity (%)

PPV(%)

NPV(%)

Accuracy (%)

69.1 64.6 80.8 83.5* 79.8 87.9** 80.8

100 100 100 100 96.4 100 96.4

100 100 100 100 100 100 98.8

48.2 44.4 56.9 63.6 48.1 70.0 58.7

76.0 72.4 85.0 85.8* 83.4 90.6** 84.3

*

P < .05: Histology vs Histology þ IMP3. P < .05: Histology þ Cytology vs Histology þ Cytology þ IMP3 P < .05. EUS-FNA: endoscopic ultrasound-guided fine needle aspiration. PPV: positive predictive value. NPV: negative predictive value. **

Table 4B Results of EUS-FNA (PDAC and benign diseases).

Histology Cytology Histology Histology Histology Histology Histology

þ þ þ þ þ

Cytology IMP3 p53 Cytology þ IMP3 Cytology þ p53

Sensitivity (%)

Specificity (%)

PPV(%)

NPV(%)

Accuracy (%)

70.3 69.9 81.9 86.4* 75.3 90.4** 81.9

100 100 100 100 96.4 100 96.4

100 100 100 100 98.4 100 98.8

53.8 52.8 65.1 71.8 57.4 77.8 64.3

78.0 77.5 86.5 89.9* 81.7 92.8** 86.5

*

P < .05: Histology vs Histology þ IMP3. P < .05: Histology þ Cytology vs Histology þ Cytology þ IMP3 P < .05. EUS-FNA: endoscopic ultrasound-guided fine needle aspiration. PPV: positive predictive value. NPV: negative predictive value. **

IMP3 positive expression was observed in 60 (85.7%) of 70 specimens of PDAC. One of three NET cases that had liver metastasis showed IMP3-positive expression and the other two NET showed IMP3-negative expression. No SPN case was IMP3 positive. Two of three adenosquamous carcinoma were IMP3 positive. IMP3 expression was negative in all specimens of benign lesions. P53 overexpression was observed in 33 (47.1%) of 70 PDAC, two adenosquamous carcinoma, two cases diagnosed as malignancy by clinical course, and one (5%) benign lesion. The sensitivity, specificity, PPV, NPV and accuracy of the combination of cytohistology and IMP3 in all cases were 87.9%, 100%, 100%, 70.0%, and 90.6%, respectively (Table 4A). These results in PDAC and benign diseases were 90.4%, 100%, 100%, 77.8%, and 92.8%, respectively (Table 4B). The sensitivity and accuracy of cytohistological analysis combined with IMP3 staining were significantly greater than that of cytohistology alone in both all cases and in PDAC and benign diseases. The sensitivity, specificity, PPV, NPV, and accuracy of the combination of cytohistology and p53 immunostaining in all cases were 80.8%, 96.4%, 98.8% 58.7%, and 84.3%, respectively. These results in PDAC and benign diseases were 81.9%, 96.4%, 98.8% 64.3%, and 86.5%, respectively. The sensitivity, specificity, and accuracy of cytohistological analysis combined with p53 staining were not significantly different from cytopathology alone. In PDAC, among the specimens diagnosed as malignancy or suspected malignancy by HE staining, 87.5% (35/40) were positive for IMP3 staining. In addition, 12 (75.0%) of 16 samples diagnosed as atypical cells were positive for IMP3. On the other hand, only 4 (25.0%) of 16 samples diagnosed as atypical cell were positive for p53. Analysis for the relationship between immunohistochemical staining and tissue amount in PDAC showed that IMP3 positivity in 36 of 39 (92.3%) grade 4 specimens, 19 of 23 (82.6%) grade 3 specimens, and 5 of 8 (62.5%) grade 2 specimens. In contrast, p53

Table 5 Results of immunohistochemical staining and tissue amount in PDAC. Score

N

Immunohistohemical staining

IMP3

P53

Positive

%

Positive

%

1 2 3 4

6 11 24 40

0 8 23 39

e 5 19 36

e 62.5 82.6 92.3

e 3 7 23

e 37.5 30.4 59.0

Total

81

70

60

84.5

33

46.8

positivity was observed in 23 of 39 (59.0%) grade 4 specimens, 7 of 23 (30.4%) grade 3 specimens, and 8 (37.5%) grade 2 specimens (Table 5). Discussion The results of this study showed that IMP3 expression was significantly correlated with tumor differentiation in resected specimens. Prognosis of IMP3-positive patients was worse than that of IMP3-negative patients. In multivariate analysis, tumor differentiation and lymph node metastasis were significantly correlated with OS. In EUS-FNA samples, the sensitivity, specificity, and accuracy of cytohistological analysis combined with IMP3 staining were significantly higher than that of cytohistology alone, revealing that IMP3 staining was a valuable marker for differentiation of malignant from benign pancreatic disease. The overexpression of IMP3 in large tissue section obtained by surgery was 72.3% for PDAC, 50% for malignant IPMN, 20% for benign IPMN, and none for CP. As the overexpression of IMP3 was observed in benign IPMN, IMP3 positive is not identical with malignancy. However, in the cases of solid pancreatic masses,

