Pancreatic ductal adenocarcinoma (PDAC) is the fourth leading cause of cancer death, and is projected to become the second leading cause of cancer-related death in the United States by 2030.¹ Surgical resection remains the only curative treatment modality; however, most patients are ineligible for surgery because of advanced disease at diagnosis.²

To develop effective methods for early stage diagnosis of this deadly disease which could improve its outcomes, it is important to identify the fundamental genetic changes involved in PDAC carcinogenesis. Activation of the K-RAS oncogene and inactivation of the tumor suppressor genes CDKN2A/INK4A, TP53, and SMAD4 were recently reported to make key contributions to PDAC carcinogenesis.³ Cytogenetic studies of PDAC have also identified numerous complex structural and numerical alterations at the subchromosomal level, and copy number gain of 3q, 5p, 7p, 8q, 11q, 12p, 17q, 19q, and 20q, and loss of 1p, 3p, 4q, 6q, 8p, 9p, 10q, 12q, 13q, 15q, 17p, 18q, 19p, 21q, and 22q, are recurrent aberrations in PDAC.^4,5 However, cytogenetic studies of PDAC are often complicated by a strong desmoplastic reaction and inflammatory cells.⁶ Accordingly, most analyses have been of pancreatic cell lines, short-term cultures, or xenografted tumor cells; hence recorded genomic alterations could potentially have been acquired in vitro.⁷ Moreover, most reported chromosomal alterations in PDAC have been from the minority of patients who present without detectable metastases and are eligible for surgery. Also, as a substantial proportion of patients develop recurrent disease following surgery, detected genetic changes may not represent early events in carcinogenesis.⁸ Therefore, studies of PDAC precursor conditions, such as intraductal papillary mucinous neoplasm (IPMN) and pancreatic intraepithelial lesion (PanIN), are essential to identify early events in the process of PDAC carcinogenesis. Recently, an increase in the number of patients diagnosed with IPMN (the most common precursor lesion of PDAC) has been described, likely due to incidental discovery with new imaging techniques. However, only a few studies have reported the genetic alterations in these tumors because of difficulties obtaining tissue samples without using invasive procedures, such as surgery.^9-13

It is also important for clinicians to discriminate PDAC from other benign pancreatic diseases to facilitate establishment of appropriate therapeutic plans. Endoscopic ultrasound (EUS)-guided fine-needle aspiration or cytologic brushing is the current standard method for tissue acquisition and pathologic diagnosis of pancreatic lesions. However, histopathologic diagnosis is often difficult and diagnostic sensitivity is consequently unsatisfactory.^14-16 Fluorescence in situ hybridization (FISH) is a clinically useful technique, as it can be conducted using limited tissue material, including smears and brushings, and is rapid and relatively accessible.⁹ Several studies have reported the clinical utility of FISH analysis using the commercially available probe set, UroVysion (Abbott Molecular Inc, Des Plaines, IL, USA), and the Papanicolaou Society of Cytopathology adopted FISH as a useful ancillary test to complement routine cytology.^17-19 However, The UroVysion FISH analysis method was developed for diagnosis of bladder cancer using urine samples, and its probe set contains 9p21 which is directed to the CDKN2A gene, and chromosome enumeration probes (CEPs) directed to chromosomes 3, 7, and 17. To increase diagnostic accuracy for PDAC, it will be essential to identify the optimal probes targeting chromosomal regions typical of this tumor type.

Most of the previous studies concerning the chromosomal alteration of PDAC and IPMN has been conducted in Western countries, and there has been only a few reports from Asian countries.^11,12,20,21 In addition, previous studies did not show consistent chromosomal alteration, which might imply that there could be many genetic variations in the genesis and progression of PDAC and IPMN. We assume that the differences of chromosomal alteration between PDAC and IPMN might provide valuable information about carcinogenesis because IPMN is a well-known premalignant disease. The commonality of genetic alteration between the two diseases might be presumed to occur in the earlier stages of carcinogenesis, and the differences might be occur in the later steps. Besides, these differences might be also helpful in the preoperative diagnosis of PDAC. In this study, we investigated the chromosomal aberrations in PDAC and IPMN patient samples by FISH analysis using probes targeting chromosomes reported as frequently altered in PDAC, to identify the patterns and differences of chromosomal alteration in Korean PDAC and IPMN patients, and to investigate the optimal probe sets that may aid preoperative diagnosis of PDAC.

