The distinction between amphicrine-like carcinoma and mixed neuroendocrine-non-neuroendocrine neoplasms (MiNENs) is more than academic. MiNENs consist of two morphologically distinct tumor populations, each with its own identity, whereas ALCs demonstrate dual differentiation within the same neoplastic cells or within a diffusely admixed population. The current WHO Classification of Tumours, 6th edition, recognizes amphicrine differentiation as a distinct pathological phenomenon and emphasizes the importance of integrating morphology with immunohistochemical findings to establish the diagnosis. Getting this classification right matters because the two entities can behave differently and may call for different therapeutic strategies.

The case began with a 52-year-old woman referred to Kantonsspital Winterthur in Switzerland with several months of persistent upper abdominal pain, bloating, and unintentional weight loss. Her symptoms had initially improved with pantoprazole therapy but subsequently worsened. Laboratory investigations showed anemia, with reduced hemoglobin of 116 g/L, mean corpuscular volume of 62 fL, and mean corpuscular hemoglobin of 21 pg, alongside elevated liver enzymes including AST of 57 U/L, ALT of 96 U/L, GGT of 235 U/L, and alkaline phosphatase of 151 U/L. Tumor markers were largely unremarkable, with only a mild elevation in neuron-specific enolase at 15.2 μg/L, while CA 19-9, CEA, and chromogranin A remained within normal ranges.

Contrast-enhanced computed tomography revealed a mass located between the pancreatic head and the first portion of the duodenum, with an additional lesion suggestive of an enlarged lymph node cranial to the pancreatic head between the portal vein and inferior vena cava. A suspected liver metastasis in segment III and a nonspecific pulmonary nodule in the right lower lobe were also noted. Subsequent 18F-FDG PET/CT demonstrated increased metabolic activity in the pancreaticoduodenal region, with a large hypermetabolic mass measuring 2.3 by 4.8 centimeters between the portal vein and inferior vena cava, several smaller metabolically active peripancreatic lymph nodes, a left supraclavicular lymph node, multiple pulmonary nodules in both lungs, and a liver lesion in segment III all showing increased tracer uptake. Magnetic resonance imaging of the liver confirmed a highly suspicious lesion in segment II and identified three additional small lesions consistent with metastatic disease.

Endoscopic ultrasound-guided fine-needle aspiration was performed on the pancreatic mass and the portocaval lymph node, and a liver biopsy was obtained from the segment III lesion. The aspirates showed malignant cells consistent with a poorly differentiated carcinoma, forming complex cellular clusters of monomorphic to pleomorphic cells with predominantly medium-sized hyperchromatic nuclei, occasional prominent nucleoli, and foamy cytoplasm. Immunocytochemistry demonstrated strong CK19 positivity, membranous and cytoplasmic beta-catenin expression without nuclear localization, weak diffuse synaptophysin expression, focal trypsin positivity, and negative SOX11 staining.

Histopathological examination revealed a solid poorly differentiated carcinoma with a proliferative index approaching 90 percent, as measured by Ki-67 staining. Immunohistochemical analysis showed strong diffuse expression of synaptophysin, partial expression of trypsin and chromogranin A, absence of CDX2, loss of p53 expression, retained ATRX expression, and expression of BCL10 in tumor cells. The tumor showed marked mitotic activity, with up to 10 mitoses per high-power field. Based on the initial findings, including evidence of both acinar and neuroendocrine differentiation, the tumor was first diagnosed as a mixed acinar-neuroendocrine carcinoma. Detailed evaluation, however, demonstrated overlapping acinar and neuroendocrine differentiation throughout the tumor, with expression of acinar markers trypsin and BCL10 together with neuroendocrine markers synaptophysin and chromogranin A, without evidence of two spatially distinct components. According to current diagnostic criteria, the tumor was therefore reclassified as pancreatic amphicrine-like carcinoma with acinar differentiation.

