Abstract
Background/Aim: Peritoneal fluid (PF) cytology is critical for distinguishing high-grade serous carcinoma (HGSC) from metastatic disease in patients with breast carcinoma who present with peritoneal carcinomatosis (PC). Case Report: A 50-year-old woman underwent surgery and adjuvant therapy for pT1N0 grade 2/2 luminal A breast carcinoma. Sixteen months postoperatively, palliative chemotherapy was administered following a pleural biopsy and diagnosis of metastatic carcinoma. The patient developed PC despite chemotherapy. PF cytology specimens suggested metastatic carcinoma. However, we observed a papillary cellular arrangement during the review of cytology slides. HGSC was confirmed by immunocytochemistry showing positive paired box 8 (PAX8) and Wilms' tumor 1 (WT1) expression and negative GATA-binding protein 3 expression. Conclusion: In patients with breast carcinoma history, an awareness of characteristic cytomorphology of HGSC, including a papillary pattern with positive PAX8 and WT1 immunoreactivity, is essential to prevent the misdiagnosis of such cases and in ensuring accurate treatment and management.
Peritoneal fluid cytology is a widely used tool for the detection and diagnosis of disseminated malignant cells. However, peritoneal fluid specimens from patients who were treated for breast carcinoma should be examined cautiously to ensure a clear distinction between primary peritoneal high-grade serous carcinoma (HGSC) and peritoneal metastatic breast carcinoma (pMBC) (1, 2). In patients with peritoneal carcinomatosis, this diagnostic distinction between HGSC and pMBC is needed to ensure an appropriate treatment strategy (1).
According to previous studies, patients with primary peritoneal HGSC have a significantly longer survival duration than patients with pMBC after treatment with cytoreductive surgery (1, 3). In other words, the origin of peritoneal malignancy is a pivotal factor in the survival prognosis of the patient. Consequently, peritoneal fluid cytology is among the most critical types of assessment performed upon the detection of peritoneal carcinomatosis in a patient with breast carcinoma (1). We herein present a case of a 50-year-old woman with breast carcinoma who was finally diagnosed with peritoneal HGSC after several misdiagnoses of pMBC based on cytological and surgical specimens. An awareness of this repeated misinterpretation prompted us to initiate a clinicopathological review. In this report, we describe the cytopathological and immunophenotypic characteristics that are useful for the distinction of primary peritoneal HGSC from pMBC.
Case Report
Case presentation. A 50-year-old Korean woman with a family history of breast carcinoma presented with abnormal findings in her left breast that were detected during regular screening. A subsequent biopsy led to a diagnosis of invasive ductal carcinoma of the breast. Magnetic resonance imaging (MRI) revealed an irregular spiculated mass with a diameter of 1.6 cm in the upper outer quadrant of her right breast, with adjacent probable satellite nodules. A 1.1-cm-sized non-mass enhancement was also detected in the upper outer quadrant of left breast. The patient underwent bilateral breast-conserving surgery with sentinel lymph node biopsy.
The postoperative pathological examination led to the diagnosis of multiple invasive ductal carcinomas (two right-sided; one left-sided). All lesions were defined as a modified Scarff–Bloom–Richardson grade of 2. The greatest tumor dimension was 1.5 cm. Immunohistochemistry revealed that the tumors were consistent with the luminal A subtype, as indicated by uniform estrogen receptor (ER; intensity score=3; proportion score=5; total score=8) and progesterone receptor (PR) positivity (intensity score=3; proportion score=5; total score=8) and human epidermal growth factor receptor 2 (HER2; score=0) negativity. An extensive intraductal component was present. Although lymphovascular invasion was identified, no lymph node metastasis was observed. Accordingly, the pathological stage of the breast carcinoma was pT1(3)N0. Postoperatively, the patient received chemotherapy with tamoxifen and goserelin and subsequent adjuvant radiotherapy.
Sixteen months after treatment, this patient presented at the Emergency Department with resting dyspnea. Chest computed tomography (CT) revealed right pleural thickening accompanied by a large volume of effusion in the right hemithorax, suggesting a malignancy. A cytological examination of the right-sided pleural fluid revealed numerous malignant cells, leading to a diagnosis of metastatic carcinoma. A pleural biopsy was performed via video-assisted thoracoscopic surgery (VATS), and an analysis of the specimen supported the apparent diagnosis of metastatic carcinoma. Immunostaining revealed that the pleural tumor was positive for ER but negative for PR and HER2. Accordingly, the patient underwent palliative chemotherapy with paclitaxel for pMBC.
