The application of tumor origin genetic testing in differential diagnosis of breast metaplastic carcinoma

INTRODUCTION

Tumors originating from different tissues possess specific gene expression profiles resembling the tissues of origin, and tumor types can, therefore, be identified by analyzing these profiles. A method for identifying tumor tissue origin based on the expression levels of 90 genes has been developed. Variations in gene expression offer important insights for distinguishing between solid tumors and afford useful approaches for identifying tissue origin in patients with cancer of unknown primary (CUP). A multicenter clinical validation study investigating CUPs demonstrated that this method provided support for targeted treatment in 82% of cases.[1]

Metaplastic breast carcinoma represents a heterogeneous group of highly invasive variant diseases, classified into histological subtypes such as low-grade malignant squamous adenocarcinoma, fibromatosis-like metaplastic carcinoma, spindle cell carcinoma, metaplastic carcinoma with heterogeneous mesenchymal component differentiation, and mixed metaplastic carcinoma.[2] In this report, we present two clinical cases of breast tumors diagnosed through comprehensive genomic tumor analysis.

Testing to determine tumor tissue origin involved using assay kits and corresponding analysis software. Paraffin samples were sectioned at a thickness of 4 µm. Experimental procedures followed the manufacturer’s protocol for the Tissue Origin Gene Detection Kit (Hangzhou Canhelp Gene Technology Co., Ltd., China; Medical Device Registration Certificate No. NMPA20223400901). Following RNA extraction, gene expression profiling was performed using reverse-transcription quantitative polymerase chain reaction with fluorescent probes. The resulting expression profiles were computationally matched against a multicancer reference genomic database using proprietary pattern recognition algorithms. This report was composed in compliance with the Declaration of Helsinki,[3] approved by the Hospital’s Ethics Committee, and conducted after obtaining informed consent from the patients.

CASE REPORTS Patient 1

A 68-year-old postmenopausal woman presented with a mass in the right breast that had persisted for more than 1 week, without breast pain, nipple discharge, skin dimpling, erythema, or axillary lymphadenopathy. Her medical and family histories did not contribute significantly to malignancy risk. Mammography revealed an irregular, high-density mass (3.7 × 3.0 cm) with spiculated margins located in the upper outer quadrant of the right breast. Ultrasonographic findings classified the lesion as breast imaging-reporting and data system (BI-RADS) category 4C (high suspicion for malignancy). Pathological investigation of a needle biopsy indicated a malignant tumor. Microscopically, tumor cells were spindle-shaped, with the stroma exhibiting myxoid degeneration and inflammatory cell infiltration. Some cellular regions showed epithelioid morphology and atypia with infiltrative growth. Immunohistochemical findings, considered collectively, indicated (1) metaplastic carcinoma and (2) a mesenchymal malignant tumor, with definitive pathological classification recommended on complete tumor resection. On February 28, 2023, the patient underwent a radical mastectomy with preservation of the right axilla and post-operative histopathological examination of the specimen. Under low magnification, the tumor demonstrated invasive growth and uneven cellular density. Regions showed fibromyxoid changes, while others revealed high-grade sarcomatous morphology [Figure 1]. Immunohistochemical results were as follows: Cytokeratin (CK)14 (−), P63 (−), cytokeratin high molecular weight (CK-HMW) (−), GATA binding protein 3 (GATA 3) (+), vimentin (+), Ki-67 (+, >90%), estrogen receptor (ER) (−), progesterone receptor (PR) (−), human epidermal growth factor receptor 2 (HER2) (0), CK5/6 (−), P40 (−), S-100 (−), smooth muscle actin (SMA) (−), CK8/18 (−), and leukocyte common antigen (LCA) (−). Figure 2 depicts selected significant immunohistochemical findings. The tumor was originally considered a partially dedifferentiated or undifferentiated mesenchymal tumor, but metaplastic carcinoma could not be excluded due to the presence of atypical components. To definitively establish the tumor’s tissue origin, a 90-gene expression assay was performed using formalin-fixed paraffin-embedded samples. As demonstrated in Figure 3, the analysis indicated with a probability of 60.8% that the tumor originated as a sarcoma. The patient was followed up for 18 months after surgery, with no signs of progression or recurrence observed.

Hematoxylin-eosin staining for Patient 1. (a) Infiltrative growth (green arrow indicate Infiltrative borders). (b) Different regions (▲: Sparse areas; ✱: Dense areas). (c) Myxoid degeneration (green arrow). (d) Sarcomatous morphology in some areas (green arrows indicate nuclear fission). The magnification of the image was ×200. Scale bar: 50μm. Figure 1: Hematoxylin-eosin staining for Patient 1. (a) Infiltrative growth (green arrow indicate Infiltrative borders). (b) Different regions (▲: Sparse areas; ✱: Dense areas). (c) Myxoid degeneration (green arrow). (d) Sarcomatous morphology in some areas (green arrows indicate nuclear fission). The magnification of the image was ×200. Scale bar: 50μm.

