—  SHORT COURSE #21  —

Mesenchymal Neoplasms of the Female Genital Tract

Case 3 - Carcinosarcoma

Teri Longacre, Esther Oliva and Robert Soslow


Introduction
Carcinosarcoma, also referred to as "malignant mixed Mullerian tumor" or "MMMT" is a neoplasm composed of malignant-appearing epithelial and mesenchymal elements [1]. Although they can arise in any genital organ, they are found most frequently in the uterus where they represent less than 5% of malignant neoplasms. These tumors are increasingly thought of as carcinomas that demonstrate sarcomatoid differentiation [2, 3, 4, 5], although many gynecologists persist in classifying them as sarcomas. Regardless, they are extraordinarily heterogeneous, clinically aggressive neoplasms. Their morphologic heterogeneity is responsible for difficulties in accurate diagnosis.

Clinical Features
The mean age of patients with endometrial carcinosarcoma is in the 7th decade, but the age range spans from the 4th through 9th decades. The disease tends to present like other endometrial cancers, with vaginal bleeding being common. Another typical presentation of carcinosarcoma is a polypoid mass that protrudes through the cervical os.

The 5-year survival for carcinosarcoma is around 30% and the 5 year survival in surgical stage I disease (confined to uterus) is approximately 50% [6, 7, 8, 9]. This very aggressive profile contrasts with that of other high grade endometrial cancers where 5 year survivals in stage I disease are approximately 80% or better [10, 11]. This has led to toxic treatment protocols that usually include ifosphamide and cisplatin along with whole pelvic irradiation. There are no data that suggest that treating the predominant tumor component (i.e. rhabdomyosarcoma, when present) is preferable to standard therapy.

There have been numerous studies that have sought to define prognostic factors in carcinosarcoma, but many are compromised by including at least 2 groups of patients, some of whom were clinically staged and others who were surgically staged [6, 7, 8, 9]. Despite that, there is general agreement that surgical stage is the most important prognostic indicator. In common with some of the older literature, our group has recently found that the presence of heterologous elements is a statistically significant poor prognostic factor in Stage I patients [12]. 30% of patients with heterologous elements survive 5 years as compared to 80% of patients with homologous elements. Other factors that have been proposed include the grade of the carcinomatous and sarcomatous elements, the percentage of tumor demonstrating sarcomatous differentiation, the depth of myometrial invasion and the presence of lymphovascular invasion [6, 7, 8, 9].

Morphology
The cardinal rule of carcinosarcoma is that it is a biphasic tumor, composed of distinctive and separate, but admixed, malignant-appearing epithelial and mesenchymal elements. The epithelial and mesenchymal elements should not merge with one another. This definition essentially excludes from consideration all monophasic tumors such as undifferentiated carcinomas and undifferentiated sarcomas and tumors that display either malignant-appearing epithelial elements (endometrioid adenocarcinoma with spindle cell features) or malignant-appearing mesenchymal elements (Mullerian adenosarcoma).

Carcinomatous and sarcomatous elements are generally easy to find, but occasional cases show sarcoma or carcinoma almost exclusively. About one-third of carcinosarcomas harbor a carcinoma with endometrioid differentiation and two-thirds contain carcinomas that are serous or undifferentiated [12]. The former tumors may be associated with complex atypical hyperplasia and endometrial intraepithelial carcinoma may be found in the latter tumors. Clear cell, mucinous and squamous carcinomas have also been described. In our series [12], 10% of the carcinomatous components were FIGO grade 1, 10% were grade 2 and 80% were grade 3. The sarcomatous components were also heterogeneous. The homologous components of carcinosarcoma are usually spindle cell tumors without obvious differentiation; many resemble fibrosarcomas or pleomorphic sarcomas. In our experience, it is uncommon to find a carcinosarcoma containing areas that closely resemble leiomyosarcoma or endometrial stromal sarcoma. Almost all are histologically high grade. The most common heterologous elements are chondroid or rhabdomyoid (constituting something that resembles either pleormorphic rhabdomyosarcoma or embryonal rhabdomyosarcoma). Occasional cases showing lipoblastoid, osteoid, neural/neuroectodermal and angiomatoid elements have been described.