Please cite this article in press as: Senoo J, et al., Immunohistochemical analysis of IMP3 and p53 expression in endoscopic ultrasound-guided fine needle aspiration and resected specimens of pancreatic diseases, Pancreatology (2017), https://doi.org/10.1016/j.pan.2017.12.010

J. Senoo et al. / Pancreatology xxx (2017) 1e8

positivity of IMP3 strongly suggests malignancy. The majority of PDAC specimens showed moderate to strong staining with diffuse staining patterns, which was similar to the findings of a previous report [18]. These results indicated a low false-negative rate for IMP3 staining. Therefore, IMP3 may be useful as immunohistochemical marker for small samples obtained by EUS-FNA. Actually, in this study, evaluation of IMP3 expression in 10 of 11 EUS-FNA samples was the same as that of resected specimens. In addition, although islet of Langerhans sometimes showed weak staining in benign lesions, tumor cells showed moderate to strong staining. Therefore, it is not difficult to assess whether IMP3 staining is positive or negative. However, caution must be taken when evaluating section negative for IMP3 expression because malignant lesions are not always IMP3-positive. In this study, 28% of PDAC and 50% of malignant IPMN specimens were negative for IMP3 expression because of low positive rate in well-differentiated adenocarcinoma. There was a significant correlation between IMP3 expression and tumor differentiation. Of the well-differentiated tumors, 7 of 16 (43.7%) were positive for IMP3 expression. On the other hand, 27 of 31 (87.1%) in the moderately and poorly differentiated tumors were positive for IMP3 expression (Table 2). However, the relationship between IMP3 and tumor differentiation is controversial. Schaeffer et al. showed that IMP3 overexpression was significantly associated with tumor differentiation [20]. On the other hand, Wang et al. and Wachter et al. reported no significant association between IMP3 expression and tumor differentiation [18,21]. The reason might be due to different evaluation methods and number of patients, as Wang et al. performed real-time PCR using RNA samples of PDAC and Wachter et al. examined only 26 PDAC patients. Also, some reports showed an association between IMP3 and tumor grade in colorectal and lung cancers [16,17]. In this study, IMP3 expression, tumor differentiation, residual tumor classification, and lymph node status were associated with survival rate in univariate analysis (Table 2). In multivariate analysis, tumor differentiation and lymph node metastasis was significant correlated to OS. Schaeffer et al. reported the IMP3 overexpression in PDAC was correlated with poor survival [20]. No association between IMP3 expression and OS was revealed by multivariate analysis in the present study of 47 PDAC specimens, although a lower number of samples might lead to no association between IMP3 expression and prognosis in multivariate analysis. It is very important to perform pathological diagnosis before any treatment for PDAC. To obtain tumor samples, EUS-FNA is increasingly used worldwide. However, pathological diagnosis of PDAC is sometimes difficult because of the small size of the tissue samples. In such cases, immunohistochemical markers may help to differentially diagnose a malignant lesion from a benign lesion. IMP3 has been reported as a useful diagnostic marker in PDAC for analysis of core needle biopsy samples or for immunocytochemical analysis of samples obtained using a conventional FNA needle [11,18]. Small sample size often precludes histological examination, and cytology plus immunocytochemistry has been performed in many centers; however, it is sometimes challenging to differentiate PDAC with atypical cells owing to reactive inflammatory changes only by cytological examination. Unlike cytological aspirates, examination of histological specimens can aid in a more definitive diagnosis of benign diseases. As for immunocytochemistry, although it on direct smears has potential utility, several problems has been reported such as unavailability of serial section, possibility of cell loss, nonspecific staining by three-dimensional cell groups and background staining due to blood and necrotic materials [35]. A core biopsy needle is 19G to 22G, which is thicker than a FNA needle, and it is sometimes difficult to puncture a target lesion with core biopsy needle. In addition, recently, it has been reported that