Export to PPT

Immunohistochemical findings for Patient 1. (a) Cytokeratin 14 negative. (b) P63 negative. (c) Cytokeratin high molecular weight negative. (d) GATA-3 positive. (e) Vimentin positive. (f) Ki-67 positive in areas of tumor cells. (g) Estrogen receptor negative. (h) Progesterone receptor negative. (i) Human epidermal growth factor receptor 2 negative in areas of tumor cells. The magnification of the image was ×200. Scale bar: 50μm. Figure 2: Immunohistochemical findings for Patient 1. (a) Cytokeratin 14 negative. (b) P63 negative. (c) Cytokeratin high molecular weight negative. (d) GATA-3 positive. (e) Vimentin positive. (f) Ki-67 positive in areas of tumor cells. (g) Estrogen receptor negative. (h) Progesterone receptor negative. (i) Human epidermal growth factor receptor 2 negative in areas of tumor cells. The magnification of the image was ×200. Scale bar: 50μm.

Export to PPT

The 90-gene expression assay revealed that the similarity score of the sarcoma tissue from the right breast mass was 60.8 (Patient 1). Figure 3: The 90-gene expression assay revealed that the similarity score of the sarcoma tissue from the right breast mass was 60.8 (Patient 1).

Export to PPT

Patient 2

A 65-year-old female presented in January 2023 with pain and a palpable mass in the right breast without any obvious precipitating factor. Ultrasonography revealed a solid mass in the right breast graded as BI-RADS category 5. Magnetic resonance imaging indicated multiple abnormal signal shadows at approximately the one o’clock position of the right breast, graded as BI-RADS category 6. On February 14, 2023, the patient underwent a radical mastectomy with the right axillary preservation. Pathological evaluation indicated metaplastic carcinoma with a mucilaginous chondroid matrix, confirmed based on microscopic morphological features and immunohistochemical staining. Immunohistochemical analysis demonstrated CK7 (+), P63 (scattered +), SRY-box transcription factor 9 (SOX-9) (+), ER (<1%), PR (<1%), HER2 (0), calponin (−), S-100 (+), SOX-10 (+), isocitrate dehydrogenase 1 (IDH1) (+), transcriptional repressor GATA binding 1 (TRPS1) (+), and Ki-67 (30%) positivity [Figure 4]. To clearly define tumor origin in Patient 2 and serve as control validation for Patient 1, the 90-gene expression assay was performed using formalin-fixed paraffin-embedded tumor samples. The result indicated a similarity of 68.4% to breast cancer genetic signatures [Figure 5]. The patient was followed for 18 months after surgery with no evidence of progression or recurrence.

Hematoxylin and eosin staining and immunohistochemical findings of Patient 2. (a-c) Areas differing from typical regional metaplastic carcinoma. (d-f) Cytokeratin-7 positive, P63 positive, and SOX-9 positive. (g-i) Estrogen receptor negative, progesterone receptor negative, and human epidermal growth factor receptor 2 negative in tumor cell areas. The magnification of the image was ×200. Scale bar: 50μm. Figure 4: Hematoxylin and eosin staining and immunohistochemical findings of Patient 2. (a-c) Areas differing from typical regional metaplastic carcinoma. (d-f) Cytokeratin-7 positive, P63 positive, and SOX-9 positive. (g-i) Estrogen receptor negative, progesterone receptor negative, and human epidermal growth factor receptor 2 negative in tumor cell areas. The magnification of the image was ×200. Scale bar: 50μm.

Export to PPT

The 90-gene expression assay indicated a similarity of 68.4% between the metaplastic carcinoma and breast cancer (Patient 2). Figure 5: The 90-gene expression assay indicated a similarity of 68.4% between the metaplastic carcinoma and breast cancer (Patient 2).

