Epithelioid sarcoma (ES) is rare with a poor prognosis and for which a loss of INI1 expression has been recently reported. We report a study of 106 cases with clinical, histologic, and immunohistochemical data, including INI1 expression, and follow-up data. Of the 106 cases, 70 were the conventional subtype and 36 the large cell subtype. INI1 was negative in 86 cases (81.1%): 57 (81%) of 70 conventional and 29 (81%) of 36 large cell subtypes. Treatment modalities were available for 76 and follow-up for 80 patients. Of the 80 patients, 43 (54%) experienced metastasis and 25 (31%) died of the disease. Univariate analysis indicated that tumor size and mitotic index were significant for metastasis-free survival, whereas proximal location, tumor size, tumor multifocality, and mitotic index were significant for overall survival. Loss of expression of INI1 is frequent in the conventional and large cell subtypes of ES and can be used as a diagnostic marker, but it has no prognostic impact.
Epithelioid sarcoma (ES) is a rare soft tissue sarcoma first described by Enzinger1 in 1970. It typically occurs in the distal extremities of young adults as a slow-growing nodule within the dermis or subcutis or in the deep soft tissues. This conventional or distal form is composed of spindle-to-polygonal epithelioid cells arranged in nodules that commonly exhibit central necrosis.2 In 1997, a proximal form arising in the deep part of the pelvis, perineum, and genital tract and proximal extremities was described.3 It consists of large epithelioid carcinoma–like and/or rhabdoid cells and has a more aggressive clinical course than distally located tumors. Moreover, it has a possible link with malignant rhabdoid tumor.
Biallelic inactivation of the SMARCBI/INI1 tumor suppressor gene located on chromosome 22 has been identified as the specific genetic event of malignant rhabdoid tumors, whatever their location.4,5 Recently, deletion of the SMARCBI/INI1 gene with gene inactivation was reported in the proximal type of ES.6 Immunohistochemical expression of INI1 was recently studied in 260 epithelioid malignant tumors, including 96 cases of ES, and showed a loss of expression in about 90% of distal and proximal ES cases.7
We report 106 cases of ES with clinical, histologic, and immunohistochemical data, including INI1 expression, and follow-up data.
Materials and Methods
Cases coded as ES were retrieved from the pathology departments at Institut Bergonié, Bordeaux, France; Institut Gustave Roussy, Villejuif, France; Centre Léon Bérard, Lyon, France; Centre Georges François Leclerc, Dijon, France; Centre Oscar Lambret, Lille, France; and the University Institute of Pathology, Lausanne, Switzerland.
All cases were histologically reviewed and a diagnosis of ES was based on histologic and immunohistochemical studies according to the latest World Health Organization classification of soft tissue and bone tumors.8 For the study, 113 samples from 106 patients were retained: 83 samples from primary tumor, 22 from local recurrence, and 8 from metastasis; for 99 patients, there was 1 sample, and for 7 patients, there were 2 samples (5 with the primary tumor and local recurrence and 2 with the primary tumor and metastasis).
Clinical data, treatment, and follow-up information were obtained from the referring pathologists (see “Acknowledgments”). Treatment and follow-up data were obtained for 76 patients.
We prepared 4-μm-thick sections from formalin-fixed, paraffin-embedded tissue blocks and stained them with H&E and safranin. Immunostaining was performed in all cases using the following antibodies: vimentin (monoclonal V9, dilution 1:100; Dakopatts, Glostrup, Denmark), pankeratin AE1/AE3 (monoclonal AE1/AE3, dilution 1:50; Dakopatts), epithelial membrane antigen (monoclonal E29, dilution 1:100; Dakopatts), CD34 (monoclonal Qbend-10, dilution 1:30; Novocastra Laboratories, Newcastle upon Tyne, England), S-100 protein (polyclonal, dilution 1:500; Dakopatts), and INI1 (monoclonal Baf47, dilution 1:50; BD Becton Dickinson, Le Pont de Claix, France). Staining procedures were performed using the LSAB kit (Dakopatts) with an automated immunostainer. Heat-induced epitope retrieval was used for all antibodies except S-100 protein. All steps were performed at room temperature, and diaminobenzidine was used as the chromogen. Appropriate positive and negative control samples were used throughout.
The clinical end points examined were the time to metastasis (metastasis-free survival) and the time to death of disease (disease-free survival). Clinicopathologic characteristics that were examined for their potential prognostic value included age, sex, distal vs proximal location, tumor size, tumor depth, tumor multifocality, tumor necrosis, conventional vs large cell subtype, mitotic index, presence vs absence of rhabdoid cells, and CD34 and INI1 status.