Most carcinosarcoma metastases are epithelial only or, less commonly, epithelial predominant. It is uncommon, but not extraordinary, to find sarcoma-only metastases from carcinosarcoma.

Immunohistochemistry
The immunophenotype of carcinosarcoma should closely follow that of the individual elements present in the tumor. For example, the serous component should express keratins and p53, while the rhabdomyoblastic component should express desmin and myogenin. It is a well known diagnostic pitfall that the sarcomatous component of carcinosarcoma can express keratins, but the same is true of leiomyosarcoma, in particular. Given this and the adherence to the cardinal rule of carcinosarcoma (see above), immunohistochemistry should not be needed for diagnosis and should only be used in exceptional circumstances. I essentially restrict the use of immunohistochemistry to cases where I want to support my impression of a heterologous element, particularly rhabdomyosarcoma. I will rarely perform immunohistochemistry to support my impression of a malignant epithelial component (p53 in epithelium that looks neoplastic and probably malignant) or to support a morphologically distinct epithelial component where the overwhelming impression is that of a monophasic undifferentiated tumor. Pleomorphic, monophasic neoplasms with patchy keratin expression should not necessarily be diagnosed as carcinosarcomas.

Differential Diagnosis
As mentioned earlier, most monophasic tumors should not be considered candidates for a carcinosarcoma diagnosis, although many pathologists believe that uterine pleomorphic rhabdomyosarcomas represent carcinosarcomas in which the epithelial component has not been sampled or has been overgrown by sarcoma. This leaves the biphasic tumors.

The four most important biphasic tumors to consider in the differential diagnosis are adenosarcoma, endometrioid adenocarcinoma with spindle cell elements [13, 14], so-called dedifferentiated endometrial carcinoma [15, 16], and composite or collision tumors.

Mullerian adenosarcoma vs carcinosarcoma
Mullerian adenosarcomas are typically polypoid tumors with the appearance of phyllodes tumor. The epithelial component is benign appearing, whereas the mesenchymal component is malignant, although it is usually low-grade in appearance. Therefore, the typical adenosarcoma does not bear resemblance to carcinosarcoma, although occasional carcinosarcomas can contain areas that are histologically indistinguishable from adenosarcoma [17]. Some adenosarcomas display sarcomatous stromal overgrowth; the sarcomatous component of these tumors is usually high grade and, on occasion, can efface the epithelial components. It is a natural tendency to consider carcinosarcoma (or leiomyosarcoma) in such a case, but, again, carcinosarcoma requires elements, at least focally, that are histologically carcinoma. Very rare adenosarcomas can manifest complex atypical hyperplasia in the epithelial compartment, leading to confusion with carcinosarcoma [18]. The phyllodes-like growth pattern, the absence of overt carcinoma, and epithelium showing low histologic grade are all against a carcinosarcoma diagnosis. Last, be aware that adenosarcoma can contain heterologous elements.

Endometrioid adenocarcinoma with spindle cell elements vs carcinosarcoma
Endometrioid adenocarcinoma with spindle cell elements has long been recognized, although it was only recently given a name and an in-depth description [13, 14]. In this tumor, the endometrioid elements, frequently showing squamous metaplasia, fuse imperceptibly with spindle cell elements that are never histologically high grade. In most cases, the endometrioid component is no more than FIGO grade 2 and the spindle cell component is cellular and sometimes mitotically active, but not markedly atypical. If there is confusion between this entity and carcinosarcoma, use the tumor grade and the presence or absence of "element fusion" to inform the decision. Carcinosarcoma contains easily separable epithelial and mesenchymal elements whereas this type of endometrioid adenocarcinoma shows seamless fusion of the two (i.e. element fusion). Be aware here as well that endometrioid adenocarcinomas can contain chondroid and osteoid elements. Heterologous elements by themselves do not signify carcinosarcoma.

