Pulmonary sarcomatoid carcinoma in the anterior mediastinum mimicking thymic carcinoma with rapid widespread metastases

Abstract

Rationale:

Pulmonary sarcomatoid carcinoma (PSC) is a rare and aggressive subtype of non–small-cell lung cancer. Marked histologic heterogeneity, nonspecific imaging findings, and limited representation in small biopsy specimens can cause diagnostic error, particularly when a pulmonary lesion abuts the anterior mediastinum and mimics a thymic tumor.

Patient concerns:

A 67-year-old woman presented to a local hospital in Hangzhou with chest pain, low-grade fever, and intermittent hemoptysis. Imaging showed a mass abutting the anterior mediastinum and adjacent lung parenchyma. Core needle biopsy showed predominantly spindle cells with few atypical epithelioid cells in a lymphocyte-rich background and was interpreted as thymic carcinoma. She was subsequently transferred to Zhejiang Hospital for further evaluation.

Diagnoses:

Continued progression during docetaxel–cisplatin chemotherapy and persistent diagnostic uncertainty prompted video-assisted thoracoscopic wedge resection. The resected tumor showed markedly atypical spindle cells, pan-cytokeratin and vimentin positivity, thyroid transcription factor 1 negativity, and an elevated Ki-67 proliferation index. Programmed death-ligand 1 testing showed a tumor proportion score of approximately 10% and a combined positive score of approximately 20%. KRAS p.A146T and NRAS p.Q61K mutations were identified. Integrated histomorphologic, immunohistochemical, and molecular assessment established the diagnosis of PSC.

Interventions:

The patient initially received docetaxel plus cisplatin. Because of continued disease progression, declining performance status, and myelosuppression and gastrointestinal adverse effects during platinum-based chemotherapy, cisplatin was discontinued. The patient subsequently received dose-reduced docetaxel plus pembrolizumab as an individualized treatment strategy intended to balance continued systemic therapy with tolerability.

Outcomes:

The disease progressed rapidly with metastases to the lungs, bone, soft tissue, liver, pancreas, and adrenal gland. The patient died of respiratory failure secondary to widespread tumor dissemination.

Lessons:

PSC should be considered when an apparent anterior mediastinal mass shows spindle cell morphology or discordant clinicoradiologic and pathologic findings. Adequate representative tissue and integrated radiologic, morphologic, immunohistochemical, and molecular assessment are essential. Rapid progression despite docetaxel plus pembrolizumab indicates that programmed death-ligand 1 positivity alone may not reliably predict immunotherapy benefit, while the clinical significance of rare KRAS/NRAS co-mutations remains uncertain.

1. Introduction

Pulmonary sarcomatoid carcinoma (PSC) is a rare and highly aggressive histologic variant of non–small-cell lung carcinoma (NSCLC), accounting for only approximately 0.1% to 0.4% of all primary lung cancers.^[1] Histologically, PSC comprises a heterogeneous group of poorly differentiated NSCLC showing sarcomatoid differentiation.^[2] According to the 2021 World Health Organization classification, the sarcomatoid carcinoma category includes pleomorphic carcinoma, giant cell carcinoma, spindle cell carcinoma, pulmonary blastoma, and carcinosarcoma.^[3] Owing to its pronounced morphologic diversity, absence of specific radiologic features, and propensity to yield atypical or even misleading histologic components on needle biopsy, the clinical diagnosis of PSC is particularly challenging.^[4]

PSC can arise in any region of the lung and more commonly presents as a peripheral rather than a central lesion. However, when the tumor is located near the mediastinum, its imaging findings are more likely to be mistaken for thymic neoplasms, lymphoma, or infectious diseases.^[5,6] An anterior mediastinal mass exhibiting spindle cell proliferation with a lymphocytic background and focal epithelial components is particularly prone to misdiagnosis as thymic carcinoma. Definitive diagnosis relies on a comprehensive assessment of gross specimens, an extended immunohistochemical panel, and molecular testing; in routine practice, however, limited biopsy material often precludes complete immunophenotypic characterization, leading to diagnostic error at initial presentation and delay in appropriate treatment.

This report presents a case of PSC that was initially misdiagnosed as thymic carcinoma on needle biopsy. Through a detailed review of the radiologic features, pathologic findings, immunohistochemical profile, and molecular testing results, we analyze the causes of misdiagnosis and key diagnostic pitfalls. We further discuss the clinical characteristics and therapeutic challenges of PSC, with the aim of enhancing clinicians’ awareness of this rare tumor and improving the accuracy of early diagnosis and management.