After six cycles of palliative chemotherapy, she presented at the Emergency Department with abdominal distension. Abdominopelvic CT revealed diffuse peritoneal carcinomatosis and a large volume of ascites. The initial cytological analysis of the peritoneal fluid, as well as six subsequent analyses, yielded a diagnosis of metastatic carcinoma. At that time, polymerase chain reaction and sequencing analyses of the breast cancer 1 (BRCA1) and BRCA2 genes revealed no pathogenic mutations.
Twenty-seven months after the initial diagnosis of primary breast carcinoma, a seventh sample of peritoneal fluid was submitted for cytological examination. At that time, a gynecological pathologist examined the liquid-based cytology slides and raised a suspicion of primary peritoneal HGSC. Immunocytochemical staining to detect paired box 8 (PAX8), Wilms tumor 1 (WT1), p53, and GATA-binding protein 3 (GATA3) revealed diffuse and strongly positive staining for the former two proteins but negative staining for the latter two. These results were consistent with tubo-ovarian or peritoneal serous carcinoma. Subsequent immunostaining revealed a complete loss of p53 expression, indicative of HGSC. Finally, a diagnosis of primary peritoneal HGSC was made based on the absence of tubo-ovarian lesions on radiologic images. At that time, we repeated the BRCA mutational analysis using a multiplex ligation-dependent probe amplification method and detected a heterozygous deletion of exons 1-14 of BRCA1. Based on these results, the chemotherapy regimen was changed to paclitaxel and carboplatin for the treatment of primary peritoneal HGSC.
The patient's serum cancer antigen-125 concentration decreased rapidly following the initiation of chemotherapy (from 302.9 to 44.1 U/ml), indicating a reduced tumor burden. Two months after the change in the chemotherapy regimen, chest CT revealed no change in the extent of pleural thickening despite the disappearance of the left pleural effusion. Abdominopelvic CT revealed that although the retroperitoneal lymph node metastasis had decreased markedly in size, the mild peritoneal thickening and ascites remained unchanged. Persistent left hydronephrosis indicative of ureteral metastasis was also observed. Taken together, these findings suggested that patient had achieved a partial response.
Cytopathological findings. All liquid-based pleural and peritoneal fluid cytology specimens exhibited high cellularity. A low-power microscopic analysis revealed that the tumor cells were arranged in a predominantly papillary pattern (Figure 1A) and had formed variably sized papillary fragments and clusters with or without fibrovascular cores (Figure 1B). In some areas, individual tumor cells containing large, round-to-oval, hyperchromatic nuclei with pleomorphism and conspicuous nucleoli were admixed with mesothelial cells and inflammatory cells (Figure 1C). Mitoses were frequently observed (Figure 1D), and atypical mitotic figures were present.
The peritoneal fluid cytology specimens collected after chemotherapy contained multiple highly abnormal anaplastic nuclei that were more than five times larger than the adjacent tumor cell nuclei (Figure 1E). Most nuclei were surrounded by irregular nuclear membranes, and some displayed chromatin patterns with a smudged and degenerated appearance. Occasionally, the abnormal nuclear shape even mimicked the appearance of a multinucleated giant cell. Cells harboring these abnormal nuclei also possessed intracytoplasmic microvacuoles (Figure 1F), the aggregates of which resulted in a foamy appearance. Some tumor cells also contained large intracytoplasmic vacuoles that resembled perinuclear halos. Cyst-like spaces were observed between the individual tumor cells (Figure 1G).
Immunostaining of cell blocks and liquid-based preparation slides was performed as previously described (4-15). Immunocytochemically, the tumor cells exhibited diffuse and strong nuclear immunoreactivity for PAX8 (Figure 1H) and WT1 (Figure 1I), indicating a Müllerian origin and serous type, respectively. However, malignant cell nuclei completely lacked p53 expression (Figure 1J), indicating a truncating TP53 mutation. The lack of immunoreactivity for GATA3 (Figure 1K) excluded the possibility of a metastasis originating from the breast carcinoma.