1. Patients and sample collection

Prospectively collected tumor tissue samples from patients who underwent pancreatic resection at the Seoul National University Hospital (Seoul, Korea) from April 2015 to July 2016 were investigated. The patients had PDAC (n=48) or IPMN (n=17). Among those with IPMN, two, four, and five patients had mild, moderate, and severe dysplasia, respectively, while six had invasive IPMN. Invasive IPMN was defined when there as the presence of an invasive carcinoma derived from (arising in the area of) IPMN pathologically.²² All 65 patients were scheduled for surgery with a suspicion of malignancy in radiological evaluations. Preoperative histologic diagnoses using EUS-guided fine-needle aspiration were performed in 25 of 48 PDAC (52.1%) and in 5 of 17 IPMN patients (29.4%), and were possible to make a diagnosis of malignancy in 19 of 25 PDAC (76.0%), and in four of five IPMN patients (80.0%).

The demographic and pathologic characteristics of the patients, and FISH results for chromosomes 7q, 17p, 18q, 20q, and 21q, and a chromosome enumeration probe 18 (CEP18), were investigated. Pathologic staging was determined according to the seventh edition of the American Joint Committee on Cancer staging system.²³ Informed consent for tissue sample collection for research purposes was obtained from individual patients preoperatively, and the study protocol, as well as ethical issues, were reviewed and approved by the Institutional Review Board at Seoul National University Hospital (IRB numbers: H-0901-010-267 and H-1807-099-960). Each tumor sample was harvested immediately after surgical resection and stored in liquid nitrogen. Some of the harvested tumor sample was sent to a pathologist and confirmed by frozen section biopsy to ensure that the tumor tissue was properly harvested.

2. Fluorescence in situ hybridization

Tumor samples were transported to the laboratory in an icebox. For FISH examination, tissue samples were minced with a surgical scalpel and incubated in collagenase type IV (1 mg/mL) (STEMCELL Technologies, Vancouver, BC, Canada) for 20 minutes. After washing with phosphate-buffered saline, samples were filtered using 100 µm cell strainers (BD Falcon, Franklin Lakes, NJ, USA) to generate a single cell suspension, followed by centrifugation for 5 minutes at 1,200 rpm. After adding 5 mL of 0.075 M KCl to each tube, samples were incubated for 25–30 minutes in a 37°C water bath. Carnoy’s fixative (500 µL) was added, and samples were incubated for 5 minutes at room temperature. Suspensions were centrifuged for 5 minutes at 1,200 rpm, and supernatants were removed. Pellets were resuspended in 3–5 mL of Carnoy’s fixative and incubated for 20 minutes at room temperature, then suspensions were centrifuged for 5 minutes at 1,200 rpm and the supernatants were removed; this step was performed twice. Next, fixed cells were mixed with Carnoy’s fixative and dropped onto microscope slides. Air-dried slides were pretreated with 2× standard saline citrate (SSC; 300 mmol/L sodium chloride and 30 mmol/L sodium citrate) for 30 minutes at 37°C, and dehydrated with cold 70%, 85%, and 100% ethanol for 2 minutes each. Under protection from light, FISH probes were added to the prepared slides, which were then covered with coverslips and sealed with rubber cement. FISH probes used were as follows: XL Spectrum Orange (7q22)/Spectrum Green (7q36), XL ATM Spectrum Green (11q22)/TP53 Spectrum Orange (17p13), XL MALT Break Apart Spectrum Orange/Green (18q21) (MetaSystems, Altlussheim, Germany), Vysis CEP 18 (D18Z1) Spectrum Orange (Abbott Molecular), IGH Spectrum Green (14q32.33)/MAFB Spectrum Red (20q12) (Cytocell Ltd, Cambridge, UK), and Vysis RUNX1 Spectrum Green (21q22)/RUNX1T1 Spectrum Orange (8q21) (Abbott Molecular). Probes and target DNA were simultaneously denatured at 75°C for 5 minutes, then slides were hybridized for 10–16 hours at 37°C in a hybridizer (Dako, Glostrup, Denmark). After hybridization, slides were washed in 0.4× SSC at 73°C for 2 minutes, and in 0.1% Nonidet P-40/2× SSC at room temperature for 2 minutes. Chromosomes were counterstained with 10 µL of 4′-6′-diamine-2-phenylindole dihydrochloride (DAPI/Antifade) (MetaSystems). Images were analyzed using a Zeiss Axioplan 2 imaging microscope (Carl Zeiss MicroImaging GmbH, Munich, Germany) with the ISIS software (MetaSystems). Approximately 100 nuclei were scored for each probe (Fig. 1). Nuclei with ambiguous signals and cells with poor morphology were excluded from scoring. The absolute cutoff values of FISH analysis using tissue samples has not yet been established, and we adopted the binomial treatment of the data to set the cutoff values because it was one of the reliable methods to calculate the cutoff values of FISH analysis in the hematological diseases. Based on FISH analysis of 18 normal pancreatic tissue samples in a preliminary study, cutoff values for the normal range for FISH analysis were calculated using the Excel 2013 (Microsoft Corp., Redmond, WA, USA) statistical function CRITBINOM (n, p, α) with a confidence level of 95% (Table 1).²⁴ When the percentage of cells containing >2 or <2 FISH signals exceeded the cutoff value, cases were interpreted as positive for polysomy (gain) or monosomy (deletion), respectively.