2. Case presentation

A 67-year-old woman with no smoking history initially presented to a local hospital in Hangzhou with a 2-month history of chest pain, low-grade fever, and intermittent hemoptysis. She had a remote history of pulmonary tuberculosis that had been treated and cured, and she reported no other comorbidities. Chest computed tomography (CT) demonstrated a mass abutting the anterior mediastinum and adjacent lung parenchyma. A core needle biopsy of the lesion showed abundant spindle cells with a small population of atypical epithelioid cells in a lymphocyte-rich background. Given the apparent anterior mediastinal location, the predominance of spindle cell morphology, the lymphocyte-rich background, the small proportion of atypical epithelioid cells, the limited biopsy volume, and the incomplete immunohistochemical characterization, the outside-hospital pathological interpretation favored thymic carcinoma.

A few weeks later, the patient was transferred to Zhejiang Hospital, Hangzhou, Zhejiang Province, China, for further diagnostic evaluation and management. At our institution, positron emission tomography–computed tomography demonstrated a mass-like lesion in the anterior mediastinum with markedly increased fluorodeoxyglucose uptake (Fig. 1A), which, together with the pathological findings, supported a diagnosis of malignancy. Chemotherapy was initiated with a regimen of docetaxel (120 mg on day 1) combined with cisplatin (30 mg on days 1–3). During the docetaxel–cisplatin chemotherapy, the patient developed treatment-related myelosuppression and gastrointestinal adverse effects, which were managed with supportive care. She was subsequently discharged with planned follow-up.

Figure 1.

(A) PET-CT at initial evaluation showing an anterior mediastinal hypermetabolic lesion. (B) Chest CT after the first cycle of chemotherapy; the anterior superior mediastinal mass measured approximately 66 × 46 mm. (C) Chest CT before surgical biopsy; the anterior superior mediastinal mass measured approximately 70 × 48 mm, indicating progression, with enlargement and increasing solid components of multiple bilateral cavitary lesions and nodules consistent with metastases. CT = computed tomography, PET-CT = positron emission tomography–computed tomography.

Three weeks later, the patient was readmitted to receive the next cycle of chemotherapy. Repeat chest CT (Fig. 1B) demonstrated no significant reduction in lesion size. To further exclude tuberculosis infection, additional tests were conducted, and bronchoscopy revealed a small amount of active bleeding, but no acid-fast bacilli were detected. After completion of the second cycle of chemotherapy, the patient’s condition remained stable, and she was discharged again.

About 1 month later, she was rehospitalized due to worsening symptoms. Repeat chest CT (Fig. 1C) showed interval enlargement of the anterior superior mediastinal lesion despite treatment for presumed thymic carcinoma, indicating disease progression. Multiple bilateral pulmonary cavitary lesions and nodules had also increased in size and number, raising concern for metastases; low-density hepatic lesions were additionally noted. Because imaging suggested poor treatment response and the diagnosis remained uncertain, video-assisted thoracoscopic wedge resection of a left lower lobe lesion was performed to obtain adequate tissue for definitive diagnosis and to guide subsequent management.

Intraoperative frozen section analysis of the left lower lobe nodule and margin demonstrated solid and nested tumor growth, favoring a poorly differentiated carcinoma. On permanent sections, the resected left lower lobe nodule, margin region, and a dorsal basal segment nodule were interpreted as a poorly differentiated malignant tumor; epithelioid sarcoma could not be excluded, and molecular testing was recommended. The stapled pulmonary margin was negative for tumor involvement. Histopathological examination showed an ill-defined tumor composed predominantly of markedly atypical spindle cells without glandular or squamous differentiation (Fig. 2A and B). Immunohistochemical staining demonstrated positivity for pan-cytokeratin (pan-CK) and vimentin, negativity for thyroid transcription factor 1, and an elevated Ki-67 proliferation index (Fig. 2C–F). Pan-CK positivity supported the epithelial nature of the tumor, whereas the coexpression of pan-CK and vimentin, together with the spindle cell morphology, supported sarcomatoid differentiation. Programmed death-ligand 1 (PD-L1) immunohistochemistry using the 22C3 antibody clone demonstrated membranous staining in tumor cells from the present case (Fig. 3A). Evaluation of the entire tumor section yielded a tumor proportion score of approximately 10% and a combined positive score of approximately 20%. An external positive control was included to confirm assay performance and the expected membranous staining pattern (Fig. 3B). Molecular testing identified KRAS p.A146T and NRAS p.Q61K mutations, with variant allele frequencies of 0.45% and 1.36%, respectively. Integrating the histomorphological, immunophenotypic, and molecular findings, the multidisciplinary team established the final diagnosis of PSC.

