Poorly differentiated adenocarcinoma misdiagnosed as gastrointestinal stromal tumor: A case report

Abstract

BACKGROUND

Gastric submucosal tumors (SMTs) comprise a heterogeneous group of lesions arising beneath the mucosal layer, among which gastrointestinal stromal tumors (GISTs) are the most common mesenchymal neoplasms. However, some epithelial malignancies with predominant submucosal growth or atypical endoscopic manifestations may closely mimic SMTs, creating substantial diagnostic difficulty and increasing the risk of misdiagnosis when tissue sampling is limited or histologic interpretation is challenging.

CASE SUMMARY

A 58-year-old woman presented with recurrent epigastric discomfort for 1 month. Gastroscopy revealed a 1.5-cm SMT-like lesion with surface erosion and blood on the posterior wall at the junction of the gastric body and antrum. Endoscopic ultrasound demonstrated a hypoechoic lesion apparently originating from the muscularis propria, and diagnostic endoscopic submucosal dissection biopsy showed spindle cell-like morphology, leading to an initial impression of GIST. However, postoperative examination of the surgical specimen, together with immunohistochemistry, established the diagnosis of ulcerative poorly differentiated adenocarcinoma with invasion into the deep muscularis propria and metastasis in 2 of 28 regional lymph nodes. The patient subsequently received five cycles of adjuvant chemotherapy with the SOX regimen and showed no evidence of recurrence at the latest follow-up.

CONCLUSION

Suspected GIST with erosion or ulceration requires multimodal evaluation to confirm diagnosis and exclude epithelial malignancy.

Key Words: Gastrointestinal stromal tumor; Adenocarcinoma; Endoscopic submucosal dissection; Pathology; Case report

Core Tip: Gastric adenocarcinoma with predominant submucosal growth may closely mimic gastrointestinal stromal tumor. In this case, a small submucosal tumor-like gastric lesion with surface erosion and bleeding was misdiagnosed as gastrointestinal stromal tumor because endoscopic ultrasound suggested a muscularis propria-derived lesion and diagnostic endoscopic submucosal dissection biopsy showed spindle cell-like morphology. The final diagnosis of poorly differentiated adenocarcinoma was made after surgical resection. Small submucosal tumor-like lesions with mucosal abnormalities require multimodal evaluation, adequate tissue acquisition, and timely immunohistochemical assessment to reduce misdiagnosis.

INTRODUCTION

Gastrointestinal stromal tumor (GIST) is the most common mesenchymal neoplasm of the digestive tract. Approximately 60%-70% of patients with GIST present with abdominal discomfort, and 30%-40% present with bleeding[1,2]. Common diagnostic methods for gastric submucosal tumors (SMTs) include white-light endoscopy, endoscopic ultrasound (EUS), computed tomography (CT), and pathology, among which EUS plays a central role in evaluating the layer of origin and internal echogenicity[3]. Gastric adenocarcinoma is an epithelial malignancy, and poorly differentiated subtypes account for approximately 20%-54% of gastric cancers[4]. Although gastric adenocarcinoma usually arises from the mucosa, some lesions with predominant submucosal growth may mimic SMTs, creating substantial diagnostic difficulty.

We report a case of poorly differentiated adenocarcinoma initially misdiagnosed as GIST. EUS revealed a 1.5-cm submucosal lesion on the posterior wall at the junction of the gastric body and antrum. Diagnostic endoscopic submucosal dissection (ESD) biopsy initially suggested GIST, whereas postoperative pathology after surgical resection confirmed poorly differentiated adenocarcinoma. This case highlights the need for greater diagnostic caution in small SMT-like lesions with atypical mucosal abnormalities.

CASE PRESENTATION

Chief complaints

A 58-year-old woman was hospitalized owing to a 1-month history of recurrent pain in the central abdomen and epigastric region.

History of present illness

The patient presented with a 1-month history of recurrent pain in the central abdomen and epigastric region. She did not report any history of hematemesis, melena, weight loss, or fatigue. Gastroscopy and EUS revealed a submucosal lesion on the posterior wall at the junction of the gastric body and antrum, suggestive of GIST. Diagnostic ESD followed by biopsy showed spindle cell-like morphology, leading to a provisional diagnosis of GIST, for which surgery was recommended. The patient was then admitted to the surgical ward for endoscopy-guided laparoscopic gastrectomy. Postoperative pathology confirmed ulcerative poorly differentiated adenocarcinoma measuring 2.5 cm × 1.2 cm × 1 cm, with metastasis in 2 of 28 regional lymph nodes, including 2 of 12 lymph nodes along the lesser curvature.