3. Statistical analysis

Categorical variables are presented as numbers and percentages, and were compared using the Fisher exact test. Continuous variables are expressed as means with standard deviations, and were compared using the Mann-Whitney U test. All statistical analyses were conducted using SPSS version 20.0 (IBM Corp., Armonk, NY, USA), and p-values <0.05 were considered statistically significant.

1. Patient demographic and clinicopathological characteristics

The demographic and clinicopathological characteristics of the study population are detailed in Table 2. The mean age of the participants was 65.1 years, and 43 of them (66.2%) were male. Pancreaticoduodenectomy was the most common treatment method, with distal pancreatectomy the second most common. Carbohydrate antigen 19-9 (CA19-9) levels were significantly higher in the PDAC group than the IPMN group. Pathologically, most patients with PDAC had T3 disease (44/48, 91.7%) and lymph node metastases (37/48, 77.1%). However, there were no statistically significant differences between the PDAC and invasive IPMN in pT stage (p=0.080), pN stage (p=0.173), angiolymphatic invasion (p=0.413), perineural invasion (p=0.070), and venous invasion (p=1.000).

2. FISH analysis and comparison of PDAC and IPMN

At least two chromosome alterations were detected in all patients with either PDAC or IPMN using standard cutoff values (Table 1). For the PDAC group, 17p deletion was the most frequently detected alteration (46/48, 95.8%), followed by 18q deletion (40/48, 83.3%), CEP18 deletion (39/48, 81.2%), 20q gain (39/48, 81.2%), 21q deletion (37/48, 77.1%), and 7q gain (34/48, 70.8%) (Table 3). For the IPMN group, 17p deletion (16/17, 94.1%) and CEP18 deletion (16/17, 94.1%) were also the most frequently detected alterations, followed by 21q deletion, 20q gain, 18q deletion, and 7q gain. CEP18 gain was significantly more frequent in the PDAC group than the IPMN group (26/48 vs 4/17, p=0.029), and the frequency of 18q deletion was marginally significantly different between the two groups (40/48 vs 10/17, p=0.051). The patterns of chromosomal alteration were similar between invasive and noninvasive IPMN, and there were no statistical differences between the two groups in the chromosomal alterations detected by each probe. The details of frequent chromosomal alterations identified in invasive versus noninvasive IPMN were as follows: 7q gain (4/6 vs 6/11, p=1.000), 17p deletion (6/6 vs 10/11, p=1.000), 18q deletion (3/6 vs 7/11, p=0.644), 20q gain (4/6 vs 7/11, p=1.000), 21q deletion (6/6 vs 6/11, p=0.102), and CEP18 deletion (6/6 vs 10/11, p=1.000).

There were statistically significant differences between the PDAC and IPMN groups in the proportion of cells with 17p deletion (32.2±29.3 vs 16.9±20.8, p=0.019) and 18q deletion (27.7±30.2 vs 7.0±13.5, p=0.004) (Fig. 2). For the other probes, there were no statistically significant differences in the chromosomal alteration rates between the PDAC group and IPMN group.

Modification of the cutoff value (percentage of cells positive for a chromosome alteration required for a positive score) resulted in statistically significant differences in mean alteration rates between the PDAC and IPMN groups for some probes as follows: 17p deletion with cutoff values of 10.0% (36/48 vs 7/17, p=0.011) and 20.0% (27/48 vs 3/17, p=0.006); 18q deletion with cutoff values of 10.0% (29/48 vs 2/17, p=0.001) and 20.0% (23/48 vs 1/17, p=0.002); and CEP18 gain with a cutoff value of 5.0% (18/48 vs 2/17, p=0.048).