Figure 2.

Histopathologic and immunohistochemical features. (A) Low-power view shows an ill-defined tumor area stained with hematoxylin and eosin (H&E, ×4). (B) Higher-power view shows markedly atypical spindle cells without glandular or squamous differentiation (H&E, ×20). (C) Tumor cells positive for vimentin. (D) Tumor cells positive for CK (pan). (E) Tumor cells negative for TTF-1. (F) Increased Ki-67 proliferation index. CK = cytokeratin, H&E = hematoxylin and eosin, TTF-1 = thyroid transcription factor 1.

Figure 3.

PD-L1 immunohistochemical staining. (A) Representative PD-L1 membranous staining in tumor cells from the present case. (B) External positive control demonstrating the expected membranous staining pattern for PD-L1 immunohistochemistry; the control tissue was not obtained from the present case. PD-L1 = programmed death-ligand 1.

On postoperative day 10, the patient developed discomfort characterized by soreness and distension in the left thigh, accompanied by recurrent low-grade fever. Imaging studies of the left thigh (Fig. 4A) revealed multiple nodular abnormal signal lesions in the mid-to-distal segment of the left femur, with cortical bone destruction and adjacent muscular edema; similar lesions were also observed in the right femoral head and the soft tissues of both thighs, radiologically suggestive of metastatic disease. A few days later, the patient underwent ultrasound-guided needle biopsy of the left thigh mass, and the pathological findings indicated metastasis of a poorly differentiated malignant tumor.

Figure 4.

(A) Thigh magnetic resonance imaging (MRI) obtained in the early postoperative period demonstrates multifocal lesions compatible with metastases. (B) Chest CT during chemoimmunotherapy shows further enlargement of the anterior superior mediastinal mass (approximately 82 × 51 mm). (C) Abdominal CT: multiple low- and very low-density lesions in the liver and pancreas, suggestive of metastases. CT = computed tomography.

At the time of treatment modification, the patient had rapidly progressive metastatic disease and a declining performance status. Moreover, she had experienced myelosuppression and gastrointestinal adverse effects during the preceding docetaxel–cisplatin chemotherapy. The multidisciplinary team, therefore, considered that further platinum-based doublet chemotherapy would be poorly tolerated. Cisplatin was discontinued, the dose of docetaxel was reduced from 120 mg to 30 mg, and pembrolizumab (100 mg) was added in view of the positive PD-L1 expression. This regimen, consisting of dose-reduced single-agent docetaxel combined with pembrolizumab, represented an individualized empiric salvage strategy and a pragmatic compromise intended to maintain systemic treatment while reducing further treatment-related toxicity. Follow-up chest CT (Fig. 4B) showed further progression of disease, and the second cycle of combined chemotherapy and immunotherapy was initiated shortly thereafter. Due to the new onset of abdominal pain, an additional abdominal CT (Fig. 4C) was performed, revealing exudative changes around the small intestine in the right lower abdomen with suspected free intraperitoneal air, raising concern for gastrointestinal perforation. Multiple low-density lesions were also observed in the liver and pancreas, and a nodular lesion was noted in the left adrenal gland, all of which were considered metastatic foci.

A few days later, the patient experienced sudden onset of chest tightness and markedly worsened dyspnea, progressing to respiratory and circulatory failure. Given the critical condition and the patient’s refusal of invasive resuscitative measures such as chest compressions and endotracheal intubation, she ultimately died of respiratory failure secondary to widespread metastasis of pulmonary malignancy.

3. Discussion

This case highlights the deceptive nature of PSC when it presents in the anterior mediastinum, mimicking thymic carcinoma. It further illustrates the aggressive clinical trajectory of PSC harboring rare RAS co-mutations and the failure of standard chemoimmunotherapy in this context. The following sections provide an in-depth analysis of the diagnostic, molecular, and therapeutic aspects of this case.