History of past illness

The patient exhibited no pertinent medical background of significance.

Personal and family history

The patient had no smoking or alcohol usage background, nor any reported familial malignancy heritage.

Physical examination

No irregularities were detected during the physical assessment upon admission.

Laboratory examinations

The results of a complete blood count, coagulation profile (including D-dimer), thyroid panel (five tests), blood biochemistry, and tumor markers [carcinoembryonic antigen (CEA), alpha-fetoprotein, CA125, squamous cell carcinoma antigen, and CA153] were all within normal limits.

Imaging examinations

Gastroscopy revealed a submucosal elevation on the posterior wall at the junction of the gastric body and antrum, measuring approximately 1.5 cm in diameter, presenting as a hemispherical protrusion with erosions and fresh blood at its apex (Figure 1A). Biopsy pathology revealed chronic active mucosal inflammation with geographic necrosis and exudate, and focal glandular epithelial hyperplasia. EUS showed a submucosal elevation on the posterior wall at the junction of the gastric body and antrum, with ulceration at its apex (Figure 1B). With intraluminal water-injection-assisted ultrasound imaging, EUS revealed mucosal layer loss at the elevation site (Figure 1C). The lesion originated from a hypoechoic structure within the muscularis propria, with indistinct borders and no posterior acoustic shadowing. Cross-sectional measurement was not feasible. The surrounding gastric wall structures were intact. The ultrasonography findings suggested a possible GIST. Abdominal contrast-enhanced CT revealed gastric wall thickening on the posterior wall near the junction of the gastric body and antrum, with no obvious enlargement of lymph nodes in the surrounding area (Figure 1D).

Figure 1 Endoscopic and computed tomography-related images. A: Gastroscopy showing a hemispherical mucosal elevation on the posterior wall at the junction of the gastric body and antrum; B: Endoscopic ultrasound showing a submucosal elevation with ulceration at the apex on the posterior wall at the junction of the gastric body and antrum; C: Intraluminal water-injection-assisted endoscopic ultrasound imaging showing mucosal layer loss at the elevation site; D: Abdominal contrast-enhanced computed tomography showing gastric wall thickening on the posterior wall near the junction of the gastric body and antrum.

FINAL DIAGNOSIS

Postoperative pathology and immunohistochemistry confirmed ulcerative poorly differentiated adenocarcinoma (pT2N1M0, stage IIA).

TREATMENT

After the imaging investigations, the patient and her family were informed that the imaging findings suggested a possible GIST, but that the possibility of a malignancy could not be ruled out. The patient was advised to undergo a biopsy or diagnostic ESD to clarify the nature of the lesion or to proceed directly to surgical gastrectomy. After discussion, the family opted for diagnostic ESD.

Intraoperatively, a 1.5 cm submucosal lesion was identified on the posterior wall at the junction of the gastric body and antrum, with central ulceration and minimal bleeding. The lesion margins were marked using an electrosurgical knife, and a mixture of epinephrine, saline, and methylene blue was injected submucosally. The lesion showed minimal elevation. The mucosal layer was incised with an electrosurgical knife to expose the submucosal lesion, which revealed pale yellow, relatively dense tissue.

After consultation with the patient's family, the decision was made to discontinue further ESD and perform a diagnostic tissue biopsy to confirm the nature of the lesion. Histopathologic examination of the diagnostic ESD biopsy showed spindle cell-like morphology, leading to an initial impression of a spindle cell tumor (Figure 2). Based on the pathological findings and the challenge of achieving complete resection with ESD, gastrectomy was recommended.

Figure 2 Diagnostic endoscopic submucosal dissection biopsy pathology. A: Hematoxylin and eosin staining of the biopsy specimen from the posterior wall at the junction of the gastric body and antrum, showing a spindle cell-rich lesion (× 100); B: Higher-power hematoxylin and eosin view showing spindle cell-like morphology with elongated tumor cells (× 200).