The clinical features are shown in Table 1. There were 66 males and 40 females. Patient age ranged from 11 to 93 years (median, 31 years). Of the 106 patients, 76 (71.7%) were younger than 40 years. There were 76 cases that occurred in the extremities (59 in the upper extremities with 33 in the hand, 20 in the forearm, and 6 in the arm; and 17 in the lower extremities with 6 in the thigh, 5 in the leg, 4 in the foot, and 2 in the knee area) and 30 cases were proximal (8 in the proximal lower extremity, 7 in the perineal and genital areas, 7 in the head and neck, 6 in the trunk wall, 1 in the proximal upper extremity, and 1 in the mediastinum). Tumor size (known for 69 patients) ranged from 1 to 30 cm (median, 4 cm). Of the cases, 34 (32.1%) were superficially located in the dermis or subcutaneous tissue, 66 (62.3%) were below the superficial fascia, and data were unknown in 6 cases. In 80 cases (75.5%), there was a single nodule; in 17 cases (16.0%), there were multiple nodules; and data were unknown in 9 cases. Duration of the lesion before diagnosis was from 2 to 96 months with a median of 12 months, but data were known for only 19 cases. Nine patients had metastasis at diagnosis: lung, 4 patients; lymph node, 2 patients; lung and lymph node, 1 patient; and other sites (pleura and bone), 2 patients.
Tumors were composed of conventional epithelioid-to-spindle cells Image 1A and Image 1B in 70 cases (66.0%) and of large cells with prominent nucleoli in 36 cases (34.0%). One case showed a hybrid appearance of the 2 subtypes but was classified as a large cell subtype. Rhabdoid cells were present in 12 cases of the conventional subtype (17%) and in 15 cases of the large cell subtype (42%) Image 2A. Mitotic activity ranged from 1 to 50 mitoses per 10 high-power fields (median, 9 mitoses). Areas of tumor necrosis were present in 77 cases, including 49 cases of the conventional subtype (70%) and 28 cases of the large cell subtype (78%).
Immunohistochemical studies Table 2 showed that vimentin reactivity was present in all cases. Pankeratin AE1/AE3 Image 1C and epithelial membrane antigen were positive in 102 and 104 cases, respectively. CD34 was positive in 66 cases: 42 cases of the conventional subtype (60%) and 24 cases of the large cell subtype (67%). S-100 protein was focally positive in 16 cases (15.1%). Loss of expression of INI1 was shown in 86 cases (81%) with a positivity in the nuclei of nontumor cells in every case. Negativity for INI1 was shown in 57 cases (81%) of the conventional subtype Image 1D and 29 cases (81%) of the large cell subtype Image 2B. Twenty-three (85%) of 27 cases with rhabdoid cells and 63 (80%) of 79 cases without rhabdoid cells were INI1-negative. INI1 was negative in 68 (82%) of 83 primary tumors, 18 (82%) of 22 local recurrences, and 7 (87%) of 8 metastases. In 7 patients, 2 events were studied with the primary tumor for every case, a local recurrence for 5 and a metastasis for 2. In 6 cases, both events were negative for INI1, whereas in 1 case, both events (the primary tumor and a local recurrence) were positive.
* Data are given as number (percentage).
Treatment and Follow-up
Data on treatment modalities were available for 76 patients, and follow-up data were available for 80 patients. Of the tumors, 70 were initially excised, with a complete resection in 48 cases; 37 patients received radiotherapy; and 35 patients received chemotherapy. Initial incomplete resection was related to the location in 5 cases (3 in the perineum and 2 in the neck) and to tumor size in 4 cases (>80 mm). Moreover, in the group of tumors initially resected with incomplete resection, 15 (68%) of 22 were deeply located, whereas only 13 (27%) of 48 tumors initially resected with a complete resection were deeply located. Follow-up ranged from 2 to 336 months for survivors (median, 48 months). Among patients with an initial complete resection, 9 (19%) experienced at least 1 local recurrence. At least 1 distant metastasis was experienced by 43 patients (54%): lymph node for 24 patients, lung for 21 patients (7 cases both lymph node and lung metastases), and other sites for 5 patients. At last follow-up, 37 patients had no evidence of disease (6–225 months after diagnosis), 7 were alive with disease, and 25 patients died of the disease. Univariate analysis (Table 1) indicated that tumor size and mitotic index were significant for metastasis-free survival, whereas distal vs proximal location, tumor size, multifocality, and mitotic index were significant for overall survival.
In this series of ES, 81% of cases showed a loss of INI1 expression in the conventional and large cell subtypes. These results are in agreement with the data reported by Hornick et al7 on 96 cases showing a loss of INI1 expression in 92% of cases. In our series, there was no link between the presence of rhabdoid cells and the loss of INI1 expression. The frequency of loss of INI1 expression was similar in primary tumors, local recurrences, and metastases, and in patients with 2 events, results were similar in the 2 events, suggesting that this abnormality occurs early and is involved in tumor genesis rather than progression.
Immunostaining for INI1 can be used to confirm the diagnosis of ES as a complement to epithelial markers and CD34.9 In fact, absence of INI1 expression is rarely observed in other tumor types. It has mainly been reported in rhabdoid tumors10,11 and in some epithelioid malignant schwannomas.7 A partial loss of INI1 expression has also been documented in some synovial sarcomas.11 In the series reported by Hornick et al,7 INI1 expression was normal in metastatic carcinomas from different primary sites, embryonal carcinomas, metastatic melanomas, epithelioid mesotheliomas, epithelioid angiosarcomas, anaplastic large cell lymphomas, and histiocytic sarcomas. The commercially available antibody is of good quality and easy to interpret, thanks to the constant presence of an internal positive control in samples (normal vessels, inflammatory cells).
In rhabdoid tumors, immunohistochemical analysis is very sensitive and highly specific for the detection of INI1 loss of function.10,11 This loss of expression makes it possible to identify cases showing histologic features poorly compatible with rhabdoid tumor but exhibiting an age of onset and clinical behavior similar to that of conventional rhabdoid tumors. In rhabdoid tumors, this loss of expression is related to a biallelic inactivation of the hSNF5/INI1 tumor suppressor gene related to point mutations or deletions.12 A previous study reported molecular analysis of hSNF5/INI1 gene and protein expression in 11 cases of epithelioid sarcomas, with 6 proximal and 5 distal types showing an absence of protein expression in 6 cases (5 of 6 proximal types) and deletion of the gene in 5 proximal types.6 These results suggested that, as in rhabdoid tumors, the loss of INI1 expression is related to a loss of gene function by deletion. These results should certainly be confirmed in a larger series of cases.
In our series, there was no link between loss of INI1 expression and prognosis. Metastasis outcome was related to tumor size and mitotic index and overall survival to proximal vs distal location, tumor size, tumor multifocality, and mitotic index. These factors have been previously identified in several studies.2,13–17 In this series, tumor grading had no prognostic value for predicting overall survival or distant metastasis outcome.
We are grateful to the following for contributing cases and paraffin blocks or for supplying clinical information: J.J. Auger, MD, Cholet, France; D. Bakeland, MD, Guéret, France; A. Balaton, MD, Paris, France; S. Bendjaballah, MD, Beauvais, France; V. Bland, MD, Nancy, France; A. Brouchet, MD, Toulouse, France; A. Chollat-Namy, MD, Marseille, France; P. Coppe, MD, Dunkerque, France; A. Daubech, MD, Le Bouscat, France; J.C. Diana, MD, Aubagne, France; D. Douvin, MD, Orléans, France; M.C. Golaire, MD, Saint-Brieuc, France; B. Gosselin, MD, Lille; N. Henroteaux, MD, Narbonne, France; S. Labbe, MD, La Rochelle, France; B. Laillou, MD, Périgueux, France; C. Lefebvre, MD, Lille, France; C. Mannat-Guillon, MD, Le Bouscat; M. Marquet, MD, Montluçon, France; M. Massot-Bordenave, MD, Bayonne, France; J.F. Michiels, MD, Nice, France; J.P. Mompart, MD, La Rochelle; P. Noret, MD, Rennes, France; S. Patrier, MD, Elbeuf, France; J. Pelletier, MD, Tours, France; P. Richard, MD, Amiens, France; E. Salles, MD, Tarbes, France; G. Valantin, MD, Nancy; C. Valence, MD, Ajaccio, France; A. Villejoubert, MD, Bourges, France; and N. Visy, MD, Reims, France.
Upon completion of this activity you will be able to:
describe the link between SMARCBI/INI1suppressor gene and human tumors.
define the proximal form of epithelioid sarcoma.
list the most important prognostic factors of epithelioid sarcoma.
The ASCP is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians. The ASCP designates this educational activity for a maximum of 1 AMA PRA Category 1 Credit™ per article. This activity qualifies as an American Board of Pathology Maintenance of Certification Part II Self-Assessment Module.
The authors of this article and the planning committee members and staff have no relevant financial relationships with commercial interests to disclose.
Questions appear on p 285. Exam is located at www.ascp.org/ajcpcme.