The pathogenesis of PSC has not yet been fully elucidated. Current studies suggest that its development is closely linked to epithelial–mesenchymal transition. The epithelial–mesenchymal transition endows epithelial cells with mesenchymal-like properties, enhancing their motility and invasiveness, and serves as a key biological basis for metastasis and treatment resistance in various malignancies.^[7,8] In addition, PSC frequently harbors driver mutations in TP53, EGFR, KRAS, NRAS, and other genes, suggesting malignant clonal evolution from a common progenitor cell. These molecular abnormalities not only aggravate the biological aggressiveness of PSC but also partially account for its resistance to conventional therapies.^[7,9,10]

Early diagnosis of PSC is highly challenging. On the one hand, this type of tumor lacks characteristic radiologic features and often presents as cavitary nodules or scattered consolidative lesions in both lungs, which can easily be mistaken for thymic carcinoma, lymphoma, or infectious diseases such as pulmonary tuberculosis and fungal infections.^[11–13] On the other hand, because of its complex tissue architecture, needle biopsy often fails to capture representative components, resulting in a high rate of misdiagnosis. Even for experienced pathologists, adequate tumor sampling is crucial for the diagnosis of PSC, particularly in the setting of bulky masses.^[5] It has been reported that only about half of PSC cases can be correctly diagnosed pathologically via bronchoscopy or CT-guided lung biopsy, whereas the remaining patients require analysis of large surgical specimens to establish a definitive diagnosis.^[14] In pathological diagnosis, immunohistochemistry plays a pivotal role in the identification of PSC. PSC is characterized by the coexistence of epithelial and mesenchymal components, with a typical immunophenotype of dual positivity for CK/Pan and vimentin, and expression of markers such as thyroid transcription factor 1 and P63 in a subset of cases.^[15] In the present case, the coexpression of pan-CK and vimentin in the resected specimen supported epithelial and sarcomatoid differentiation, respectively. The negative thyroid transcription factor 1 result did not exclude a pulmonary origin. These findings emphasize that immunohistochemical results should be interpreted in conjunction with tumor morphology, radiologic localization, and adequate tissue sampling. Importantly, certain immunomarkers, such as P63 and CK19, may also be expressed in thymic carcinoma, increasing the risk of misinterpretation when only limited biopsy material is available.^[16,17]

In the present case, the initial misdiagnosis was likely multifactorial rather than being attributable solely to insufficient biopsy volume. First, the marked intratumoral heterogeneity of PSC created a substantial risk of sampling error. The core biopsy predominantly contained spindle cells, with only a small population of atypical epithelioid cells, and therefore failed to demonstrate the full spectrum of pleomorphic epithelial and sarcomatoid differentiation. Second, the lesion appeared radiologically to be centered in, or closely associated with, the anterior mediastinum, while its pulmonary origin was not clearly established. This atypical location naturally favored consideration of a thymic epithelial malignancy. Third, the imaging findings were nonspecific and overlapped with those of thymic carcinoma, particularly because the mass broadly abutted both the anterior mediastinum and the adjacent lung parenchyma. Fourth, the initial immunohistochemical characterization was incomplete, limiting the ability to systematically distinguish PSC from thymic carcinoma and other spindle cell malignancies. In addition, certain epithelial markers may be expressed in both PSC and thymic carcinoma, further increasing the risk of diagnostic misinterpretation in a small biopsy specimen. Thus, the initial error resulted from the interaction of sampling bias, tumor heterogeneity, anatomic ambiguity, radiologic overlap, and incomplete immunophenotypic assessment. Examination of the larger surgical specimen permitted more representative morphologic evaluation and a broader immunohistochemical and molecular workup, ultimately supporting the diagnosis of PSC.

The molecular profile of this patient – harboring concurrent KRAS p.A146T and NRAS p.Q61K mutations – provides crucial insight into the tumor’s biological aggression. While KRAS mutations are prevalent in NSCLC (approx. 30%), they predominantly affect codons 12 and 13.^[18,19] The KRAS p.A146T variant identified here is exceptionally rare (<1% of NSCLC) and, unlike classical variants, is more commonly associated with digestive tract malignancies. Furthermore, KRAS and NRAS mutations are typically mutually exclusive; their coexistence in this patient represents a rare “molecular anomaly” likely driven by intense evolutionary pressure. This unique “double hit” on the mitogen-activated protein kinase pathway may effectively amplify downstream signaling, potentially explaining the patient’s catastrophic disease progression and resistance to therapy, a trajectory distinct from classical G12C-driven tumors.^[19–24]