The patient and her family consented to surgery. One week after ESD, laparoscopic gastrectomy was performed under endoscopic guidance. Intraoperatively, the serosa over the posterior wall at the junction of the gastric body and antrum appeared smooth, and endoscopy aided localization. The resected specimen measured roughly 2.5 cm × 1.5 cm (Figure 3A). Frozen section analysis identified poorly differentiated carcinoma (tumor size 2.5 cm × 1.2 cm × 1 cm) in the posterior wall at the junction of the gastric body and antrum, infiltrating the deep muscular layer (Figure 3B). After discussing the findings with the family, a radical gastrectomy was performed, including subtotal gastrectomy with gastrojejunostomy and peri-gastric lymph node dissection.

Figure 3 Surgical specimens. A: Gross view of the resected stomach showing a lesion measuring approximately 2.5 cm × 1.2 cm × 1 cm; B: Opened gastrectomy specimen showing an ulcerative carcinoma located on the posterior wall at the junction of the gastric body and antrum, with clear anatomical orientation.

Histopathological examination of the surgical specimen demonstrated infiltrative tumor growth and periodic acid-Schiff-positive mucin within tumor cells and glandular lumina (Figure 4). Histopathological examination of the resected stomach revealed an ulcerative poorly differentiated adenocarcinoma on the posterior wall at the junction of the gastric body and antrum, measuring 2.5 cm × 1.2 cm × 1 cm, predominantly poorly cohesive, with invasion into the deep muscularis propria and focal perineural invasion. A total of 28 lymph nodes were examined, including 7 along the greater curvature, 5 in the infrapyloric region, 12 along the lesser curvature, and 4 in group 8. Metastatic carcinoma was identified in 2 of 12 lymph nodes along the lesser curvature, with no metastasis in the lymph nodes along the greater curvature (0/7), infrapyloric region (0/5), or group 8 (0/4). Some ectopic pancreatic tissue was also noted.

Figure 4 Surgical histopathology. A: Periodic acid-Schiff (PAS) staining showing infiltrative tumor growth in the gastric wall (× 100); B: Higher-power PAS staining showing PAS-positive mucin in tumor cells and glandular lumina (× 200).

Immunohistochemistry for tumor markers in the main tumor revealed: Anti-epithelial antigen 1/anti-epithelial antigen 3 (+), cytokeratin (CK) antibody monoclonal 5.2 (+), CD31 (vascular endothelial +), CD34 (vascular endothelial +), CD56 (-), CEA (partial +), CK7 (+), CK8/18 (+), chromogranin A (scant +), D2-40 (lymphatic vessel +), desmin (smooth muscle +), epithelial membrane antigen (+), H-caldesmon (smooth muscle +), human epidermal growth factor receptor 2 (-), Ki67 (MIB-1) (60% +), mucin-5AC (+), mucin-6 (+), P53 (-), S100 (neural invasion), Syn (faint +), Topo II (10% +), MutL homolog 1 (+), MutS homologs 2 (+), MutS homologs 6 (+), PMS1 homolog 2 (+). Immunohistochemistry of the metastatic lymph nodes revealed: Anti-epithelial antigen 1/anti-epithelial antigen 3 (+), CK7 (+), CK8/18 (+), human epidermal growth factor receptor 2 (-). Epstein-Barr virus-encoded RNA in situ hybridization was negative, indicating an absence of Epstein-Barr virus, and all four mismatch-repair proteins were positive, confirming intact mismatch-repair function.

The pathological features observed in the initial ESD biopsy and the final surgical resection specimen are summarized and compared in Table 1. After receiving postoperative fluid replacement, nutritional support, and other symptomatic treatments, the patient was discharged. After surgery, the patient received five cycles of adjuvant chemotherapy with the SOX regimen.

Table 1 Comparison of pathological findings between endoscopic submucosal dissection biopsy and surgical specimen.

Item	ESD biopsy specimen (initial assessment)	Surgical resection specimen (final diagnosis)
Specimen type	Limited endoscopic biopsy (partial submucosal tissue)	Full-thickness gastrectomy specimen
Tumor architecture	Fragmented tissue with spindle cell-like components	Infiltrative growth involving mucosa, submucosa, and muscularis propria
Cellular morphology	Elongated cells with hyperchromatic nuclei; limited structural context	Poorly differentiated adenocarcinoma with marked atypia
Mucosal involvement	Not adequately assessed	Definite ulceration with mucosal invasion
Depth of evaluation	Superficial and limited	Complete assessment of tumor depth
Immunohistochemistry	Not performed at initial stage	Tumor cells positive for epithelial markers (anti-epithelial antigen 1/anti-epithelial antigen 3, CK antibody monoclonal 5.2, CK7, CK8/18, epithelial membrane antigen, partial CEA); CD34 staining limited to vascular endothelial cells; CD117 and discovered on GIST-1 not performed
Diagnostic interpretation	Suggestive of gastrointestinal stromal tumor	Confirmed poorly differentiated adenocarcinoma

ESD: Endoscopic submucosal dissection; CK: Cytokeratin; CEA: Carcinoembryonic antigen; GIST: Gastrointestinal stromal tumor.