Comprehensive genomic profiling using next-generation sequencing with the FoundationOne CDx platform identified several alterations, including the RAF1 fusion, IRF2 loss, PBRM1 exon 30 loss, SMAD4 L533P, and TP53 Q52*. Genomic analysis demonstrated microsatellite stability and a tumor mutational burden of 5 mutations per megabase. Testing of predictive biomarkers by immunohistochemistry showed no evidence of mismatch repair deficiency. The KANK4::RAF1 fusion represents the principal molecular feature of this case. RAF1 rearrangements have been described in a subset of pancreatic acinar cell carcinomas and are considered recurrent genomic events in that tumor type. These rearrangements are thought to promote constitutive activation of the MAPK signaling pathway through disruption of normal RAF1 regulatory mechanisms while preserving the C-terminal RAF1 kinase domain. In this case, the identified fusion was predicted to retain the RAF1 kinase domain, providing a biologically plausible mechanism for activation of RAF1-dependent downstream signaling.

First-line palliative chemotherapy with FOLFIRINOX, a regimen combining folinic acid, 5-fluorouracil, irinotecan, and oxaliplatin, was initiated because it is considered effective against both tumor components. After four cycles, the patient developed a marked elevation in liver transaminases. Endoscopic retrograde cholangiopancreatography demonstrated obstruction of the distal common bile duct caused by the pancreatic head mass, with associated dilation of the intrahepatic and extrahepatic bile ducts. A papillotomy was performed and a covered metal stent was placed, leading to normalization of liver enzyme levels. Restaging computed tomography after four cycles showed disease progression, with an increase in hepatic, pulmonary, and lymph node metastases. In light of the rapid progression, a predominance of the neuroendocrine carcinoma component was suspected, and second-line palliative chemotherapy with carboplatin and etoposide was initiated. After three cycles, the primary pancreatic tumor and liver metastases remained stable, but pulmonary and lymph node metastases continued to progress. Due to persistent upper abdominal pain refractory to analgesics, CT-guided percutaneous neurolysis of the coeliac plexus with alcohol was performed.

Given the presence of a RAF1 fusion and emerging evidence from published case reports, third-line palliative treatment with trametinib, a MEK inhibitor, was initiated. RAF1 fusion proteins may activate downstream MAPK signaling through RAF-MEK-ERK pathway stimulation, providing a biological rationale for therapeutic inhibition of this pathway. However, therapy had to be discontinued after only one month because of significant treatment-related toxicity, including diarrhea, nausea, fatigue, and cutaneous adverse effects. The short duration of exposure precluded meaningful assessment of treatment efficacy, meaning this case neither supports nor refutes the effectiveness of MEK inhibition in RAF1-rearranged pancreatic neoplasms. Nevertheless, together with emerging reports of responses to MAPK pathway inhibitors in selected pancreatic acinar carcinomas, the case illustrates the potential clinical relevance of identifying actionable genomic alterations in rare pancreatic tumors. Following treatment discontinuation, the focus of care shifted to best supportive care, and the patient died 14 months after the initial diagnosis. No autopsy was performed.

The authors emphasize that pancreatic neoplasms exhibiting combined acinar and neuroendocrine differentiation are exceptionally rare and remain diagnostically challenging. Additional pathogenic alterations involving TP53, SMAD4, PBRM1, and IRF2 were detected, and although alterations in these genes have been implicated in tumor progression, chromatin remodeling, and immune regulation in various malignancies, their specific contribution to the biology of pancreatic ALC with acinar differentiation remains uncertain. Given the single-case nature of this report, no conclusions regarding their biological significance, prognostic relevance, or potential impact on treatment response can be drawn. The identification of the rare KANK4::RAF1 fusion broadens the molecular landscape of these uncommon tumors and supports consideration of comprehensive genomic profiling in the diagnostic evaluation of rare pancreatic neoplasms, since such profiling may reveal potentially actionable molecular alterations even in tumor entities for which established targeted treatment strategies are currently lacking.