3. Associations of clinicopathologic features and results of FISH analysis

Analysis of relationships between the commonly identified chromosomal alterations and clinicopathological factors demonstrated that CEP18 gain was significantly more frequent in older patients (21/28 vs 7/20, p=0.024) and those with lymph node metastasis (23/37 vs 3/11, p=0.041) (Table 4); however, no other probes exhibited any significant associations with clinicopathological factors.

For patients with PDAC, overall 1- and 2-year survival rates were 67.2% and 61.6%, respectively. During follow-up (median, 15.5 months; range, 0 to 26 months), recurrence was diagnosed in 26 of 48 patients (54.2%). The majority of recurrence was diagnosed within 12 months (21/26, 80.8%), and the median time to recurrence was 5.0 months. Patients with recurrence had comparable carcinoembryonic antigen and CA19-9 levels and pathologic findings, including T stage, N stage, differentiation, angiolymphatic invasion, perineural invasion, and venous invasion, to those without recurrence (Table 5). There were also no significant differences between patients with recurrence and without recurrence in 7q gain, 17p deletion, 18q deletion, CEP18 deletion, 20q gain, 21q deletion, or CEP18 gain.

4. Preoperative diagnosis of PDAC

To clarify whether the investigated chromosomal alterations were useful for discriminating between PDAC and IPMN, we evaluated the diagnostic accuracies of various patterns of chromosomal alterations (Table 6). Analysis including the five common chromosomal aberrations in the PDAC group (7q gain, 17p deletion, 18q deletion, 20q gain, and 21q deletion) and CEP18 gain, which were significantly more frequent than in the IPMN group, identified a mean of 4.6±0.9 (range, 2 to 6) chromosomal alterations in the PDAC group, with 3.7±1.1 (range, 2 to 6) in the IPMN group (p=0.004). Selection of single probes resulted in higher diagnostic accuracy, with values of 79.3% (sensitivity, 87.8%; specificity, 58.8%; and relative risk, 2.5) for 17p deletion (cutoff value, 10.0%), and it slightly increased to 80.0% when 17p deletion (cutoff value, 10.0%) combined with 18q deletion (standard cutoff value: sensitivity, 97.9%; specificity, 29.4%; and relative risk, 4.8) or 18q deletion with a 10.0% cutoff value (sensitivity, 89.6%; specificity, 52.9%; and relative risk, 2.3).

When we analyzed the 25 PDAC patients who had preoperative histologic diagnosis, all six patients (100.0%) who had not been diagnosed of malignancy showed positive results by FISH analysis whether the diagnostic criteria was set to the 17p deletion (cutoff value, 10.0%) combined with 18q deletion (standard cutoff value), or 18q deletion (cutoff value, 10.0%). The 18 (94.7%), and 16 (84.2%) of 19 patients who had preoperative diagnosis of malignancy showed positive results when the diagnostic criteria was set to the 17p deletion (cutoff value, 10.0%) combined with 18q deletion (standard cutoff value), or 18q deletion (cutoff value, 10.0%), respectively.

The results of the present study demonstrate that chromosomal alterations are very frequent in tumor samples from patients with both PDAC and IPMN. The chromosomal aberration patterns in IPMN were similar to those in PDAC, there were no significant differences in most probe sets, and comparisons of invasive and noninvasive IPMNs also demonstrated no significant differences between these groups. These results imply that similar early genetic alterations may be implicated in the development of both IPMN and PDAC, although they may be partly attributable to the fact that more than half of patients had severe dysplasia or invasive IPMN. Some previous cytogenetic studies support this assumption. Fujii et al.¹¹ conducted PCR-based microsatellite analysis of 13 IPMN specimens and found frequent loss of heterozygosity at 6q, 8p, 9p, 17p, and 18q with ratios of 31% to 62%. Fritz et al.¹⁰ investigated 20 IPMN specimens by microarray-based comparative genomic hybridization analysis and reported frequent loss of chromosomes 2, 4q, 5q, 6q, 8p, 10q, 11q, 13q, 15q, 18q, and 22q with ratios of 38.5% to 76.9%, and gains of chromosomes 7 and 19q in half of specimens from invasive IPMN or IPMN with severe dysplasia. Both studies identified chromosomal aberrations also frequently identified in PDAC.^4,5 Nevertheless, the chromosomal changes in IPMN have not been fully elucidated and further studies are warranted, as most previous studies have been based on small numbers of tissue samples.