Histopathological findings. The VATS pleural biopsy specimen revealed diffusely growing tumor cells and a lack of papillary architecture. Irregularly shaped and sized clusters of tumor cells were delineated by infiltrating borders containing obviously desmoplastic stroma (Figure 2A). The individual tumor cells exhibited both cytoplasmic and nuclear alterations and possessed relatively distinct borders with abundant eosinophilic cytoplasm. In several foci, the cytoplasm was either intensely eosinophilic or clear, with variably-sized vacuoles. The tumor cell nuclei displayed marked pleomorphism and membrane irregularity. Consistent with the cytopathological features, some tumor cell nuclei were more than 5-fold larger than those in the adjacent tumor cells, and cells harboring the former nuclei thus resembled multinucleated giant cells (Figure 2B). These cells also had a degenerated appearance, with apparent chromatin clumping or smudging. Frequent mitotic figures and a few atypical mitoses were observed. Consistent with the cytology specimens, occasional intracytoplasmic vacuoles were detected. Coagulative tumor cell necrosis was evident in some areas of the biopsied specimen, and scattered nests or clusters of surviving tumor cells were observed adjacent to the necrotic areas (Figure 2C). Lymphatic tumor emboli were also identified. The stroma was densely fibrotic in some areas and contained inflammatory cells and fibrinous debris. Immunostaining of the pleural tumor yielded results consistent with the immunocytochemical findings. Specifically, the tumor cells were uniformly positive for PAX8 (Figure 2E) and WT1 (Figure 2F), and exhibited a complete lack of p53 expression (Figure 2G). The surrounding inflammatory cells and reactive mesothelial cells exhibited patchy p53 expression with a variable level of staining intensity (i.e. wild-type p53 expression pattern). Again, the tumor cells were negative for GATA3 (Figure 2H), as well as cytokeratin 5/6 and D2-40, which excluded the possibility of MBC and malignant mesothelioma.
Discussion
In this report, we have described our experience with a case of primary HGSC that was repeatedly misdiagnosed as pMBC in a patient who had previously been treated for primary breast carcinoma. Female breast carcinoma accounted for approximately one-fourth of all newly diagnosed carcinoma cases worldwide in 2018 (16). In the Republic of Korea, the incidence of breast carcinoma has increased continuously since 2000, and this malignancy is among the top five carcinomas affecting women of all ages (17). The bone is the most common site of first distant metastasis of breast carcinoma, followed by the lungs, brain, and liver (18). In contrast, MBC rarely involves the peritoneal cavity. Although only a few articles have reported the incidence of pMBC, this disease is undoubtedly a significant cause of morbidity and mortality in patients with breast carcinoma (19, 20).
Lobular histology, higher histological grade, and non-luminal A subtype have been identified as significant pathological predictors of the development of MBC (21). Moreover, MBC with peritoneal involvement occurs more frequently in patients with advanced-stage primary carcinoma or a history of multiple nonperitoneal metastases (3). In our case, the patient had early-stage (pT1N0) primary breast carcinoma with a histological grade of 2, and the tumor exhibited diffusely strong ER and PR expression (i.e. luminal A subtype). Moreover, she had no prior history of recurrence or nonperitoneal metastasis. Taken together, these findings suggested a low likelihood of pMBC.
In a patient with history of carcinoma, the primary site of a malignant effusion is often the site of a pre-existing malignancy. Consequently, we did not consider the possibility of a primary peritoneal HGSC during several rounds of cytological analysis. However, the features of our case demonstrate the importance of a pathological focus on the cytomorphological features, as the possibility of a second primary site of malignancy cannot be excluded completely. In such cases, a pattern recognition approach can facilitate the differential diagnosis. For example, on a liquid-based cytological slide, pMBC cells in a malignant effusion would generally present with a cannonball appearance or as acinar groups and single cells, although single linear formations of signet ring-like cells have been observed in cases involving a primary invasive lobular carcinoma of the breast. Specimens of hormone receptor-positive breast carcinoma contain tumor cells with low-to-intermediate-grade nuclear atypia and a low mitotic rate. A papillary structure with fibrovascular cores or psammomatous calcifications is rarely observed in histological primary breast carcinomas other than invasive micropapillary carcinomas. In contrast, cellular specimens of primary peritoneal HGSCs usually comprise predominantly papillary clusters with variable sizes and occasional psammoma bodies, although cannonballs or acinar groups may be observed (22). HGSC tumor cells contain large, pleomorphic nuclei with hyperchromasia and conspicuous nucleoli and exhibit markedly increased mitotic activity and atypical mitotic figures. In our case, all the cytology specimens were hypercellular and predominantly contained arborizing papillary fragments, as well as large tumor cells with severe nuclear pleomorphism, multiple prominent nucleoli, and frequent mitoses, some of which showed atypical figures.