7

the diagnostic histological specimens could be obtained by conventional EUS-FNA needles [36,37]. Therefore, in this study, we evaluated the utility of IMP3 expression in histological samples obtained using a conventional FNA needle. In this study, the sensitivities of cytology, histology, and cytohistology were 64.6%, 69.1%, and 80.8% respectively, which were relatively lower as compared with previous studies [2e6]. The diagnosis of “atypical cell” was 21.1% (21/99) by cytology and 18.6% (18/97) by histology. There were many cases diagnosed as atypical cell mainly due to the small sample sizes, which may result in lower sensitivity of cytological and histological analyses. However, 12 (75.0%) of 16 PDAC cases histologically diagnosed as atypical cell were positive for IMP3. Therefore, the sensitivity and accuracy of IMP3 immunohistochemical staining were greater than for routine cytohistology alone. Next, the relationship between immunohistochemical expression and tissue quantity was analyzed. In this study, 79.0% (64/81) of tissue samples of PDAC obtained by EUS-FNA were scored as 3 or 4, which is similar to that reported in a previous study (72.2e90%) [5,9]. In PDACs, 82.6% and 92.3% of specimens, which were scored as 3 or 4, were positive for IMP3 expression. Even in specimens that received a score of 2 (limited diagnosis), 5 (62.5%) of 8 were positive for IMP3 expression. Therefore, IMP3 staining should be considered valuable to distinguish malignant from benign lesions in small tissue samples, which are difficult to evaluate by HE staining alone. In this study, the sensitivity and accuracy of p53 immunohistochemical staining of EUS-FNA specimens were not significantly improved, as compared to routine staining analysis. The criteria of p53 expression was strictly set to avoid false-positive diagnoses. These strict criteria may have led to the lack of significant differences in sensitivity and accuracy. The criteria for p53 positivity or negativity are controversial. Some reports defined nuclei staining of more than 30%e60% as positive for p53 expression [25,27]. The results of the present study showed the potential diagnostic efficacy of p53 immunohistochemical staining, which may help to arrive at a more confident diagnosis in combination with routine analysis. There were some limitation to this study that should be addressed. First, this was a retrospective study. Second, HE staining and IMP3 or p53 immunohistochemical staining were not performed at the same time. Third, immunohistochemical staining was sometimes performed after evaluation by HE staining, thus the sliced sections may have been too small to perform immunohistochemical staining. In such cases, diagnosis was insufficient. In conclusion, the results of this study demonstrated the utility of IMP3 as an immunohistochemical marker for histological analysis of solid pancreatic masses obtained by EUS-FNA. References [1] Li D, Xie K, Wolff R, Abbruzzese JL. Pancreatic cancer. Lancet 2004;363: 1049e57. [2] Puli SR, Bechtold ML, Buxbaum JL, Eloubeidi MA. How good is endoscopic ultrasound-guided fine-needle aspiration in diagnosing the correct etiology for a solid pancreatic mass?: A meta-analysis and systematic review. Pancreas 2013;42:20e6. [3] Chen J, Yang R, Lu Y, Xia Y, Zhou H. Diagnostic accuracy of endoscopic ultrasound-guided fine-needle aspiration for solid pancreatic lesion: a systematic review. J Canc Res Clin Oncol 2012;138:1433e41. [4] Itoi T, Sofuni A, Itokawa F, Irisawa A, Khor CJ, Rerknimitr R. Current status of diagnostic endoscopic ultrasonography in the evaluation of pancreatic mass lesions. Dig Endosc 2011;23:17e21. [5] Iglesias-garcia J, Dominguez-munoz E, Lozano-leon A, Abdulkader I, LarinoNoia J, Antunez J, et al. Impact of endoscopic ultrasound-guided fine needle biopsy for diagnosis of pancreatic masses. World J Gastroenterol 2007;13(2): 289e93. [6] Fuccio L, Hassan C, Laterza L, Correale L, Pagano N, Bocus P, et al. The role of Kras gene mutation analysis in EUS-guided FNA cytology specimens for the differential diagnosis of pancreatic solid masses: a meta-analysis of

Please cite this article in press as: Senoo J, et al., Immunohistochemical analysis of IMP3 and p53 expression in endoscopic ultrasound-guided fine needle aspiration and resected specimens of pancreatic diseases, Pancreatology (2017), https://doi.org/10.1016/j.pan.2017.12.010

