Diagnostic dilemmas in hepatoid adenocarcinoma of the stomach: Navigating clinical and pathological loopholes

Abstract

Hepatoid adenocarcinoma of the stomach (HAS) is a rare and highly aggressive gastric cancer subtype characterized by hepatic morphological features and often elevated alpha-fetoprotein (AFP) levels. Despite its distinct biology HAS frequently eludes early diagnosis due to nonspecific gastrointestinal symptoms, radiological mimicry of hepatocellular carcinoma, and variable expression of diagnostic biomarkers. Imaging modalities like CT and magnetic resonance imaging offer limited specificity, often misinterpreting hepatic metastases as primary liver tumors. While elevated AFP is a supportive clue, it lacks sensitivity and specificity, particularly in AFP-negative HAS. Histologically, HAS poses challenges with subtle or focal hepatoid differentiation, especially in limited biopsy samples. Immunohistochemical variability, especially inconsistent staining of markers like SALL4, Glypican-3, and HepPar-1, further complicates accurate classification. Misdiagnosis can lead to inappropriate management and a worse prognosis. The lack of standardized diagnostic criteria globally exacerbates variability in reporting and treatment decisions. This review outlined the multidimensional diagnostic hurdles in HAS and underscored the necessity of an integrated, multidisciplinary approach. Enhanced tissue sampling, combined immunohistochemical panels, and emerging molecular tools may collectively improve diagnostic accuracy and clinical outcomes for this aggressive gastric malignancy.

Key Words: Hepatoid adenocarcinoma; Gastric cancer; Alpha-fetoprotein; Histopathology; Diagnostic pitfalls

Core Tip: Hepatoid adenocarcinoma of the stomach is a rare but aggressive tumor that mimics hepatocellular carcinoma both morphologically and serologically. Diagnostic delays occur due to nonspecific symptoms, radiological ambiguities, and variable biomarker profiles. A multidisciplinary and standardized diagnostic approach is critical for improving early detection and clinical outcomes.

INTRODUCTION

Hepatoid adenocarcinoma (HAC) of the stomach (HAS) is a rare and aggressive subtype of gastric carcinoma that displays morphological and functional characteristics similar to hepatocellular carcinoma (HCC). It was first described by Aizawa et al[1] in 1985 and is defined by its hepatoid differentiation, meaning its tumor cells resemble hepatic parenchymal cells, and often by the production of serum alpha-fetoprotein (AFP). Despite accounting for less than 1% of all gastric cancers, HAS is clinically significant due to its disproportionately poor prognosis, frequent liver metastases, and early systemic dissemination[2]. Patients diagnosed with HAS often present at advanced stages, and the 5-year survival rate is markedly lower compared with conventional gastric adenocarcinomas[3].

The biological behavior of the tumor is largely attributed to its high vascularity, tendency for hepatic and lymphatic invasion, and a propensity to mimic other malignancies, especially HCC. A crucial problem is that HAS lacks a universally accepted set of diagnostic criteria that has led to inconsistencies in diagnosis, underreporting, and heterogeneity in clinical management[4]. Morphologically, HAS may present as a pure hepatoid tumor or as a mixed type with adenocarcinomatous and hepatoid components. Immunophenotypically, it often expresses hepatic markers such as AFP, Glypican-3, and SALL4, but variability in marker expression frequently complicates diagnosis[5].

Given these challenges, HAS is often misclassified or diagnosed late, missing critical windows for surgical resection or targeted therapy. Moreover, over-reliance on serum AFP or imaging alone often leads to incorrect assumptions, especially in cases with hepatic involvement in which patients may be wrongly diagnosed with primary liver tumors. This review aimed to synthesize the multifaceted diagnostic challenges associated with HAS across clinical, imaging, pathological, and immunohistochemical (IHC) domains, emphasizing the urgent need for diagnostic vigilance and standardization[4,5].

MOLECULAR FEATURES

HAC displays distinct molecular alterations that while sharing some features with HCC diverge in critical genomic and transcriptomic aspects. Key molecular characteristics of HAC include frequent TP53 mutations, often associated with genomic instability and aggressive tumor behavior. In contrast CTNNB1 mutations, which activate WNT/β-catenin signaling, are more commonly observed in HCC and may be less prevalent in HAC, suggesting a divergence in tumorigenic pathways. Moreover, AFP gene amplification has been reported in some HAC cases, potentially contributing to the high serum AFP levels observed clinically and providing a molecular correlate to this diagnostic biomarker. In-depth analyses have also identified aberrant activation of oncogenic pathways such as PI3K/AKT, MAPK/ERK, and MYC although these findings require validation in larger cohorts. Elucidating these molecular signatures can aid in distinguishing HAC from morphologically similar neoplasms and offer targets for precision therapy[2,5,6].