Subject of Research: A rare pancreatic amphicrine-like carcinoma with acinar differentiation harboring a KANK4::RAF1 gene fusion

Article Title: Case Report: Pancreatic amphicrine-like carcinoma with acinar differentiation harboring a KANK4-RAF1 gene fusion

Article References: Kammerer, S., Bode, P., Schreiber, H., & Ortega Sanchez, G. (2026). Case Report: Pancreatic amphicrine-like carcinoma with acinar differentiation harboring a KANK4-RAF1 gene fusion. Pathology & Oncology Research, 32, Article 1612537. https://doi.org/10.3389/pore.2026.1612537

Image Credits: AI Generated

DOI: 10.3389/pore.2026.1612537

Keywords: pancreatic cancer, amphicrine-like carcinoma, KANK4-RAF1 fusion, RAF1 rearrangement, acinar differentiation, MiNEN, trametinib, comprehensive genomic profiling, MAPK pathway, rare tumors, Case, Report

The patient at the center of the report was a 61-year-old man with a history of bullous pulmonary emphysema who underwent a routine chest computed tomography scan in June 2020. The scan revealed a small mass, measuring 15 by 19 millimeters, in the upper segments of his left lung, along with a second smaller lesion in the right lung. Over the following months the left-sided mass grew steadily, and by November 2021 it had invaded the anterior chest wall and the soft tissues of the neck, destroying the first rib. An ultrasound-guided biopsy of the lesion yielded a diagnosis of metastasis from a poorly differentiated adenocarcinoma, thought at the time to originate from a hepatocellular carcinoma.

What followed was a grim demonstration of the tumor’s chemoresistance. The patient received successive lines of chemotherapy—cisplatin with docetaxel, then irinotecan, then gemcitabine, then cisplatin with etoposide—and every regimen failed, with the primary lesion expanding relentlessly from 22 by 23 millimeters to 112 by 64 millimeters over the course of treatment. A palliative course of external beam radiotherapy delivered to the primary tumor and to involved lymph nodes in the left supraclavicular region produced only a modest 29 percent shrinkage. By the time the patient was reassessed, the tumor had become a Pancoast-type cancer of the lung apex, encasing the left subclavian and vertebral arteries and eroding the first rib and adjacent vertebrae.

A careful re-examination of the biopsy material, however, changed the trajectory of the case. Pathologists identified epithelioid cells arranged in nested clusters, expressing cytokeratins 7 and 19 and the hepatocellular marker HepPar-1, while lacking p63, TTF-1, and CDX2—an immunophenotype characteristic of hepatoid adenocarcinoma. Two findings stood out. First, the tumor had lost expression of PMS2, one of the four essential proteins of the DNA mismatch repair system, while retaining MSH6. Second, the tumor cells expressed PD-L1, the molecular target of checkpoint-inhibiting drugs, with a combined positive score of 17. The investigators noted that the specimen was tiny, containing fewer than 100 tumor cells, which placed the PD-L1 assessment at the borderline of reliability, but the overall picture was enough to justify a bold therapeutic decision.

In October 2022, one month after completing radiotherapy, the patient began treatment with pembrolizumab, an antibody that blocks the PD-1 receptor and reinvigorates anti-tumor T cells. The effect was rapid and substantial. After just three infusions, the apical tumor in the left lung had regressed by 45.4 percent, shrinking to 74 by 45 millimeters. Treatment continued, and after 18 cycles the lesion measured 60 by 46 millimeters. More remarkably, a surgical biopsy of a residual supraclavicular lesion performed in November 2023 found only loose connective tissue and foci of necrosis, with no tumor cells whatsoever—a complete morphological response. The right lung lesion, considered a metastasis, diminished by 77.1 percent during therapy, although it was never biopsied, so systemic response is probable but not histologically confirmed.

To understand why the tumor had lost PMS2 and why it might be so immunogenic, the team turned to high-throughput sequencing. Sequencing of the patient’s constitutional DNA from blood leukocytes revealed no pathogenic variants in mismatch repair genes, and MLPA analysis found no deletions or aberrant methylation of MLH1, MSH2, PMS2, or MSH6, effectively excluding Lynch syndrome, the hereditary condition most commonly associated with mismatch repair deficiency. The answer lay in the tumor itself. Panel sequencing of tumor DNA, performed at a median depth of 1515-fold, uncovered biallelic somatic PMS2 mutations—a truncating variant, p.Arg421Ter, and a missense variant, p.Arg107Trp—that together explained the loss of PMS2 protein expression.

