Non-Metastatic Pure Pancreatic Hepatoid Carcinoma with Genomic and Transcriptomic Analysis: A Case Report

Abstract

Background

Pancreatic hepatoid carcinoma (PHC) is a rare and poorly characterized malignancy, with approximately 50 reported cases in the literature. Pure PHC, which lacks any features of adenocarcinoma or neuroendocrine differentiation, is a subset of this population and is extremely rare. Data on its management and genomic findings are limited, and further characterization may provide helpful information in caring for these patients.

Case Presentation

A 42-year-old female was found to have a large, well-circumscribed mass in the body of the pancreas with elevated serum alpha-fetoprotein. Imaging demonstrated an 8.5 cm pancreatic lesion with no other disease. Endoscopic ultrasound-guided biopsy revealed a bile-producing carcinoma with morphological and immunohistochemical features consistent with hepatocellular carcinoma. She underwent a subtotal pancreatectomy and splenectomy, and pathologic evaluation confirmed a well-differentiated pure PHC confined to the pancreas. Whole genome and transcriptome analysis showed microsatellite stability, an elevated tumour mutation burden, copy number alterations in Chr19p13.3, and no mutations typically seen in pancreatic ductal adenocarcinoma (PDAC). These findings support the diagnosis of a pure hepatoid carcinoma with clinical, histopathological, and genomic characteristics resembling hepatocellular carcinoma. She remains disease free at 9 months without adjuvant therapy.

Conclusion

This case demonstrates further characterization of a rare pancreatic lesion and illustrates the importance of integrated histopathological and genomic analyses in characterizing rare malignancies. Additionally, our findings suggest that pure PHC may be a distinct entity rather than a variant of PDAC.

Introduction

Carcinoma arising outside of the liver and showing hepatocellular differentiation (hepatoid carcinoma) is rare. Hepatoid carcinoma has been described in many locations including stomach, oesophagus, pancreas, lung, urinary tract and is frequently associated with elevated alpha-fetoprotein (AFP) levels [1, 2]. Pancreatic hepatoid carcinoma (PHC) is very rare and often forms part of a carcinoma of mixed histologic type, such as mixed hepatoid-ductal adenocarcinoma of the pancreas, which usually has a poor prognosis. Few examples of pure PHC have been reported previously in the literature and are poorly understood. Herein we report the clinical, pathologic, and molecular features of a large, non-metastatic, pure PHC that underwent whole-exome sequencing and transcriptomic profiling as part of the Personalized OncoGenomics Program clinical trial. To ensure a high quality report, the CARE Checklist has been completed for the case report, attached as online supplementary material (for all online suppl. material, see https://doi.org/10.1159/000544747).

Case Presentation

The patient is a 42-year-old Chinese female with no prior medical or surgical history, aside from a caesarean section. She denied other symptoms including nausea, vomiting, fatigue, weight loss, pancreatic insufficiency, or jaundice. She presented to her family physician after 18 months of intermittent epigastric discomfort. She takes no medications and has no allergies. She has no prior history of smoking, alcohol, or recreational drug use. She has no family history of malignancy, liver or pancreatic diseases. She has no concerning environmental exposures. Given the persistent discomfort, blood work and a CT scan were ordered to investigate her symptoms.