CLINICAL PRESENTATION AND DIAGNOSTIC DELAY

The clinical presentation of HAS is nonspecific and overlaps with more common gastric disorders, contributing significantly to diagnostic delay. Most patients present with symptoms such as epigastric discomfort, nausea, vomiting, gastrointestinal bleeding, or weight loss, features common to gastritis, peptic ulcer disease, or conventional gastric carcinoma[6]. These vague manifestations often result in initial misdiagnosis or delayed referrals for endoscopic evaluation and biopsy. A review of pooled case series by Zeng et al[3] showed that patients with HAS frequently underwent extensive evaluations before an accurate diagnosis was made due to its ability to mimic benign conditions or standard adenocarcinomas[7].

Moreover, the tendency of HAS to metastasize early, especially to the liver, can mislead clinicians into suspecting a primary hepatic malignancy, particularly in AFP-positive patients. This delay in clinical suspicion and failure to include HAS in early differential diagnosis contributes to poorer prognostic outcomes[6,7].

IMAGING PITFALLS IN HAS DETECTION

Imaging modalities such as contrast-enhanced CT, magnetic resonance imaging (MRI), and positron emission tomography CT are routinely used for the staging of gastric cancers but have limited reliability in identifying HAS. Radiologically, HAS tumors often exhibit hypervascularity and intense enhancement in the arterial phase, resembling HCC[8]. This similarity is especially problematic in patients who present with hepatic metastases as their initial manifestation, leading to a common misinterpretation of the lesion as primary HCC with gastric metastasis rather than the reverse[8,9].

MRI features may include mosaic enhancement and tumor necrosis, but these are not specific. Positron emission tomography CT may fail to detect HAS lesions due to low fluorodeoxyglucose uptake, particularly in tumors with well-differentiated hepatoid features[9]. As a result imaging findings should always be interpreted alongside histological and serological data to avoid misdiagnosis.

SERUM AFP: A DOUBLE-EDGED SWORD

Serum AFP is a widely known tumor marker and was originally considered a hallmark of HAS. However, it functions as a double-edged sword. Approximately 70%-80% of patients with HAS have elevated serum AFP levels, but this marker lacks both specificity and sensitivity. Elevated AFP levels can also occur in HCC, yolk sac tumors, and even benign liver conditions such as cirrhosis and hepatitis[10].

More critically, a subset of HAS cases are AFP-negative, making AFP an unreliable solitary marker. The over-reliance on AFP levels may misguide clinicians toward a hepatic origin, particularly when combined with radiological findings of liver lesions. Therefore, while AFP can support the diagnosis, it must always be corroborated with histological and IHC evidence[10,11].

DNA METHYLATION PROFILING

Epigenetic profiling, particularly DNA methylation analysis, has emerged as a powerful tool for tumor classification and has shown promise in differentiating morphologically overlapping cancers. Though comprehensive methylation profiling studies specific to HAC are currently sparse, emerging data suggest that methylation signatures could effectively distinguish HAC from primary HCC and other AFP-producing neoplasms. DNA methylation-based classifiers, like those developed for central nervous system tumors and sarcomas, may be adapted to include hepatoid carcinomas, improving diagnostic precision. Preliminary studies have hinted at hypermethylation of tumor suppressor genes such as RASSF1A and CDKN2A, paralleling findings in gastric adenocarcinoma, one of the common primary sites of HAC. Integrating such methylation data into diagnostic workflows could allow earlier and more accurate classification of tumors with hepatoid morphology[4,5].

HISTOPATHOLOGICAL CHALLENGES

The histopathological identification of HAS is nuanced. The defining feature is the presence of hepatocytes resembling tumor polygonal cells with abundant eosinophilic cytoplasm and a trabecular or solid growth pattern. However, these characteristics are not unique to HAS and may overlap with poorly differentiated gastric adenocarcinomas or metastatic HCC[11].

Another challenge is the presence of mixed histology. Many HAS tumors contain both adenocarcinomatous and hepatoid components, often making the hepatoid area difficult to detect, especially in superficial biopsies. Pathologists must be adept at identifying these subtle changes and should consider further sampling or resection if hepatoid differentiation is suspected[11,12].