The present study identified that using a modified cutoff value of 10%, deletions in 17p13 (TP53) and 18q21 (SMAD4/DPC4) were significantly more frequent in PDAC than IPMN. Both the TP53 and SMAD4/DPC4 genes are well-known tumor suppressors reportedly inactivated in more than 50% of PDACs.³ Previous studies of genetic or protein loss of TP53 and SMAD4 revealed rising incidence with increasing PanIN grade.^25,26 The present study supports the previously proposed tumor progression model for PDAC, which postulates that genetic changes at these loci may be involved in the late steps of carcinogenesis.²⁷ However, the difference of chromosomal alteration between the PDAC and IPMN might be due to the difference of pathway between PanIN- and IPMN-derived carcinogenesis. The deletion rates of chromosome 17p and 18q have been reported to range from 80% to 100%, and 56% to 88% in PDAC or PanIN with high grade dysplasia, and from 73% to 100%, and 54% to 100% of IPMN with high grade dysplasia or invasion, respectively.^12,13,21,28 However, some other studies reported that the deletion of 17p and 18q even in PanIN-1 in 87% and 50%, respectively, and the SMAD4/DPC4 gene was inactivated only 3% of IPMN.^13,29 There have been only a small number of studies concerning the chromosomal alteration of the precursor lesions of PDAC, and future studies are necessary to elucidate how it is involved at any stage of the carcinogenesis.

The present study also determined that CEP18 gain was significantly more frequent in PDAC; however, little is known about the significance of changes in chromosome 18 copy number, particularly gain, in pancreatic carcinogenesis. On the contrary, some previous cytogenetic studies reported consistent frequent loss of chromosome 18 in PDAC, which was also identified in the present study.^4,6,8 Gain of chromosome 18 has been reported in lymphoproliferative diseases, including acute lymphocytic leukemia, multiple myeloma, and non-Hodgkin’s lymphoma; however, its role in carcinogenesis and clinical significance has yet to be elucidated.³⁰ For pancreatic disease, Miyabe et al.¹² reported that polysomy 18 (CEP18) was significantly more frequent in invasive IPMN, and may be involved in malignant transformation of IPMN, along with polysomy 7 and P16/TP53 deletion. Further studies are necessary to clarify the significance in carcinogenesis and clinical impact of CEP18.

The present study did not reveal clear associations between chromosomal aberrations and clinicopathologic features, including disease recurrence and patient survival, other than for CEP18 gain, which was more frequent in older patients and those with lymph node metastasis. This may be because this study was performed on small patient populations with relatively short-term follow-up. Moreover, approximately 90% of patients had T3 disease and about 80% had lymph node metastases, which could mask the effects of chromosomal alterations. However, there are some reports of a relationship between chromosomal alterations and clinicopathologic prognostic factors. Gutiérrez et al.³¹ reported that changes of chromosomes 7, 17q, 18q21, and 20 were significantly more frequent in advanced TNM stage tumors, and that numerical changes of chromosomes 4 and 9q34, together with gains of chromosome 8q24, were associated with reduced overall survival of patients. Stoecklein et al.³² reported that chromosome 17 ploidy level was negatively associated with disease free survival and overall survival.

Although EUS-guided cytology and core needle biopsy have been the primary tools for diagnosis of PDAC, they (particularly cytology) have been discredited because of low diagnostic sensitivity. The diagnostic yields of pancreatic EUS-guided fine-needle aspiration and core needle biopsy indicate sensitivities for these techniques of 54% to 96%, and 71% to 99%, respectively, and the present study showed similar sensitivity (76.0%).^14,15 The main limitation of cytology is false-negative results in patients with PDAC, which can be attributed to various factors, including difficulties in cytologic interpretation of specimens with inflammatory cells, induced by adjacent chronic pancreatitis or recent instrumentation; paucicellular specimens, which harbor few or no malignant cells; and well-differentiated carcinomas, which are difficult to discern.³³ FISH can be used to analyze limited tissue material, including small biopsies, and samples from brushing or aspiration cytology, and has the ability to detect chromosomal alterations common in malignant tumors. This technique has been increasingly used in research and clinical practice for detection of pancreatobiliary malignancy in cytology specimens.^17-20