OUTCOME AND FOLLOW-UP

The patient underwent follow-up with contrast-enhanced abdominal CT every 6 months and gastroscopy annually after surgery. At the latest follow-up in March 2026, no tumor recurrence had been detected.

DISCUSSION

SMTs are a heterogeneous group of lesions, among which GISTs are the most common mesenchymal neoplasms of the stomach[5]; however, accurate preoperative differentiation between GIST and epithelial malignancies with atypical growth patterns remains challenging, even with EUS and tissue acquisition[6].

Previous reports have shown that gastric adenocarcinoma may occasionally present as an SMT-like lesion and be clinically suspected or misdiagnosed as GIST, as summarized in Table 2. Compared with previously reported cases, the present case is clinically distinctive because it combined several deceptive features in a single small lesion: Surface erosion and bleeding, a hypoechoic lesion apparently arising from the muscularis propria on EUS, and misleading spindle cell-like morphology on diagnostic ESD biopsy. These features created a convincing but incorrect preoperative impression of GIST and illustrate the diagnostic workflow challenges that may arise in atypical gastric SMT-like lesions[7-9].

Table 2 Previous cases of gastric adenocarcinoma suspected or misdiagnosed as gastrointestinal stromal tumor.

Ref.	Age/sex	Endoscopic findings	EUS findings	Diagnostic method	Final diagnosis
Hu et al[24], 2023	64/F	3-cm round SMT-like mass with smooth mucosa	Heterogeneous hypoechoic mass (MP), subtle layer obliteration	ESD	Moderately to poorly differentiated adenocarcinoma
Chen et al[25], 2019	48/F	2-cm SMT near cardia with central ulceration and exposed vessel	Not described	Intraoperative frozen section	Moderately differentiated adenocarcinoma
Liu et al[26], 2025	60/M	2-cm SMT-like mass at the lesser curvature, smooth overlying mucosa	2.0 cm × 0.8 cm well-defined hypoechoic mass (MP)	ESD	Poorly differentiated adenocarcinoma
Lee and Oh[27], 2020	64/F	2-cm lobulated SMT with normal mucosa	19 mm × 11 mm heterogeneous hypoechoic mass (MP), well-circumscribed	Intraoperative frozen biopsy	Signet ring cell adenocarcinoma
Kim et al[28], 2012	57/M	10 cm × 12 cm intraluminal protruding lesion with normal mucosa	Heterogeneous mass (MP) with hyperechoic deposits and necrotic zones	EUS-guided trucut biopsy	Signet ring cell adenocarcinoma
Shin et al[29], 2015	59/F	8-cm protuberant SMT with slight ulceration	Not described	Postoperative pathology	Poorly differentiated tubular adenocarcinoma
Shin et al[29], 2015	68/M	4.8-cm protruding SMT with central ulceration	Not described	Laparoscopic biopsy	Mucinous adenocarcinoma
Li et al[30], 2016	44/M	1-cm sessile polypoid SMT-like lesion with erosion	0.7 cm × 0.8 cm well-defined hypoechoic mass (MP)	ESD with IHC	Poorly differentiated adenocarcinoma

EUS: Endoscopic ultrasound; SMT: Submucosal tumor; MP: Muscularis propria; ESD: Endoscopic submucosal dissection; F: Female; M: Male; IHC: Immunohistochemistry.