Initial laboratory investigations revealed a normal complete blood count and metabolic panel. Serum ALT was minimally elevated at 45 U/L and ALP was elevated at 141 U/L. GGT, total bilirubin, and lipase were normal. Serum tumour markers including CA19-9, CEA, CA125, and CA15-3 were all normal; however, serum AFP was elevated at 52.9 µg/L. A CT abdomen and pelvis with venous phase (Fig. 1a) was performed and demonstrated a well-circumscribed 8.6 × 7.9 × 6.5 cm heterogeneously enhancing pancreatic body mass with solid and cystic components. There was no pancreatic duct dilation or signs of pancreatic atrophy. The mass moderately compressed the splenic vein, with no evidence of left-sided portal hypertension or splenic vein thrombosis. A gadolinium-enhanced MRI abdomen demonstrated a predominantly T1 and T2 hypointense lesion with small areas of T1 hyperintensity suggestive of intralesional haemorrhage or proteinaceous material (Fig. 1b). The case was discussed at multidisciplinary rounds given its unusual appearance. Endoscopic ultrasound demonstrated a pancreatic tail mass with well-defined borders and a heterogeneous appearance. Endoscopic ultrasound however did not demonstrate cystic components. Fine-needle biopsy of the mass revealed a bile-producing carcinoma, morphologically and immunohistochemically identical to hepatocellular carcinoma (HCC). By immunohistochemistry, the tumour was positive for arginase-1 but negative for cytokeratin 7 and chromogranin-A. A chest CT scan revealed no evidence of metastatic disease, and a review of prior abdominal imaging demonstrated no abnormality or tumour within the liver.

Fig. 1.

a CT abdomen and pelvis portal venous phase images of PHC. b MRI abdomen and pelvis T2 phase images of PHC.

The case was brought forward to the local multidisciplinary tumour board, which concluded this mass to be a PHC given no other foci of HCC on imaging and no mixed histology on the biopsy specimens. A formal oncologic resection was recommended, and she successfully underwent an open subtotal pancreatectomy and splenectomy without tumour rupture (Fig. 2). The intraoperative frozen section from the pancreatic neck margin was negative for tumour. The non-tumoral pancreas was soft with a small non-dilated pancreatic duct. There was no fracture of the gland during resection. Two closed suction drains were left in the resection bed. Her hospital course was generally uncomplicated aside from a low-volume biochemical pancreatic leak. Because of this, she was discharged home with both drains on post-op day 8.

Fig. 2.

Subtotal pancreatectomy and splenectomy with PHC.

She was seen in the clinic 3 weeks postoperatively and was doing well. Her incisions had healed well, and she was only experiencing mild bloating symptoms and fatigue that did not significantly affect her life. Her drains were removed at this initial visit as there was no further output to indicate a biochemical leak. An AFP level taken before the visit was 5.2 µg/L. She was seen 1 month later and noted her bloating and fatigue were improving. She had also returned to her preoperative weight. A PET scan was performed and demonstrated no areas of concern aside from mild uptake in the endometrium, which was biopsied and shown to be benign. Given the pathologic and molecular findings following the surgical resection of the PHC, medical oncology decided against offering any systemic or targeted therapy. The patient continues on active surveillance and is now disease free, approximately 9 months post-resection.

The subtotal pancreatectomy contained an 8.5 cm, well-circumscribed, lobulated, and entirely solid tan-brown mass with fibrous septa (Fig. 3). Other than its location, the gross appearance of the mass was typical of HCC. No areas of peri-tumoral fibrosis, cystic change, or mucin production were identified. The mass was confined to the pancreas but compressed the anterior and posterior pancreatic surfaces. The specimen included 37 benign, regional peripancreatic lymph nodes.

Fig. 3.

Circumscribed, lobulated, tan-brown tumour in the mid-body of the pancreas – post-fixation.

On microscopic examination, the mass had an entirely solid and trabecular architecture, with tumour lobules divided by fibrous septa and areas of intra-tumoral haemorrhage. The mass contained numerous capillaries, with a hepatic sinusoidal arrangement. The tumour cells were polygonal, with abundant eosinophilic cytoplasm and with nuclei containing open chromatin and prominent, single or multiple nucleoli. An estimated 10–20% of the tumour cells contained macrovesicular lipids. Focal bile production and numerous large, eosinophilic globules resembling alpha-1 antitrypsin were seen within the tumour. The morphologic appearances of all sections of the tumour were identical to moderately differentiated HCC (Fig. 4).

Fig. 4.