IHC: INCONSISTENCIES AND AMBIGUITIES

IHC is crucial in the diagnostic workup of HAS, but the interpretation of IHC panels can be inconsistent. The most frequently used markers, i.e. AFP, HepPar-1, Glypican-3, and SALL4, have varying expression profiles. SALL4, in particular, has emerged as the most reliable marker for hepatoid differentiation due to its consistent nuclear staining pattern in HAS cells[12].

However, SALL4 is also expressed in germ cell tumors, and HepPar-1 and Glypican-3 are not specific to HAS, leading to diagnostic overlap. In addition the absence of AFP staining does not rule out HAS. The lack of a standardized IHC panel across institutions further contributes to diagnostic ambiguity, underscoring the need for consensus on a minimal set of markers required for diagnosis.

BIOPSY LIMITATIONS AND TUMOR HETEROGENEITY

Endoscopic biopsies are often inadequate for diagnosing HAS because the hepatoid component tends to lie deeper in the submucosa or beyond. Superficial biopsies may capture only the conventional adenocarcinoma component, leading to missed diagnosis. This is particularly problematic in tumors with a heterogeneous architecture or mixed histology. Multiple and deeper biopsies, preferably guided by endoscopic ultrasound, are necessary to obtain representative samples. Even in surgical resections thorough sampling is needed to evaluate the full spectrum of differentiation as limited sections may fail to reveal the hepatoid areas[13].

DIFFERENTIAL DIAGNOSIS: A SOURCE OF MISCLASSIFICATION

The differential diagnosis of HAS includes metastatic HCC, yolk sac tumors, and poorly differentiated gastric adenocarcinomas. Distinguishing HAS from metastatic HCC is particularly challenging, especially in patients presenting with liver lesions and elevated AFP. In such cases an incorrect diagnosis may lead to a misdirected treatment plan targeting primary liver cancer instead of gastric origin. Yolk sac tumors can also present with SALL4 and AFP positivity, especially in young patients, adding another layer of complexity. A multidisciplinary approach integrating clinical history, radiological imaging, endoscopic evaluation, and comprehensive pathology is essential to accurately classify the tumor[14].

IMPACT OF LACK OF STANDARDIZED DIAGNOSTIC CRITERIA

One of the most critical challenges is the absence of standardized diagnostic criteria for HAS. Some studies define HAS based on histological morphology alone while others require concurrent AFP elevation. This lack of consensus leads to inconsistencies in reporting, complicates epidemiological studies, and results in heterogeneity in treatment decisions. Standardized criteria encompassing morphological, immunohistochemical, and possibly molecular features are necessary to ensure consistency in diagnosis, facilitate early detection, and enable accurate comparisons across clinical studies[15].

CLINICAL CONSEQUENCES OF MISDIAGNOSIS

Misdiagnosing HAS can have severe clinical consequences. Patients may receive suboptimal therapy due to underestimation of tumor aggressiveness or incorrect origin. HAS tends to metastasize early, and delays in diagnosis can result in the loss of opportunities for curative surgical resection. Furthermore, patients may be excluded from clinical trials designed for specific histological or molecular subtypes if HAS is not recognized[16].

ARTIFICIAL INTELLIGENCE APPLICATIONS

Artificial intelligence (AI) has revolutionized diagnostic pathology and radiology by enhancing pattern recognition, standardization, and diagnostic speed. In the context of differentiating HAC from HCC, AI-driven tools such as convolutional neural networks and radiomics offer the potential to analyze subtle morphological and imaging features that might elude human interpretation. For example, AI models trained on histopathological whole-slide images can learn to recognize the glandular-hepatoid hybrid morphology typical of HAC while radiomic analysis can extract high-dimensional imaging features from CT or MRI to distinguish between hepatic and extrahepatic tumors with hepatoid differentiation. Digital pathology platforms that integrate IHC with AI can also assist in biomarker quantification and classification, contributing to improved interobserver consistency. Although still in early stages for HAC, these technologies could form the backbone of precision diagnostic systems in the near future[17,18].