From a practical perspective, small gastric SMT-like lesions with surface erosion, ulceration, or bleeding should not be regarded as typical low-risk GISTs without further caution[10]. Previous reports have described central ulceration or depression as recurrent features of SMT-like gastric cancer[11,12], and erythematous surface change has also been noted as a characteristic endoscopic finding[13,14]. The 2023 Chinese consensus also emphasizes careful assessment of mucosal surface abnormalities on white-light endoscopy, because erosion or ulceration in an SMT-like lesion may increase concern for malignancy[15]. In addition, indistinct lesion margins on EUS and CT are atypical for benign or low-risk GISTs, suggesting that reliance on imaging findings alone may be insufficient in atypical lesions[16-18]. Therefore, mucosal surface abnormalities in an SMT-like lesion should be considered clinically important warning signs, even when the lesion is small and EUS suggests origin from the muscularis propria. Biopsy should not be delayed in such lesions, and early immunohistochemical evaluation should be considered when biopsy material is limited or clinicopathologic discordance is present[19].

A key issue highlighted by this case is that immunohistochemistry was not performed on the initial diagnostic ESD biopsy despite the spindle cell-like morphology. In our setting, immunohistochemical testing in a primary hospital required a relatively long turnaround time, and the preoperative EUS findings together with the initial biopsy morphology strongly suggested GIST, which contributed to the lack of early immunohistochemical evaluation and may have delayed the final diagnosis.

The choice of tissue-acquisition method also deserves comment. In retrospect, EUS-guided fine-needle biopsy could have been considered as an alternative preoperative diagnostic approach. However, according to both the 2023 Chinese consensus and the European Society of Gastrointestinal Endoscopy technical review[15,20], the diagnostic performance of EUS-fine needle aspiration/fine needle biopsy is influenced by lesion size, and in lesions smaller than 2 cm, obtaining adequate tissue for accurate pathologic diagnosis may be more difficult. In addition, the Chinese consensus indicates that, for lesions considered suitable for complete endoscopic resection, preoperative tissue sampling is not always mandatory in experienced centers[15]. Thus, although EUS-fine needle biopsy might have improved preoperative diagnostic confidence in this case, it cannot be assumed that it would have provided a more reliable diagnosis than diagnostic ESD biopsy.

The indication for diagnostic ESD in this case also warrants discussion in light of the 2023 Chinese consensus[15], which emphasizes that treatment decisions should integrate lesion size, location, symptoms, metastatic risk, and the feasibility of complete resection rather than rely on tumor size alone. Although tumor location could be assessed preoperatively, mitotic activity and molecular characteristics could not be reliably evaluated before adequate tissue acquisition. Therefore, preoperative decision-making in this case relied mainly on lesion size, gastric location, EUS impression, imaging evidence of nodal or distant disease, and the presumed feasibility of complete endoscopic resection. In this case, these assessable preoperative factors included the small lesion size (approximately 1.5 cm), gastric location, a muscularis propria-derived appearance on EUS suspicious for GIST, no definite nodal or distant metastasis on CT, and the presumed feasibility of complete endoscopic resection. Importantly, even if the lesion had ultimately proved to be GIST, definitive resection would still likely have been required in this case because the lesion was considered to have malignant potential and complete removal remained the preferred management strategy[21]. However, the presence of surface erosion and bleeding made this lesion a borderline SMT-like lesion rather than a typical low-risk GIST and should have prompted greater diagnostic caution.

The final diagnosis of poorly differentiated adenocarcinoma was established after examination of the surgical specimen, which provided full-thickness tissue architecture, demonstrated ulcerative growth with invasion into the deep muscularis propria, and revealed lymph node metastasis. In addition, the tumor and metastatic lymph nodes showed a consistent epithelial immunophenotype, including positivity for anion-exchanger 1/anion-exchanger 3, CK antibody monoclonal 5.2, CK7, and CK8/18, with partial CEA expression in the primary tumor. Although classical GIST markers such as CD117 and discovered on GIST-1 were not performed, and CD34 positivity was limited to vascular endothelial cells rather than tumor cells, the combined morphologic and immunophenotypic findings favored poorly differentiated adenocarcinoma over GIST[22,23]. A limitation of this case is that classical GIST markers were not assessed in either the initial ESD biopsy or the surgical specimen; therefore, exclusion of GIST was based on the overall clinicopathological context rather than on a complete GIST-specific immunohistochemical panel.

CONCLUSION

This case highlights the limitations of relying on imaging findings and limited biopsy morphology alone in the diagnosis of gastric SMTs. In suspected GIST, particularly when surface erosion or ulceration is present, multimodal diagnostic evaluation integrating endoscopy, EUS, adequate tissue acquisition, and timely immunohistochemical assessment is essential for establishing the correct diagnosis and excluding epithelial malignancy.