Microscopic section demonstrating tumour with morphologic appearances identical to HCC, with solid and trabecular morphologic arrangement and with macrovesicular lipid (H&E. ×100 original magnification). Top right inset: eosinophilic globules, resembling alpha-1 antitrypsin, within the tumour (H&E. ×400 original magnification). Bottom right inset: tumoral bile production at centre (H&E. ×400 original magnification).

By immunohistochemistry, the tumour was diffusely positive for AE1/AE3 cytokeratin cocktail, patchily but strongly for arginase-1, and entirely negative for hepatocyte-specific antigen. Immunohistochemical markers of neuroendocrine differentiation (synaptophysin and chromogranin-A) and of acinar differentiation (trypsin and BCL10) were negative. The tumour also lacked immunohistochemical evidence of pancreatic pseudopapillary neoplasm, having an intact, membranous pattern of e-cadherin expression and negativity for progesterone receptor. Immunohistochemical staining for SWI/SNF family proteins (SMARCB1, SMARCA4, SMARCA2, and H3K27me3) was normal.

Whole-exome sequencing and transcriptomic profiling were performed as part of the Personalized OncoGenomics Program clinical trial. The tumour was found to be microsatellite stable and to have a whole genome tumour mutation burden (TMB) of 9.61 mutations/Mb. Several regions of chromosome 19 contained copy number amplification, including Chr19p13.3, an area frequently amplified in HCC. The chromosome 19 amplicon contained multiple genes of interest (Table 1), including AKT2, PIK3R2, BRD4, and NOTCH3, some of which are part of ongoing clinical trials, but none of which are associated with an FDA-approved therapy. High outlier expression of LRP6, the gene for a co-receptor on the WNT signalling pathway, was detected with a z score of 4.2 and 5.7 compared to TCGA_LIHC and TCGA_PAAD, respectively. LRP6 is frequently over-expressed in tumours with increased WNT pathway signalling, including HCC. A duplication event affecting SMARCA4 and resulting in a ZNF791-SMARCA4 fusion was detected, and raised a possibility of deregulation of the SWI/SNF complex, despite intact immunohistochemical staining for SWI/SNF proteins. No genetic alterations typical of pancreatic ductal adenocarcinoma were detected, such as gain-of-function mutations in KRAS or loss-of-function alterations in TP53 or SMAD4. Overall, the genomic and transcriptomic findings were considered compatible with HCC.

Table 1.

Variants of interest in the chromosome 19 amplicon

Gene	Chromosome band	Variant type	Copy number change	Z score against TCGA_LIHCa	Z score against TCGA_PAADb
AKT2	chr19:q13.2	Amplification	+5	1.4	1.5
PIK3R2	chr19:p13.11	Amplification	+11	1.7	2.6
BRD4	chr19:p13.12	Amplification	+11	3.8	4.8
EPOR	chr19:p13.2	Amplification	+5	−0.9	−1.9
JAK3	chr19:p13.11	Amplification	+11	−1.2	−3.8
NOTCH3	chr19:p13.12	Amplification	+11	1.2	−0.3
PRKACA	chr19:p13.12	Amplification	+11	3.7	5.4
SMARCA4, ZNF791	chr19:p13.2, p13.13	Structural variant	+5, +4	0.6, 3.4	−0.2, 3.3

^aThe Cancer Genome Atlas Liver Hepatocellular Carcinoma Collection. ^bThe Cancer Genome Atlas Pancreatic Adenocarcinoma.

Discussion

PHC is rare, with approximately 50 cases reported in the English literature [3–7]. PHC is presently classified by the World Health Organization as a histologic variant of PDAC and is arbitrarily defined as pancreatic carcinoma with at least 50% hepatoid differentiation by morphologic and immunohistochemical assessment [8]. PHC may be “pure” or part of a tumour of mixed histologic type, including elements of PDAC, acinar cell carcinoma, neuroendocrine neoplasm, or serous cystadenoma. The number of reported cases of PHC is too small to make any conclusive statement about its prognosis. However, when PHC forms part of a mixed tumour with acinar carcinoma, neuroendocrine carcinoma, or PDAC, it frequently has a poor prognosis as it tends to behave similarly to its mixed component and may require systemic chemotherapy [3, 6]. Cases of pure PHC closely resemble primary HCC and may have an excellent prognosis, without need for additional therapy following surgical excision of the primary lesion [4, 9].

