Genomic profiling of intestinal/mixed‐type superficial non‐ampullary duodenal epithelial tumors

Abstract Background and Aim The mechanism underlying carcinogenesis and the genomic features of superficial non‐ampullary duodenal epithelial tumors (SNADETs) have not been elucidated in detail. In this study, we examined the genomic features of incipient SNADETs, such as small lesions resected via endoscopic treatment, using next‐generation sequencing (NGS). Methods Twenty consecutive patients who underwent endoscopic treatment for SNADETs of less than 20 mm between January and December 2017 were enrolled. Targeted genomic sequencing was performed through NGS using a panel of 160 cancer‐related genes. Furthermore, the alteration/mutation frequencies in SNADETs were examined. Results The maximum size of the SNADETs examined in this study was 12 mm in diameter. Five SNADETs were classified as low‐grade dysplasia (LGD) tumors, while 14 SNADETs were classified as high‐grade dysplasia tumors. Only one carcinoma in situ was detected. NGS data for 16 samples were obtained. APC alterations were detected in 81% of samples (13/16). KRAS, BRAF, and TP53 alterations were detected in 25% (4/16), 18.8% (3/16), and 6.3% (1/16) of cases, respectively. Conclusion We detected APC alterations in most small SNADETs resected via endoscopic treatment, from LGD to carcinoma samples. Even in SNADETs classified as small LGD exhibited KRAS and BRAF alterations.


Introduction
Superficial non-ampullary duodenal epithelial tumors (SNADETs) are defined as adenomas and superficial adenocarcinomas, including carcinoma in situ (CIS) and submucosal invasive cancer of the non-ampullary duodenal area. 1 Duodenal epithelial tumors are extremely rare, with a reported prevalence of 0.4% in patients undergoing esophagogastroduodenoscopy. 2 However, the detection rate of duodenal carcinoma has been increasing owing to the widespread use of endoscopy. 1,3 Recently, diagnostic methods based on magnified endoscopy with narrow-band imaging (NBI) or endocytoscopy have been reported. 4,5 In addition, the number of resected SNADETs has been increasing owing to improvements in endoscopic treatment. 1 Subsequently, our understanding of the clinical and pathological features of SNADETs has been improving. 6,7 However, relationships among the genomic profile and prognosis of SNADETs have not been clarified.
In colorectal cancer (CRC), the adenoma-carcinoma sequence describes the process of carcinogenesis. 8 APC plays a principal role in CRC development as a tumor suppressor gene. Extensive studies of associations between gene alterations in key driver genes and CRC metastasis 9 have demonstrated the significant roles of alterations in KRAS, TP53, SMAD4, and BRAF. Similar mechanisms to those in CRC, such as the adenomacarcinoma sequence, may contribute to the pathogenesis of duodenal adenocarcinoma. 10 Genomic analyses of duodenal tumors have reported APC, KRAS, and BRAF alterations. 11,12 Recently, numerous studies on genetic alterations of advanced small bowel adenocarcinomas have been reported. 13 However, the data on SNADETs regarding genomic alterations are limited. In addition, the mechanism underlying carcinogenesis and the genomic features of SNADETs have not been elucidated in detail.
In this study, we examined the genomic features of incipient SNADETs, such as small lesions resected by endoscopic treatment, using next-generation sequencing (NGS).

Methods
Subjects and samples. Twenty consecutive patients (20 samples) who underwent endoscopic treatment for SNADETs less than 20 mm in diameter between January and December 2017 at Hokkaido University Hospital were enrolled. Cold snare polypectomy (CSP) and endoscopic mucosal resection (EMR) are generally indicated for lesions that are ≤10 mm and ≤20 mm in diameter, 6 respectively. Therefore, in this study, we included SNADETs that were less than 20 mm in diameter. None of the patients had any family history of cancer, familial adenomatous polyposis (FAP), or Peutz-Jeghers syndrome. SNADETs were removed by endoscopic treatment (EMR, CSP, or endoscopic submucosal dissection [ESD]).