The present study revealed the highest diagnostic accuracy for PDAC (80.0%) of FISH tests positive for 17p deletion or 18q deletion, with a cutoff value of 10.0%, and single probe detection of 17p deletion (cutoff value, 10.0%) had a diagnostic accuracy of 79.3%, with acceptable sensitivity (87.8%) and specificity (58.8%). We also found that these probe sets might be helpful in increasing the preoperative diagnostic accuracy for the patients who were not diagnosed with PDAC with conventional histologic examinations. Over decades, some studies have reported the clinical utility of UroVysion FISH using pancreatobiliary brushing specimens, revealing sensitivity significantly higher than that of conventional cytology for detection of malignancy.^17-20 UroVysion FISH has a diagnostic sensitivity of 34% to 58%, which is higher than that of routine cytology (8% to 40%); however, approximately half of patients with malignancy remain undiagnosed by FISH.^16-20 However, only a few FISH studies of pancreatobiliary malignancies have used probes other than the UroVysion FISH probe set. Miyabe et al.¹² reported that polysomy 7, polysomy 18, P16 deletion, and TP53 deletion were significantly more frequent in invasive IPMN, and that detection of polysomy 7 or TP53 deletion had potential value as diagnostic markers for invasive IPMN. Barr Fritcher et al.¹⁶ reported that the combination of the FISH probes, 1q21, 7p12, 8q24, and 9p21, identifies cancer cells with 93% sensitivity and 100% specificity, and has significantly higher sensitivity (64.7%) than the UroVysion probes (45.9%) or routine cytology analysis (18.8%). The FISH probe sets used in the present study, which target genes associated with PDAC, showed acceptable diagnostic accuracy, and could be useful as an adjunct to conventional histopathologic examination.

This study has some limitations. First, as our sample size was relatively small and FISH analysis was performed with relatively few probe sets, the statistical power may be limited. This may account for the failure to elucidate any correlation between chromosomal alterations and clinicopathologic factors, including disease recurrence and patient survival. Second, this study compared chromosomal alterations of PDAC to those of IPMN. Because one of the aims of this study was to find out the useful FISH probe sets which could aid preoperative histologic diagnosis, we selected the probes directed to the chromosomes which alterations had been reported relatively frequent in PDAC to increase the diagnostic sensitivity, and set IPMN as a control group to identify the discrimination power of selected probes because IPMN might harbor similar chromosomal alterations to PDAC. However, the chromosomal alterations during the PanIN-derived carcinogenesis would be different from that from IPMN-derived pathway. Because we experimented with cryopreserved tissue samples other than paraffin blocks, we could not harvest more premalignant tissue samples of PanIN and IPMN lesions with low to high grade dysplasia, which might make it possible to elucidate the differences of chromosomal alterations during the PanIN- and IPMN-derived pancreatic carcinogenesis. Nevertheless, we believe the commonality and the differences of chromosomal alteration between PDAC and IPMN identified in the present study could provide helpful information about the carcinogenesis of PDAC for conducting future studies. Lastly, this study was performed using tissue samples obtained by surgical resection, which harbored sufficient cells for analysis. This could have resulted in overestimation of diagnostic accuracy. Therefore, further studies are required to apply our findings in clinical practice using limited cytology specimens or small biopsies, and attempts to identify a more specific FISH probe set devoted to detection of chromosomal alterations typical of PDAC are warranted.

In conclusion, chromosomal alterations were frequently identified in both PDACs and IPMNs. PDACs had 17p deletion, 18q deletion, CEP18 deletion, 20q gain, 21q deletion, and 7q gain in more than 70% of patients, and IPMNs had a similar chromosomal aberration pattern; however, IPMNs had a lower positive rate. Gain of chromosome 18 and deletions in 17p and 18q may be involved in the late steps of PDAC carcinogenesis. Although there were no clear clinicopathological associations with chromosomal alterations, deletions at chromosome 17p and 18q may represent excellent diagnostic markers for PDAC.

This study was supported by the Collaborative Genome Program for Fostering New Post-Genome Industry of the National Research Foundation funded by the Ministry of Science and ICT (NRF-2017M3C9A5031597) and the Korean Health Technology R&D Project, Ministry of Health & Welfare (HI14C2640), Republic of Korea.

No potential conflict of interest relevant to this article was reported.

Data analysis and interpretation: C.S.L., K.I., W.K., J.R.K., Y.H. Data acquisition: K.I., Y.H. Drafting the manuscript: C.S.L. Critical revision of the manuscript, study supervision: D.S.L., S.W.K., J.Y.J. Study concept and design: J.Y.J.