RECOMMENDATIONS FOR IMPROVING DIAGNOSTIC ACCURACY

To enhance diagnostic accuracy a structured diagnostic workflow should be adopted. Clinical suspicion should be raised in patients with gastric lesions and elevated AFP or hepatic metastases. Imaging should be interpreted cautiously, and deep biopsies should be prioritized. IHC panels should include SALL4, HepPar-1, Glypican-3, and AFP. Multidisciplinary tumor boards can facilitate diagnostic consensus. Future directions should also explore the role of molecular diagnostics and AI in pathology to help overcome human limitations and provide standardized assessment algorithms[17]. All the domains are summarized in Table 1.

Table 1 Diagnostic challenges and key features in hepatoid adenocarcinoma of the stomach.

Domain	Challenges/features	Clinical implications
Clinical presentation	Nonspecific symptoms: Epigastric discomfort, weight loss, GI bleeding. Mimics benign conditions and conventional gastric cancer	Delayed diagnosis and referral; missed opportunity for early intervention
Imaging	CT/MRI may resemble HCC due to arterial phase hyperenhancement; PET-CT has low sensitivity in well-differentiated tumors	High risk of misdiagnosing metastatic liver lesions as primary HCC
Serum AFP	Elevated in 70%-80% of HAS but also seen in HCC, yolk sac tumors, and benign liver conditions; some HAS cases are AFP-negative	AFP is supportive but unreliable alone; can lead to diagnostic misdirection
Histopathology	Mixed histological features; hepatoid areas may be focal and missed in small biopsies; overlaps with other poorly differentiated tumors	Requires experienced pathological evaluation and possibly repeated/deep biopsies
IHC	Variable expression of markers (AFP, HepPar-1, Glypican-3, SALL4); SALL4 is more reliable but not entirely specific	IHC is essential but must use a panel approach; no single definitive marker
Biopsy limitations	Superficial biopsies often miss deeper hepatoid components; tumor heterogeneity adds complexity	Multiple and deeper biopsies (e.g., EUS-guided) recommended for accurate sampling
Differential diagnosis	Overlaps with metastatic HCC, yolk sac tumors, and poorly differentiated gastric adenocarcinomas	Multidisciplinary approach is crucial to avoid misclassification and mistreatment
Lack of standardized criteria	No universally accepted diagnostic benchmarks (e.g., reliance on morphology vs AFP vs IHC); variable institutional practices	Inconsistent diagnosis, hindered research comparisons, and variable treatment strategies
Misdiagnosis consequences	Incorrect treatment strategy (e.g., targeting liver primary), delayed curative options, clinical trial exclusion	Worsened prognosis and lost therapeutic opportunities
Recommendations	Use structured workflow: Clinical suspicion; imaging + deep biopsy; full IHC panel (SALL4, AFP, HepPar-1, Glypican-3); and multidisciplinary review; explore molecular diagnostics	Improves early detection, diagnostic accuracy, and patient outcomes

IHC: Immunohistochemistry; MRI: Magnetic resonance imaging; HCC: Hepatoid adenocarcinoma; PET: Positron emission tomography; EUS: Endoscopic ultrasound; AFP: Alpha-fetoprotein; GI: Gastrointestinal; HAS: Hepatoid adenocarcinoma of the stomach.

CLINICAL DIAGNOSTIC APPROACH

A structured clinical algorithm can greatly aid in the accurate diagnosis of HAC, which often presents diagnostic challenges due to its morphological overlap with HCC and other adenocarcinomas. A systematic approach should begin with clinical history and risk factors, followed by serum biomarker assessment (most notably AFP, which is elevated in the majority of HACs but also in HCC). Imaging studies (CT, MRI) can localize lesions and assess liver involvement. Histological examination remains central with emphasis on hepatoid differentiation. IHC is critical for confirmation with markers like HepPar1, AFP, SALL4, Glypican-3, and CK7/CK20 panels aiding in differential diagnosis. Incorporating molecular profiling (e.g., TP53 mutations, AFP amplification) and emerging epigenetic or AI-based tools into a diagnostic flowchart could improve accuracy. A schematic algorithm would offer clinicians a visual guide integrating these multiple diagnostic layers to streamline decision-making[15-17] (Table 1).

CONCLUSION

HAS represents a diagnostic challenge due to its overlapping features with other malignancies, variable biomarker expression, and lack of standardized criteria. The diagnostic process is further hampered by nonspecific clinical symptoms, imaging ambiguity, and biopsy limitations. An integrative diagnostic approach involving clinical, radiological, histological, and IHC data is essential. Standardization in diagnostic criteria and the development of molecular tools may pave the way for more accurate and timely diagnosis, ultimately improving patient outcomes.