This study was approved by the institutional review board of Hokkaido University Hospital (clinical research approval number 017-0417). Written informed consent was obtained from each participant. All experiments were performed in accordance with the ethical guidelines of the 2013 Declaration of Helsinki.
Specimen handling. All resected specimens were routinely fixed in 10% buffered formalin for 24 h at room temperature. Thereafter, the specimens were serially sliced at a width of approximately 2 mm and embedded in paraffin following routine methods. All sections were cut to a thickness of 3 μm and stained with hematoxylin and eosin for microscopic examination. Paired peripheral blood samples were collected from each patient and stored at À80 C.
Clinicopathological assessment. Clinicopathological findings were reviewed, including age, sex, tumor location, tumor color, tumor size, tumor macroscopic type, resection method, histological type, and phenotype of the resected specimen. Macroscopic typing of SNADETs was based on the Japanese Classification of Colorectal, Appendiceal, and Anal Carcinoma. 14 According to endoscopic features, the samples were classified into the elevated (0-I), superficial elevated (0-IIa), or superficial shallow or depressed types (0-IIc). Mixed patterns were diagnosed when more than one component was observed. Histological evaluations were performed by two expert pathologists (Satoshi Nimura and Yoshihiro Matsuno) who were blinded to the genomic analysis, clinical information, and endoscopic diagnosis. Histopathological diagnosis was based on the revised Vienna classification. 15 Adenomas of the gastrointestinal tract can be categorized as low-grade dysplasia (LGD; category 3) and high-grade dysplasia (HGD; category 4.1). Adenomas were subclassified into low-grade (equivalent to adenomas with mild to moderate atypia) and high-grade (equivalent to adenomas with severe atypia) according to their degrees of structural and/or cytological atypia. CIS showed obvious structural atypia and nuclear atypia. Representative examples of these adenomas and CIS are shown in Figures 1, 2, and 3.
Genomic DNA extraction from tumor tissues and blood cells. Each resected specimen was sectioned into five slices (8-μm-thick slices), and macroscopic trimming was performed to obtain as many cancer cells as possible for more than 50% tumor cellularity. Genomic DNA was extracted from formalin-fixed paraffin-embedded (FFPE) tissue samples using a GeneRead DNA FFPE Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. Genomic DNA was extracted from the blood samples using a genomic DNA extraction kit (Katayama Chemical, Osaka, Japan). The concentration and purity of genomic DNA samples were determined using a NanoDrop system (Life Technologies, Carlsbad, CA, USA) and Qubit dsDNA HS Assay Kit (Life Technologies) designed to be accurate for sample concentrations of 10-100 ng/mL. Genomic DNAs from the FFPE tissue and blood samples were stored at À80 C until analysis.
Library construction and NGS. Multiplex PCR was performed using a GeneReadDNAseq Panel PCR Kit V2 (Qiagen) and Human Comprehensive Cancer Panel (Qiagen), which included 160 cancer-related genes. Finally, an optimized library was constructed using a Gene Read DNA Library I Core Kit (Qiagen). The library was analyzed using an Agilent DNA 1000 Kit Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Library preparation was achieved within two working days. The enriched libraries were sequenced to obtain paired-end reads (2 Â 150 bp) using the MiSeq NGS platform (Illumina, San Diego, CA, USA), resulting in a mean depth of >500Â. The sequencing data were analyzed using an original bioinformatics pipeline, GenomeJack, tuned for clinical sequence examination, "CLUHRC" (Mitsubishi Space Software Co., Ltd., Tokyo, Japan). 17 Statistical methods. The results were analyzed using Prism version 6 (GraphPad Software, Inc., La Jolla, CA, USA). Data are expressed as means AE standard errors of the mean.  Parameters were compared between two groups by Fisher's exact test or Student's t-test. Differences were considered statistically significant when P < 0.05.