To our knowledge, we are reporting the first case of pure PHC to undergo complete genomic and transcriptomic profiling. In this case, the PHC is genetically similar to HCC, particularly given its amplification of Chr19p13.3, which is a common event in HCC and in hepatoid gastric carcinoma [9]. The high outlier expression of LRP6, which encodes a co-receptor of the WNT signalling pathway, is typical of HCC and various other tumours with increased WNT/β-catenin signalling [10]. Genetic and transcriptomic findings typical of PDAC, neuroendocrine carcinoma, and acinar cell carcinoma were not found.

The ZNF791-SMARCA4 gene fusion identified in this case has not been previously reported in pancreatic carcinoma or in HCC. We hypothesize that the gene fusion might disrupt the function of SMARCA4, one of the genes encoding the SWI/SNF protein complex, which has an important function in chromatin remodelling, and which is frequently mutated in various cancers. Cancers with variations in SWI/SNF genes have higher TMB and appear to have a better response to immune checkpoint inhibitor (ICI) therapy than cancers with normal SWI/SNF genes [11]. The TMB in the present PHC (9.61 mutations/Mb) is almost double the average TMB among cancers with normal SWI/SNF genes (5.6 mutations/Mb) but substantially lower than the average TMB among cancers with variations in the SWI/SNF genes (25.8 mutations/Mb). However, intact immunohistochemical staining for SWI/SNF proteins did not confirm functional deregulation of the SWI/SNF protein complex in this case.

If the presently reported PHC were to recur, ICI therapy would be a treatment option. High TMB (which can be defined as at least 10 mutations/Mb) has been associated with a favourable response to ICI therapy. For example, results from the KEYNOTE-158 clinical trial involving pembrolizumab showed a significantly higher objective response rate among advanced solid tumours with high TMB compared to those with low TMB [12]. Certain ICI-containing treatment regimens have been shown to improve overall survival in first-line treatment of HCC and are FDA approved in second-line treatment of HCC [13, 14]. Good responses to treatment regimens containing ICI therapy have also been reported in 2 cases of metastatic PHC [3, 5].

Mattiolo and colleagues recently reported 2 cases of PHC, which are morphologically and immunohistochemically HCC like and which had a good prognosis following surgical resection. In those cases, a next-generation sequencing panel containing 174 genes identified mutations in CTNNB1 exon 3 and copy number alterations in chromosomes 1, 18, 20, and 21 [7]. The authors suggested that at least some cases of PHC can be considered a distinct tumour entity rather than a variant of PDAC. Based mainly on the genetic findings, the authors advocated that pancreatic hepatoid tumours with CTNNB1 exon 3 mutations and loss of heterozygosity on chromosome 21 should be considered a new variant of pancreatic SPN, if they have a roundish shape, contain hyaline globules, and are lymph node negative.

Based on the findings in the presently reported case, we agree that PHC, which resembles well to moderately differentiated HCC on histologic and immunohistochemical analysis, is confined to the pancreas, is lymph node negative, and lacks genetic features of PDAC or other aggressive forms of pancreatic carcinoma, should be considered a distinct entity rather than a variant of PDAC. However, genetic findings in the present case, and the cases previously reported by Mattiolo and colleagues suggest that it may be more appropriate to consider PHC a form of ectopic HCC rather than a new variant of pancreatic SPN. CTNNB1 mutations are frequently found both in SPN and in HCC. But HCC is known for genomic instability, frequently having copy number alterations in chromosomes 1, 18, 19, 20, 21, and others [15], similar to what is reported in the present PHC and in those previously reported by Mattiolo and colleagues.