Export to PPT

DISCUSSION

Clinically, CUPs are histologically confirmed metastatic malignancies whose primary tumor sites cannot be identified by standard diagnostic evaluations or imaging studies. Histologically, CUPs are categorized predominantly as well- or moderately-differentiated adenocarcinomas (50–70%), followed by poorly differentiated carcinoma and adenocarcinomas (20–30%), with the remainder consisting of squamous cell carcinomas (5–8%) or poorly differentiated or undifferentiated malignant tumors (2–3%).[1] Together, CUPs account for 3–5% of newly diagnosed cancers worldwide each year. Including cancers initially described as of unknown origin but later identified, this proportion increases to approximately 12–15% of all newly diagnosed malignant tumors.[4] Recently, a real-time polymerase chain reaction-based method termed “90 gene expression analysis” was developed to classify 21 tumor types based on gene expression signatures. Researchers constructed a pan-cancer transcriptome database containing 5434 samples and employed support vector machine-recursive feature elimination to select 12 predictive genes per tumor type, thus creating a robust 90-gene classifier to support tumor identification. This method predicts tumor origin and supports molecular categorization in metastatic cancers.[5] In the era of precision oncology, novel molecular strategies such as tissue-derived molecular profiling based on gene expression offer improved diagnostic accuracy and optimal therapeutic management for patients with CUP.

Most previous studies and reports have focused on tumors of unknown origin[6-8]; therefore, we present two cases related to the diagnosis of metaplastic breast carcinoma. Metaplastic carcinoma with heterologous mesenchymal differentiation is characterized by biphasic components consisting of mesenchymal elements (chondroid, osteoid, rhabdoid, or glial differentiation) and carcinomatous areas (glandular/tubular structures, solid epithelial nests, or squamous lesions). The mesenchymal component exhibits a range of cytological atypia, from mild dysplasia to high-grade sarcomatous features resembling soft-tissue sarcoma morphology, historically termed as “matrix-producing carcinoma,” when an epithelialto-mesenchymal transition occurs abruptly without spindle cell intermediates. Notably, the identification of a chondro-osseous matrix within the lesion fulfills diagnostic criteria for this entity, even in the presence of occasional transitional spindle cell components.[2] Although epithelial components can usually be identified, extensive sampling may be necessary to distinguish these tumors from primary sarcomas. Immunohistochemistry typically demonstrates high-molecular-weight CKs within epithelial regions; when histomorphological and immunohistochemical ambiguity persists, genetic profiling for tumor origin can aid diagnostic clarification.

In the two breast tumor cases presented in this article, both were primary. The tumor type in Patient 1 could not be conclusively determined by histopathology and immunohistochemistry, although genetic testing suggested a sarcomatous origin. Patient 2 was confirmed postoperatively as having metaplastic breast carcinoma, with genetic profiling indicating breast cancer signatures. Breast sarcoma, arising in mesenchymal tissue, is uncommon (<1% of breast cancers). Given limited clinical research, treatment largely adheres to standard sarcoma protocols, primarily surgical resection supplemented by chemotherapy, radiotherapy, and targeted therapy. Complete surgical excision is critical for achieving local control of breast sarcoma. At present, no randomized controlled trials are available comparing treatment approaches for metaplastic breast carcinoma due to its rarity. Therapeutic strategies for metaplastic breast carcinoma continue to include surgery, chemotherapy, and radiotherapy. As research on metaplastic breast carcinoma advances, novel targeted therapies and immunotherapies – including phosphoinositide 3-kinase (PI3K) inhibitors, poly ADP-ribose polymerase (PARP) inhibitors, and programmed death 1/programmed cell death-ligand 1(PD-1/PD-L1) inhibitors – show promising long-term clinical development and have the potential to significantly contribute to patient treatment.[9,10] Consequently, precise identification of tumor tissue origin is critical for targeted treatment and improving patient prognosis. The 90-gene expression assay not only helps clarify primary tumor origin and facilitates accurate differential diagnosis, but it also holds potential as a valuable molecular diagnostic approach. Due to limited case numbers, the relative prognoses associated with different treatments could not be compared between the two cases in this report, and additional cases should be gathered in future research.

SUMMARY

Tumor-origin genetic testing provides important support for differentiating metaplastic carcinoma from malignant mesenchymal tumors, particularly as an adjunctive diagnostic tool when histological morphology and immunohistochemical markers yield indeterminate results.

AVAILABILITY OF DATA AND MATERIALS

All data generated or analyzed during the generation of this report are included in this article. Further inquiries can be directed to the author.

ABBREVIATIONS

CK14: Cytokeratin 14

CK7: Cytokeratin-7

CUP: Cancer of unknown primary

ER: Estrogen receptor

HER2: Human epidermal growth factor receptor 2

PR: Progesterone receptor

SVM-RFE: Support vector machine recursive feature elimination

AUTHOR CONTRIBUTIONS

BBR: Writing-original draft, writing-review, and editing; BZ: Writing-review and editing and data curation; XNJ: Data curation; HZ: Methodology and resources; and QXX: Conceptualization and resources. All authors read and approved the final manuscript. All authors meet the authorship status of ICMJE.

Comments (0)

No login
gif