Comparison of gene alteration profiles in LGD and HGD/CIS. The 16 analyzed libraries were divided into two groups (5 LGD and 11 HGD/CIS). There were no significant differences between the rates of APC alterations in the LGD (4/5, 80%) and HGD/CIS groups (9/11, 81.9%), KRAS alterations in the LGD (2/5, 40%) and HGD/CIS groups (2/11, 18.2%), BRAF  Comparison of gene alteration profiles between low-grade dysplasia (LGD) and high-grade dysplasia (HGD)/carcinoma in situ (CIS). The 16 samples analyzed using next-generation sequence were divided into two groups (5 LGD and 11 HGD/CIS). There were no significant differences between the alteration frequencies of APC, KRAS, BRAF, and TP53 in the LGD and HGD/CIS groups. Parameters were compared between two groups using Fisher's exact test. Differences were considered statistically significant if P < 0.05.

Discussion
We observed a high frequency of APC alterations in SNADETs (i.e. 81%). Additionally, there were no significant differences between the rates of APC alterations in the LGD group (80%) and the HGD/CIS group (81.9%). Kojima et al. reported an APC alteration frequency of 54.5% in duodenal adenoma. 12 APC plays a critical role in CRC development as a tumor suppressor gene, and its gene product inhibits Wnt/β-catenin signaling. 18 Based on a gene set enrichment analysis, Sakaguchi et al. 11 found a strong association between expression profiles in duodenal adenomas/adenocarcinomas and colorectal adenomas after Cre-lox APC knockout. These findings suggest that upregulation of the Wnt/β-catenin pathway is a major factor in the initial stages of duodenal adenoma/adenocarcinoma carcinogenesis. Our results further support the key role of APC in duodenal adenomas/adenocarcinomas. In CRC, BRAF and KRAS alterations typically arise at the adenoma stage of the adenoma-carcinoma sequence, 19,20 following an initial APC alteration. KRAS and BRAF encode proteins belong to the Ras-Raf-MEK-ERK signaling pathway. The activation of this pathway is considered a molecular switch, leading to cell growth and proliferation. 21 Alterations in KRAS and BRAF are associated with a risk of developing advanced neoplasia 22 and contribute substantially to CRC metastasis. 9 In the present study, KRAS, BRAF, and TP53 alterations were detected in 25, 18.8, and 6.3% of patients, respectively. Surprisingly, we detected KRAS and BRAF alterations in 40% (2/5) and 20% (1/5) of LGD lesions, respectively. These findings are consistent with a previous study showing that one in five cases of LGD (20%) harbor a KRAS alteration. 12 It has been reported that even in cases of LGD, large SNADETs of ≥20 mm in diameter exhibit a high risk of progression to adenocarcinoma. 23 There were no histological differences between LGD tumors with KRAS or BRAF alterations and those without alterations within wild-type sequences.
TP53 is a key driver gene in CRC progression and is frequently detected in small bowel advanced adenocarcinoma. 13 In this study, one case of CIS had a TP53 alteration. These results support the hypothesis that the accumulation of genetic alterations after an initial APC might cause progression from adenoma to carcinoma in SNADETs. Considering our results and those of previous reports, 11 SNADET progresses according to an adenoma-carcinoma sequence, similar to colorectal tumors. Additionally, more than half of the LGD SNADETs (60%; 3/5) already had KRAS or BRAF alterations, which might result in progression to HGD or carcinoma.
This study had several limitations. It included a relatively limited number of samples (20 samples) and did not include submucosal invasive cancer samples. There were no lesions with a gastric phenotype in the collected samples. Additionally, we performed genome sequencing analysis using the Human Comprehensive Cancer Panel (Qiagen), which included 160 cancer-related genes. Therefore, we could not analyze other gene alterations and epigenomic changes in SNADETs. These limitations should be considered when interpreting the study results. Therefore, further studies that include a larger number of cases and lesions with a gastric phenotype are needed in the near future.
In conclusion, in the incipient SNADETs, such as small lesions resected by endoscopic treatment, we detected APC alterations in most SNADETs from LGD to carcinoma samples. Even in SNADETs classified as small LGD (<12 mm in diameter), KRAS and BRAF alterations were present in few samples.