But the most consequential discovery was a third variant: POLE p.Pro286Arg, a mutation in the exonuclease domain of DNA polymerase epsilon, the enzyme responsible for proofreading newly synthesized DNA during replication. This substitution, with a variant allele frequency of 13 percent, is a well-known hotspot mutation in endometrial and colorectal cancers, where it disables the polymerase’s proofreading function and unleashes an avalanche of secondary somatic mutations across the genome. In POLE-mutated endometrial carcinomas carrying precisely this substitution, tumor mutational burdens ranging from 78 to 277 mutations per megabase have been reported, with a median of 169; in colorectal cancers with exonuclease domain mutations, burdens averaging 115 mutations per megabase have been documented. Such hypermutated tumors generate abundant neoantigens, making them exceptionally visible to the immune system and highly responsive to checkpoint blockade.

The sequencing also revealed a constellation of additional oncogenic alterations: two TP53 mutations (p.Arg282Trp and p.Arg175Cys), a truncating ARID1B variant, and a PTEN p.Leu320Ser substitution—a mutational profile strikingly similar to that seen in POLE-mutated, so-called multiple-classifier endometrial cancers, where proofreading deficiency secondarily drives inactivation of mismatch repair genes and TP53. The authors propose a coherent mechanistic sequence: the POLE P286R mutation crippled DNA proofreading, the resulting hypermutation disabled PMS2 through biallelic somatic hits, and the accumulated mutational load—combined with PD-L1 expression—rendered the tumor exquisitely sensitive to pembrolizumab. Although the small sequencing panel, spanning only 2.9 kilobases, prevented precise measurement of the tumor mutational burden, the hypermutated phenotype was strongly anticipated from the spectrum of variants found.

The significance of this finding is amplified by the broader genomic context of hepatoid adenocarcinoma of the lung. Only about 191 cases have been described in the literature, and the most comprehensive genomic profiling to date, encompassing 32 sequenced cases, identified 31 recurrently mutated genes including TP53, STK11, SMARCA4, CDKN2A, KRAS, and CDK8. Crucially, reported tumor mutational burdens in this disease have been low, ranging from 1.69 to 8.67 mutations per megabase, and most tumors are microsatellite-stable—features that ordinarily predict poor responsiveness to immunotherapy. Moreover, STK11 inactivation, common in these tumors, is known to create an immunologically cold, T-cell-excluded microenvironment, and the KRAS-STK11 co-mutation is considered among the most immunosuppressive genomic combinations in all of lung cancer. Against this backdrop, a hepatoid adenocarcinoma driven by POLE proofreading deficiency represents an entirely different biological entity, one whose immunogenicity rivals that of the best-responding hypermutated tumors.

The clinical outcome speaks for itself. To the date of last follow-up in June 2026, the patient had received 61 cycles of pembrolizumab—more than three and a half years of continuous disease control—with no adverse events and a good quality of life, a stark contrast to the fulminant progression that characterized his chemotherapy era. The authors suggest that the synergy between radiotherapy and checkpoint inhibition, initiated just one month apart, may have contributed to the response, and they argue that POLE exonuclease domain mutations should be regarded as a tissue-agnostic predictor of high tumor immunogenicity and checkpoint blockade effectiveness. For a tumor type with no standardized treatment algorithm and a historically dismal prognosis, the message of this single case is clear: individual molecular profiling, even in the rarest of cancers, can reveal therapeutic vulnerabilities that no empirical regimen could ever uncover.

Subject of Research: POLE P286R-mutated hepatoid adenocarcinoma of the lung responding to immune checkpoint inhibition

Article Title: A POLE P286R‐Mutated Hepatoid Adenocarcinoma of the Lung Showing Substantial Response to Immune Checkpoint Inhibition

Article References: Chernev, A., Kaurtseva, A., Abasov, R., Enikeev, R., Gegelia, N., Abdrakova, G., Usman, N., Druy, A., & Raskin, G. (2026). A POLE P286R ‐Mutated Hepatoid Adenocarcinoma of the Lung Showing Substantial Response to Immune Checkpoint Inhibition. Cancer Reports, 9(9), Article e70661. https://doi.org/10.1002/cnr2.70661

Image Credits: AI Generated

DOI: 10.1002/cnr2.70661

Keywords: hepatoid adenocarcinoma, lung cancer, POLE P286R, pembrolizumab, mismatch repair deficiency, PMS2 loss, PD-L1, tumor mutational burden, immunotherapy, DNA polymerase epsilon, rare tumors, case report