Immunocytochemistry can help to identify the cell type and origin of a metastatic tumor. Previous studies indicate that the expression of GATA3, gross cystic disease fluid protein-15, and mammaglobin in tumor cells favors MBC (23), whereas the expression of PAX8 favors a tubo-ovarian or peritoneal origin and WT1 positivity indicates the serous subtype (24). The predictive value of these markers for distinguishing between the breast or peritoneal/ovarian origin of malignant cells has been well-demonstrated (25, 26). Both the lymph node and distant metastatic sites of MBCs often exhibit ER/PR/HER2 expression patterns that differ from those of the corresponding primary breast carcinoma (27, 28). Furthermore, similar hormone receptor expression patterns were observed in primary breast carcinoma and HGSC tissues of approximately one-third of cases in a previous study (1). In summary, although the ER/PR/HER2 immunoprofile is therapeutically important, it may not facilitate the differential diagnosis between HGSC and MBC (1).
In contrast to the cytological specimens, we did not observe the papillary structure in VATS biopsy specimens collected from our case. We additionally observed nuclear and cytoplasmic alterations that are not commonly observed in HGSC. The nuclear features included extreme enlargement with an irregular outline, a highly abnormal outline that mimicked a multinucleated giant cell, and a degenerated chromatin pattern with occasional clumping or smudging. The cytoplasm was abundant and contained intracytoplasmic vacuolations and microvacuoles, the aggregation of which caused a foamy appearance. Some areas contained intensely eosinophilic cytoplasm. The concomitant increases in the nuclear and cytoplasmic volumes resulted in a low nucleus-to-cytoplasm ratio, characteristic of the histological features of surviving tumor tissues in response to chemotherapy. Our observation of the same cytomorphology in the ascitic fluid after chemotherapy further confirmed that these nuclear and cytoplasmic alterations were caused by chemotherapy. We therefore attributed the misinterpretation of the pleural lesions as pMBC to the absence of the typical architectural pattern of HGSC, as well as the chemotherapy-induced alterations in cellular morphology. Our observations suggest that when examining recurrent or metastatic tumors in a patient who has received postoperative chemotherapy, the pathologist should recognize the morphological changes attributable to treatment, as these may increase the challenges associated with diagnosis and determination of the histological type and grade.
In summary, previous studies have emphasized the importance of differentiating between pMBC and HGSC via peritoneal fluid cytology in a patient with a history of breast carcinoma. We hereby present an additional case in which HGSC was repeatedly misinterpreted, as we believe that our experience further underscores the need for awareness regarding the diagnostic hallmarks of HGSC, such as a dominant papillary pattern with positive immunoreactivity for PAX8 and WT1 and a lack of GATA3. Although peritoneal HGSC is rare, this awareness would reduce the likelihood of misdiagnosis and ensure accurate management, and would ultimately improve patient survival.
Acknowledgements
This research was supported by a grant from the National Research Foundation of Korea (NRF), funded by the Korean government (MSIT) (2018R1C1B5043725). All Authors would like to thank Dr. Sung-Im Do (Department of Pathology, Kangbuk Samsung Hospital, Seoul, Republic of Korea) for her valuable comments and support.
Footnotes
Authors' Contributions
All Authors made substantial contributions to the conception and design of the study; the acquisition, analysis, and interpretation of the data; drafting of the article; critical revision of the article for important intellectual content; and the final approval of the version to be published.
Conflicts of Interest
None of the Authors have any conflicts of interest to declare regarding this study.
- Received March 12, 2020.
- Revision received April 1, 2020.
- Accepted April 8, 2020.
- Copyright© 2020, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved