| Japanese Journal of Clinical Oncology | Pages |
Introduction
Case Report
Pathological Findings
Macroscopic Findings
Microscopic Findings
Immunohistochemical Findings
Discussion
References
Dedifferentiated Chordoma: a Case Report
INTRODUCTION
Chordomas are rare neoplasms, occurring as only 3% of all primary bone tumors (1). They are usually restricted to the middle line of the body, and approximately 50% of chordomas arise in the sacrococcygeal region (1). Histologically, chordomas show little cytological atypia. Nevertheless, in 30% of cases, metastases to the lung, skin, lymph node, liver and bone have been reported (1). The prognosis for chordoma patients is poor, the five-year survival rate being only 58% (1). Dedifferentiated chordoma, i.e. chordoma with a sarcomatous component, is so rare that only 31 cases have been reported in the English literature (2-22).
In this paper, we report a rare case of dedifferentiated chordoma in a 50-year-old male. The tumor was initially diagnosed as a leiomyosarcoma, but, after it recurred, study of the resected specimen showed that it was a dedifferentiated chordoma.
CASE REPORT
A 50-year-old male visited a physician for examination of hemorrhoids in April 1994. A tumor was identified by rectal palpation on the dorsal side of the rectum. Abdominal magnetic resonance imaging (MRI) showed a 7 × 5 cm mass with a distinct boundary located in front of the sacrum and compressing the rectum forward. Contiguous with this tumor was a 4-cm-sized one showing the same intensity, but in the sacrococcygeal region (Fig. 1). These tumors presented a low signal intensity on T1-weighted image, and nodular patterns of a high signal on T2-weighted image. Gadolinium-enhanced T1-weighted image demonstrated heterogeneously enhanced tumors. Laboratory studies at that time revealed a normal blood cell count and chemistry profile. A leiomyosarcoma was suspected and, in June 1994, a trans-sacral and transperitoneal operation was carried out. A 5-cm tumor, with a distinct boundary, was located in front of the sacrum. However, the tumor in the sacrococcygeal region was unable to be totally resected because of continuous bleeding. After discharge, the patient suffered from rectovesical failure, and, for this reason, a surgical change of the urinary tract and a proctostomy were carried out ambulatory. There were no signs of recurrence at that time.
Figure 1. Abdominal MRI of the primary tumor shows a well-circumscribed, dumbbell-shaped mass with high signal intensity on Gd-enhanced, T1-weighted image (arrows). In July 1997, abdominal computed tomography (CT) showed a recurrent mass in front of the sacrum. A huge mass with low-density areas was identified between the sacrum and rectum (Fig. 2). A contiguous soft-tissue mass and destructive lesions were also found in the perirectal muscles and sacrum, respectively. In January 1998, total extirpation of the tumor, pelvic exenteration and partial extirpation of the sacrum were carried out at the National Cancer Center Hospital, Tokyo. There have been no signs of recurrence five months after the operation. Figure 2. Abdominal CT of the recurrent tumor reveals a 24 × 14 × 7 cm mass with low-density areas between the sacrum and rectum, and a contiguous soft-tissue mass in the perirectal muscles and multiple osteolytic lesions in the sacrum. Macroscopic findings were not available in the primary tumor. The recurrent tumor measuring 24 × 14 × 7 cm had displaced the muscular wall of the urinary bladder and rectum. However, the tumor had not yet invaded the mucosa of either of them. The cut surface of the tumor showed a multilocular appearance with hemorrhage and necrosis (Fig. 3). The resected sacral lesion measured 6 × 3 cm, was whitish and had a multinodular pattern on its cut surface (Fig. 3).
Figure 3. (A) The sagitally cut surface of the recurrent tumor shows a multilocular appearance with hemorrhage and necrosis displacing the muscular wall of the rectum and bladder (right). Arrows indicate the mucosa of the rectum. (B) Whitish, multinodular lesions on the cut surface of the sacrum. In the primary lesion resected in June 1994, the tumor had proliferated expansively to invade the sacrum, and the tumor lobules were separated by bands of connective tissue and sometimes by the remaining cortical bone (Fig. 4). The tumor was arranged in a trabecular pattern with an abundant myxoid background. The tumor cells were cuboidal to polygonal, with abundant, clear or eosinophilic, vacuolated cytoplasm. Physaliferous cells and signet ring cells were also present. The nuclei were vesicular, with a regular contour, and occasionally had a pleomorphic appearance. Some epithelial cells had become spindle shaped, and were arranged in a fascicular pattern, simulating leiomyosarcoma (Fig. 5). Finely granular cytoplasmic glycogen, that was diastase sensitive, was demonstrable by PAS staining in both the epithelial and spindled components. No sarcomatous component was found in any slide of the primary lesion. Figure 4. The primary tumor in the sacrum shows a pattern of conventional chordoma with strands of vacuolated epithelial cells surrounded by mucinous stroma (H & E; original magnification ×100). Figure 5. Some chordoma cells in the primary tumor have become spindle shaped and are arranged in a fascicular pattern, simulating leiomyosarcoma (H & E; original magnification ×200). In the recurrent lesion resected in January 1998, cystic change with hemorrhage and necrosis were present, and the tumor cells had proliferated along the cystic wall (Fig. 6). The tumor cells were arranged in a sheet-like pattern. Most tumor cells were spindle shaped to polygonal, with abundant eosinophilic cytoplasm, and contained bizarre, hyperchromatic nuclei (Fig. 7). Many mitotic figures including atypical mitoses (over 20/high-power field) were seen. Large multinucleated cells with pleomorphic nuclei were also seen. Lace-like osteoid was intermingled with the tumor cells in small foci (Fig. 8). On the other hand, microscopic features of the sacral lesion were similar to those of the primary lesion. Nuclear pleomorphism was noted in some tumor cells. The spindle cells, those seen in the conventional chordoma component, were not identified in the recurrent sacral tumor. We could not detect a boundary between the conventional chordoma area and sarcomatous area. Figure 6. Most of the recurrent lesions outside the sacrum showed a cyst-forming histological appearance consisting of pleomorphic tumor cells (H & E; original magnification ×100). Figure 7. The recurrent lesion outside the sacrum is histologically characterized by pleomorphic cell sarcoma mimicking malignant fibrous histiocytoma. Note an abnormal mitosis (arrow) (H & E; original magnification ×400). Figure 8. Lace-like osteoid between neoplastic cells in the recurrent lesion (H & E; original magnification ×400). In the primary tumor, most of the typical physaliferous cells and spindled-shaped cells were strongly positive for cytokeratin (Fig. 9) and epithelial membrane antigen (EMA), and diffusely positive for S-100 protein and vimentin. A few tumor cells showed nuclear immunoreactivity for proliferation-associated antigen, Ki-67 (MIB-1). The MIB-1 labeling index was approximately 2-3% (Fig. 10), and the tumor cells were negative for p53 protein. Figure 9. Cytokeratin immunoreactivity in the primary tumor, particularly at the boundary between spindled and epithelioid (left upper) components (LSAB method; original magnification ×200).
Figure 10. (A) Approximately 2-3% of tumor cells in the primary chordoma lesion show nuclear immunostaining for proliferation-associated antigen, Ki-67 (MIB-1) (LSAB method; original magnification ×400). (B) MIB-1 labeling of the recurrent sarcomatous lesion is more than 40% (LSAB method; original magnification ×200). In the sacral region of the recurrent tumor, most tumor cells were positive for cytokeratin, EMA, S-100 protein and vimentin. In contrast, in the sarcomatous region of the recurrent tumor, spindle-shaped and polygonal cells were diffusely positive for vimentin, and focally reactive for S-100 protein, but negative for cytokeratin, EMA, desmin, actins and CD34. The MIB-1 labeling index of the sarcomatous area was greater than 40% (Fig. 10), but that of the sacral area was less than 1%. Both these areas were negative for p53 protein. The incidence of dedifferentiated chordoma has been reported to be 6-9% of all the chordomas (15,23). Of 31 cases of dedifferentiated chordoma reviewed in the English literature, the patients comprised 18 males and 13 females with ages ranging from 9 to 71 (mean 50.9) years. Of these, 20 cases (64.5%) were in the sacrococcygeal region, six (19.3%) in the vertebra and five (16.1%) in the skull base (2-22). Thus, dedifferentiation appears to be more common than in sacrococcygeal chordomas, or in vertebral or spheno-occipital tumors. The period until the discovery of secondary high-grade sarcoma after the initial diagnosis of chordoma has been reported to average 56.6 months (2-22). However, in this case, the period was slightly shorter, at 36 months. The prognosis for reported cases has been poor, with most patients suffering local recurrence or lung metastasis and dying within one or two years after diagnosis. In this case, the primary tumor was found in 1994 as a large soft-tissue mass in the pelvic cavity rather than in the sacrum, and showed a focal histological appearance of spindle-shaped cells arranged in a fascicular pattern. This finding led to misdiagnosis as a leiomyosarcoma at the first medical consultation. The recurrent tumor mostly exhibited microscopic features of pleomorphic cell sarcoma mimicking malignant fibrous histiocytoma (MFH) or osteosarcoma in the pelvic cavity, although it contained typical areas