8

J. Senoo et al. / Pancreatology xxx (2017) 1e8

prospective studies. Gastrointest Endosc 2013;78:596e608. [7] Jahng AW, Reicher S, Chung D, Varela D, Chhablani R, Dev A, et al. Staining for p53 and Ki-67 increases the sensitivity of EUS-FNA to detect pancreatic malignancy. World J Gastrointest Endosc 2010;2:362e8. [8] Kamata K, Kitano M, Yasukawa S, Kudo M, Chiba Y, Ogura T, et al. Histologic diagnosis of pancreatic masses using 25-gauge endoscopic ultrasound needles with and without a core trap: a multicenter randomized trial. Endoscopy 2016;48:632e8. [9] Kudo T, Kawakami H, Hayashi T, Yasuda I, Mukai T, Inoue H, et al. High and low negative pressure suction techniques in EUS-guided fine-needle tissue acquisition by using 25-gauge needles: a multicenter, prospective, randomized, controlled trial. Gastrointest Endosc 2014;80:1030e7. e1. [10] Van Heek T, Rader AE, Offerhaus GJ, McCarthy DM, Goggins M, Hruban RH, et al. K-ras, p53, and DPC4 (MAD4) alterations in fine-needle aspirates of the pancreas: a molecular panel correlates with and supplements cytologic diagnosis. Am J Clin Pathol 2002;117:755e65. [11] Zhao H, Mandich D, Cartun RW, Ligato S. Expression of K homology domain containing protein overexpressed in cancer in pancreatic FNA for diagnosing adenocarcinoma of pancreas. Diagn Cytopathol 2007;35:700e4. [12] Lu D, Vohra P, Chu PG, Woda B, Rock KL, Jiang Z. An oncofetal protein IMP3: a new molecular marker for the detection of esophageal adenocarcinoma and high-grade dysplasia. Am J Surg Pathol 2009;33:521e5. [13] Hoffmann NE, Sheinin Y, Lohse CM, Parker AS, Leibovich BC, Jiang Z, et al. External validation of IMP3 expression as an independent prognostic marker for metastatic progression and death for patients with clear cell renal cell carcinoma. Cancer 2008;112:1471e9. €bel M, Xu H, Bourne PA, Spaulding BO, Shih IeM, Mao TL, et al. IGF2BP3 [14] Ko (IMP3) expression is a marker of unfavorable prognosis in ovarian carcinoma of clear cell subtype. Mod Pathol 2009;22:469e75. [15] Sitnikova L, Mendese G, Liu Q, Woda BA, Lu D, Dresser K, et al. IMP3 predicts aggressive superficial urothelial carcinoma of the bladder. Clin Canc Res 2008;14:1701e6. [16] Yan J, Wei Q, Jian W, Qiu B, Wen J, Liu J, et al. IMP3 predicts invasion and prognosis in human lung adenocarcinoma. Lung 2016;194:137e46. [17] Wei Q, Huang X, Fu B, Liu J, Zhong L, Yang Q, et al. IMP3 expression in biopsy specimens of colorectal cancer predicts lymph node metastasis and TNM stage. Int J Clin Exp Pathol 2015;8:11024e32. [18] Wachter DL, Schlabrakowski A, Hoegel J, Kristiansen G, Hartmann A, Riener MO. Diagnostic value of immunohistochemical IMP3 expression in core needle biopsies of pancreatic ductal adenocarcinoma. Am J Surg Pathol 2011;35:873e7. [19] Yantiss RK, Woda BA, Fanger GR, Kalos M, Whalen GF, Tada H, et al. KOC (K homology domain containing protein overexpressed in cancer). Am J Surg Pathol 2005;29:188e95. [20] Schaeffer DF, Owen DR, Lim HJ, Buczkowski AK, Chung SW, Scudamore CH, et al. Insulin-like growth factor 2 mRNA binding protein 3 (IGF2BP3) overexpression in pancreatic ductal adenocarcinoma correlates with poor survival. BMC Canc 2010;10:59. [21] Wang BJ, Wang L, Yang SY, Liu ZJ. Expression and clinical significance of IMP3 in microdissected premalignant and malignant pancreatic lesions. Clin Transl Oncol 2015;17:215e22. [22] Lok T, Chen L, Lin F, Wang HL. Immunohistochemical distinction between intrahepatic cholangiocarcinoma and pancreatic ductal adenocarcinoma. Hum

Pathol 2014;45:394e400. [23] Ibrahim DA, Abouhashem NS. Diagnostic value of IMP3 and mesothelin in differentiating pancreatic ductal adenocarcinoma from chronic pancreatitis. Pathol Res Pract 2016;212:288e93. [24] Liu H, Shi J, Anandan V, Wang HL, Diehl D, Blansfield J, et al. Reevaluation and identification of the best immunohistochemical panel (pVHL, maspin, S100P, IMP-3) for ductal adenocarcinoma of the pancreas. Arch Pathol Lab Med 2012;136:601e9. [25] Yemelyanova A, Vang R, Kshirsagar M, Lu D, Marks MA, Shih IeM, et al. Immunohistochemical staining patterns of p53 can serve as a surrogate marker for TP53 mutations in ovarian carcinoma: an immunohistochemical and nucleotide sequencing analysis. Mod Pathol 2011;24:1248e53. [26] Itoi T, Takei K, Sofuni A, Itokawa F, Tsuchiya T, Kurihara T, et al. Immunohistochemical analysis of p53 and MIB-1 in tissue specimens obtained from endoscopic ultrasonography-guided fine needle aspiration biopsy for the diagnosis of solid pancreatic masses. Oncol Rep 2005;13:229e34. [27] Oshima M, Okano K, Muraki S, Haba R, Maeba T, Suzuki Y, et al. Immunohistochemically detected expression of 3 major genes (CDKN2A/p16, TP53, and SMAD4/DPC4) strongly predicts survival in patients with resectable pancreatic cancer. Ann Surg 2013;258:336e46. [28] Kim H, Park CY, Lee JH, Kim JC, Cho CK, Kim HJ. Ki-67 and p53 expression as a predictive marker for early postoperative recurrence in pancreatic head cancer. Ann Surg Treat Res. 2015;88:200e7. [29] Shin SH, Kim SC, Hong S-M, Kim YH, Song KB, Park KM, et al. Genetic alterations of K-ras, p53, c-erbB-2, and DPC4 in pancreatic ductal adenocarcinoma and their correlation with patient survival. Pancreas 2013;42:216e22. [30] Kurahara H, Maemura K, Mataki Y, Sakoda M, Shinchi H, Natsugoe S. Impact of p53 and PDGFR-b expression on metastasis and prognosis of patients with pancreatic cancer. World J Surg 2016;40:1977e84. [31] Bosman FTCF, Hruban RH, Theise N. WHO classification of tumours of the digestive system. fourth ed. Geneva: WHO; 2010. [32] Sobin LH, Gospodarowicz MK, Wittekind C, Greene FL, Gunderson LL, Jessup JM, et al. TNM classificaton of malignant tumors. seventh ed. Hoboken, NJ: Wiley-Blackwell; 2010. [33] Riener MO, Fritzsche FR, Clavien PA, Pestalozzi BC, Probst-Hensch N, Jochum W, et al. IMP3 expression in lesions of the biliary tract: a marker for high-grade dysplasia and an independent prognostic factor in bile duct carcinomas. Hum Pathol 2009;40:1377e83. [34] Gerke H, Rizk MK, Vanderheyden AD, Jensen CS. Randomized study comparing endoscopic ultrasound-guided Trucut biopsy and fine needle aspiration with high suction. Cytopathology 2010;21:44e51. [35] Noda Y, Fujita N, Kobayashi G, Itoh K, Horaguchi J, Takasawa O, et al. Diagnostic efficacy of the cell block method in comparison with smear cytology of tissue samples obtained by endoscopic ultrasound-guided fine-needle aspiration. J Gastroenterol 2010;45:868e75. [36] Abe Y, Kawakami H, Oba K, Hayashi T, Yasuda I, Mukai T, et al. Effect of a stylet on a histological specimen in EUS-guided fine-needle tissue acquisition by using 22-gauge needles: a multicenter, prospective, randomized, controlled trial. Gastrointest Endosc 2015;82:837e44. [37] Bang JY, Hebert-Magee S, Trevino J, Ramesh J, Varadarajulu S. Randomized trial comparing the 22-gauge aspiration and 22-gauge biopsy needles for EUSguided sampling of solid pancreatic mass lesions. Gastrointest Endosc 2012;76:321e7.

Please cite this article in press as: Senoo J, et al., Immunohistochemical analysis of IMP3 and p53 expression in endoscopic ultrasound-guided fine needle aspiration and resected specimens of pancreatic diseases, Pancreatology (2017), https://doi.org/10.1016/j.pan.2017.12.010