PSC responds poorly to conventional therapies, particularly chemotherapy, and its overall prognosis is markedly worse than that of other NSCLC subtypes.^[6,25,26] Some studies have reported a 5-year survival rate of only 14.4% and a median overall survival of less than 6 months.^[2,27,28] Although PSC is frequently associated with high PD-L1 expression, providing a theoretical basis for treatment with immune checkpoint inhibitors (ICIs), therapeutic efficacy varies substantially among individuals, and robust high-level evidence remains lacking.^[29,30] In the present case, the patient had a PD-L1 tumor proportion score of 10% and combined positive score of 20. Despite treatment with dose-reduced docetaxel plus pembrolizumab, the disease continued to progress rapidly, and no meaningful clinical benefit was observed. This outcome suggests that PD-L1 expression alone may not reliably predict the response to ICIs in PSC. The rationale for retaining docetaxel after progression on the initial docetaxel–cisplatin regimen warrants clarification. During the preceding platinum-based chemotherapy, the patient had experienced myelosuppression and gastrointestinal adverse effects. With subsequent rapid disease progression and declining performance status, further platinum-based doublet chemotherapy was considered likely to be poorly tolerated. Cisplatin was therefore discontinued, docetaxel was dose-reduced, and pembrolizumab was added based on positive PD-L1 expression. This regimen represented a pragmatic compromise intended to maintain systemic antitumor treatment while reducing the risk of further toxicity. However, the decision to retain docetaxel was not based on demonstrated tumor sensitivity, and this individualized approach should not be regarded as a standard treatment strategy for PSC. Current studies indicate that the efficacy of ICIs in non–small-cell lung cancer is highly individualized and closely related to factors such as the tumor immune microenvironment, PD-L1 expression level, and tumor mutational burden.^[31,32] The concurrent KRAS p.A146T and NRAS p.Q61K mutations may reflect aberrant activation of the mitogen-activated protein kinase signaling pathway and could have contributed to the aggressive biological behavior of the tumor; however, this interpretation remains speculative. The predictive and therapeutic significance of these rare co-mutations in PSC is currently unknown, and no validated targeted treatment was available for these specific alterations in the present case.^[33–35]

It is noteworthy that PSC frequently develops widespread metastases at an early stage, involving multiple organs such as the liver, bone, adrenal glands, pancreas, soft tissues, and even the skin.^[28,36] In this case, the patient developed extensive metastases to sites including the femur and pancreas within a short period after diagnosis and ultimately died of tumor-related respiratory failure, underscoring the extremely aggressive nature and high clinical malignancy of PSC.

This case has several limitations. First, the initial diagnostic impression of thymic carcinoma resulted from the combined effects of the lesion’s apparent anterior mediastinal location, overlapping radiologic features, limited core biopsy sampling, and incomplete immunohistochemical characterization. The biopsy predominantly contained spindle cells and did not capture the full histological heterogeneity of the tumor. Second, biomarker assessment was limited to PD-L1 expression and targeted molecular testing, and the tumor immune microenvironment was not evaluated. Third, as this report describes a single patient, no causal relationship can be established between the concurrent KRAS p.A146T and NRAS p.Q61K mutations, the rapid clinical progression, and the lack of response to chemoimmunotherapy. The main implications of this report are therefore the need for adequate and representative tissue sampling and an integrated diagnostic approach combining clinical information, radiologic findings, histomorphology, and immunohistochemistry. PD-L1 expression and rare molecular alterations should be interpreted cautiously, as their predictive and therapeutic significance in PSC remains uncertain.

In summary, this case provides several clinical insights. First, early recognition of PSC should be strengthened. In particular, when mediastinal tumors show complex imaging features and confounding factors such as a prior history of tuberculosis are present, a high index of suspicion is required to avoid misdiagnosis as thymic carcinoma or an infectious lesion. Second, during the diagnostic workup, it is essential to obtain sufficient tissue samples and perform systematic immunohistochemical and molecular testing; biopsy results from needle specimens must be interpreted with great caution. Third, immunophenotypic and molecular findings should be interpreted cautiously when guiding treatment in PSC. In this patient, PD-L1 positivity and concurrent KRAS/NRAS mutations did not predict benefit from docetaxel plus pembrolizumab. The rapid progression despite chemoimmunotherapy underscores the limited predictive value of PD-L1 expression alone and the uncertain therapeutic significance of rare RAS co-mutations. Further studies are required before the benefit of biomarker-guided combination or targeted strategies can be established.

Acknowledgments

The authors thank the patient and the patient’s family for allowing us to use their data for the completion of this report.

Author contributions

Investigation: Linru Zhong.

Supervision: Linru Zhong, Xinhai Zhu.

Visualization: Linru Zhong, Mengyu Chai.

Writing – original draft: Linru Zhong.

Writing – review & editing: Linru Zhong, Mengyu Chai, Hongyu Lin, Wensuo Yang, Xinyu Ma, Xinhai Zhu.

Formal analysis: Hongyu Lin.

Funding acquisition: Xinhai Zhu.