of chordoma in the sacrum. These findings were consistent with those of dedifferentiated chordoma, or chordoma with a sarcomatous component. Of the 31 cases reported previously (2-22), the histological appearance of a sarcomatous component has been described as MFH (10 cases), fibrosarcoma (5), malignant spindle cell component (5), osteosarcoma (3), sarcomatous component (2), anaplastic sarcoma (2), spindle-shaped cell component (1), undifferentiated spindle cell component (1), a spindle cell metaplasia (1) and spindle cell sarcoma (1). From the present immunohistochemical examination, the conventional chordoma areas seen in both the primary tumor and the recurrent tumor were positive for cytokeratin, EMA, S-100 protein and vimentin. On the other hand, the sarcomatous area seen in the recurrent tumor was positive for vimentin, but negative for epithelial markers and smooth muscle markers, such as actins and desmin. These findings were compatible with the characteristic immunohistochemical profiles of dedifferentiated chordoma (12-17,19,20,23). However, it has been reported that some spindle-shaped tumor cells of dedifferentiated chordoma were positive for S-100 protein, as shown in the present case (22). Some conventional chordomas contain a spindle-cell component, and it is important to distinguish these tumors from dedifferentiated chordoma and other spindle cell tumors (12,14,15). The chordoma with a spindled epithelial component is characterized by the transition to the conventional chordoma area, spindled appearance of epithelioid cells, pleomorphism and a paucity of mitotic figures (12,14,15). In the present case, the area of spindle-cell proliferation, as well as the conventional chordoma area, was immunoreactive for cytokeratin, EMA and S-100 protein. Among the cases of chordoma with a sarcomatous component reported in the literature, some have not been adequately examined histologically and immunohistochemically. These cases may be classified as chordoma with a spindle-cell component. To obtain an exact diagnosis, the histological features must be adequately evaluated and the epithelial phenotype examined by appropriate immunohistochemical staining. Theoretically, the pathogenesis of this lesion can be explained by two hypotheses: (1) sarcoma arising as a malignant transformation of the conventional chordoma cells due to progression; (2) sarcomatous change of post-irradiated chordoma (2,4,5,7-13,15-21,23). To our knowledge, 23 cases out of 31 (74.2%) had received radiation therapy. As in our case, chordoma with a sarcomatous component has been reported to occur in the absence of previous radiation therapy; therefore, it is not clear whether irradiation is essential for a sarcoma to develop. Immunoreactivity for Ki-67 (MIB-1) and p53 protein has been identified in dedifferentiated chondrosarcomas and dedifferentiated liposarcomas (24,25). It has been reported that the positivity of these two markers in the sarcomatous area was higher than that of non-sarcomatous tumor tissue. The sarcomatous area in eight cases of dedifferentiated chondrosarcoma had an MIB-1 labeling index of 66%, and p53 protein positivity was seen in 59% (24). The sarcomatous area in three cases of dedifferentiated liposarcoma had an MIB-1 labeling index of 15-20% (25). On the other hand, in the sarcomatous area in dedifferentiated liposarcoma, 13 out of 14 cases (93%) were positive for p53 protein, the proportion of positivity being 5-20% (26). In this case, the MIB-1 labeling index of the sarcomatous area was higher than that of the conventional chordoma area, but both the conventional area and the sarcomatous area were negative for p53 protein. Meis-Kindblom et al. (27) reported that the MIB-1 labeling index of the sarcomatous area in three cases of dedifferentiated chordoma was as high as 20-75%, but only one case out of three was positive for p53 protein. This result, as well as ours, may suggest that low expression of p53 oncoprotein is characteristic of dedifferentiated chordoma, distinct from liposarcoma or chondrosarcoma, but further study in a large number of cases is needed. A correlation between the MIB-1 labeling index and the histological grade of the tumor has been demonstrated in soft-tissue sarcomas (28,29). It has been noted that in the high-labeling-index group, the five-year survival rate was much lower than in the low-labeling-index group, indicating that a high labeling index is related to poor prognosis (28,29). The tumor cells in the sarcomatous area of the dedifferentiated chordoma proliferated so aggressively that we should regard this tumor as a high-grade neoplasm. In summary, we have reported a patient who was initially diagnosed as having a leiomyosarcoma, but was later shown to have a dedifferentiated chordoma after study of the recurrent tumor. This case highlights the importance of careful study of suspected chordoma to allow early identification of sarcomatous components.
PATHOLOGICAL FINDINGS
Macroscopic Findings
A
B
Microscopic Findings
Immunohistochemical Findings
A
B
DISCUSSION
References
This article has been cited by other articles:
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 24 Nov 1998
Copyright©Japanese Journal of Clinical Oncology, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
S. A. Hanna, R. Tirabosco, A. Amin, R. C. Pollock, J. A. Skinner, S. R. Cannon, A. Saifuddin, and T. W. R. Briggs
Dedifferentiated chordoma: A REPORT OF FOUR CASES ARISING 'DE NOVO'
J Bone Joint Surg Br,
May 1, 2008;
90-B(5):
652 - 656.
[Abstract]
[Full Text]
[PDF]
R. K. Heck Jr., T. D. Peabody, and M. A. Simon
Staging of Primary Malignancies of Bone
CA Cancer J Clin,
November 1, 2006;
56(6):
366 - 375.
[Abstract]
[Full Text]
[PDF]
This Article 
Abstract
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (9)
Request Permissions
Google Scholar
Articles by Saito, A
Articles by Moriya, Y
Search for Related Content
PubMed
PubMed Citation
Articles by Saito, A
Articles by Moriya, Y
Social Bookmarking
What's this?