Conclusion

We report the first instance of complete genomic and transcriptomic profiling of a case of pure PHC, which has not recurred following surgical resection. The pathologic and genetic features of the present PHC closely resemble HCC and are similar to previously reported cases of pure PHC. We propose that PHC, which resembles well to moderately differentiated HCC, which is confined to the pancreas, which is lymph node negative, and which lacks genetic features of PDAC and other aggressive forms of pancreatic carcinoma, could be considered ectopic HCC rather than a variant of PDAC. Cases of PHC like the one presently reported may have a good prognosis, similar to HCC completely resected from non-cirrhotic liver. Therapeutic regimens for HCC, including ICI therapy, may be efficacious in cases of PHC that recur or metastasize.

Patient Perspective

I am a generally healthy individual who maintains a regular workout routine and follows a balanced diet. Prior to mid-2022, I did not experience any significant health issues. However, I began experiencing stomach pain accompanied by a high fever, which was unusual for me in mid-2022. Despite the fever subsiding after 3 days, I felt all the symptoms were gone and my stomach back to normal.

In early 2023, I noticed irregularities in my menstrual cycle, including mild bleeding outside of my regular period. By mid-2023, my menstrual cycle ceased entirely. Concerned about these symptoms, I sought medical advice from my family doctor at that time since I also observed a noticeable mass in my abdomen, contributing to my discomfort and bloating sensation. During this period, I occasionally experienced stomach discomfort, frequent burping, and pain. Following several months of unsuccessful medication treatment for my stomach issues, I underwent a CT scan in late 2023, which revealed the presence of a tumour. Throughout this period, I experienced low-grade fevers and a diminished appetite, particularly exacerbated by fried or oily foods.

Acknowledgments

This work would not be possible without the participation of our patients and families, the POG team, the GSC platform and the generous support of the BC Cancer Foundation, Genome British Columbia (project B20POG), and the Terry Fox Research Institute Marathon of Hope Cancer Centres Network. We also acknowledge contributions towards equipment and infrastructure from Genome Canada and Genome BC (projects 202SEQ, 212SEQ, 262SEQ, 12002), Canada Foundation for Innovation (projects 20070, 30981, 30198, 33408 and 35444), and the BC Knowledge Development Fund. The results published here are in part based upon data generated by the following projects and obtained from dbGaP (http://www.ncbi.nlm.nih.gov/gap): The Cancer Genome Atlas managed by the NCI and NHGRI (http://cancergenome.nih.gov) and Genotype-Tissue Expression (GTEx) Project, supported by the Common Fund of the Office of the Director of the National Institutes of Health (https://commonfund.nih.gov/GTEx).

Statement of Ethics

This study protocol was reviewed and approved by University of British Columbia – BC Cancer Research Ethics Board (REB #H20-02317). The patient has provided written consent for publication of their case including medical details and associated photos.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

The funder had no role in the design, data collection, data analysis, and reporting of this study. This study was not supported by any sponsor or funder.

Author Contributions

S.P.: conceptualization, methodology, data collection, writing – original draft, and writing – edits. V.C.: methodology, data collection, writing – original draft, and writing – edits. C.F.: writing – original draft and writing – edits. R.G.: conceptualization and writing – edits. D.S.: conceptualization, writing – edits, and supervision. S.Y.: writing – edits. M.M. and J.J.: data collection and writing – edits. M.B.: conceptualization, methodology, writing – edits, and supervision. D.O.: conceptualization, methodology, data collection, writing – original draft, writing – edits, and supervision.

Funding Statement

The funder had no role in the design, data collection, data analysis, and reporting of this study. This study was not supported by any sponsor or funder.

Data Availability Statement

Project data can be made available upon reasonable request. Data are not available publicly as they are within the BC Cancer personalized oncogenetics database that contains identifying information and is being used in other ongoing projects. Please contact Lindsay Zibrik (lzibrik@bcgsc.ca) if data are needed.