"Dedifferentiated" endometrial carcinoma vs carcinosarcoma
"Dedifferentiated" endometrial carcinoma is a recently described entity that includes a well or moderately differentiated endometrioid adenocarcinoma juxtaposed with an undifferentiated carcinoma [15, 16]. These tumors usually look biphasic and the undifferentiated component is distinct from the differentiated component, theoretically leading to confusion with carcinosarcoma. The undifferentiated component characteristically shows only focal or weak keratin expression, potentially compounding difficulties with the differential diagnosis. In contrast to carcinosarcoma, the undifferentiated component is made of small rounded cells instead of spindle shaped or obviously pleomorphic cells.

Collision tumor vs carcinosarcoma
Last are the composite or collision tumors. These are extremely uncommon. We have seen the following very rare examples: endometrial adenocarcinoma with adenosarcoma; endometrial adenocarcinoma with endometrial stromal nodule; endometrial adenocarcinoma with metastatic poorly differentiated carcinoma; endometrial adenocarcinoma with lymphoma.

A Managerial Approach

A) Monophasic tumors:
As stated previously, monophasic tumors should not be diagnosed as carcinosarcoma. However, either the mesenchymal or epithelial components of carcinosarcoma may predominate in small sample biopsies or scant curettage specimens. Therefore, the possibility of carcinosarcoma should be mentioned whenever a highly pleomorphic sarcoma, a high-grade carcinoma that is difficult to subclassify, or heterologous elements are encountered in a small sampling. Although a diagnosis of carcinosarcoma often cannot be made given the material at hand, the final diagnosis in such cases can be deferred to the hysterectomy specimen. From a clinical perspective, the next logical intervention is hysterectomy, regardless of tumor type.

B) Biphasic tumors:
To reiterate, a carcinosarcoma diagnosis can only be made confidently when distinct, high grade epithelial and mesenchymal components are found. Carcinosarcoma is also a reasonable diagnosis when both elements are high grade, but fused. When both elements are low grade and merge with one another, the diagnosis of endometrioid adenocarcinoma with spindle cell elements is warranted. Tumors showing high grade epithelial components and separate low grade spindle cell areas are difficult to subclassify as are tumors with low grade epithelial components and separate high grade mesenchymal zones. Fortunately, these tumors are uncommon. In general, the higher grade component should probably drive the diagnosis. A reasonable differential diagnosis for the latter tumors includes adenosarcoma with low grade carcinomatous elements and high grade sarcomatous elements, and an unusual carcinosarcoma. The type of epithelial elements present and whether they predominated would probably influence my final decision.

What constitutes "low" and "high grade" appearing mesenchymal components and the extent of mesenchymal differentiation sufficient for a carcinosarcoma diagnosis are probably not reproducible. These are the guidelines (admittedly subjective and empirical) that we use for separating "high grade" from "low grade" mesenchymal proliferations in the context of a biphasic uterine tumor: cellularity greater than that seen in typical ovarian fibroma; easily found mitotic figures in 3 to 5 fields at intermediate magnification; any atypical mitotic figure; nuclear hyperchromasia; nuclear pleomorphism recognizable at scanning magnification; 3X nuclear enlargement relative to non-neoplastic myometrial muscle fibers or non-neoplastic endometrial stroma. "High grade" mesenchymal elements usually feature almost all of these elements, but atypical mitotic figures can be scarce or absent. Regarding the extent of mesenchymal differentiation sufficient for a carcinosarcoma diagnosis, we would hesitate to make an outright diagnosis of carcinosarcoma if the malignant mesenchymal component could not be circled with a pen and measured; it would preferably measure more than 2 or 3 mm. If it did not, we would diagnose adenocarcinoma with spindle cell elements and describe the focal high grade mesenchymal appearance in a note. Of course, these practices are valid only when the lesion is well sampled; they are reasonable only when the pathologist and clinician understand each other's thresholds for diagnosis and therapeutic decisions. As patients with uterine carcinosarcoma are frequently entered onto protocols that involve treatment with highly toxic agents and are stigmatized as having one of the most aggressive gynecologic cancers, it is prudent to reserve an outright diagnosis of carcinosarcoma for lesions demonstrating absolutely classic morphology.

References
  1. McCluggage WG, Haller U., Kurman R.J., Kubik-Huch R.A. Mixed epithelial and mesenchymal tumours. In: Tavassoli F.A., Devilee P., editors. World Health Organization Classification of Tumours. Pathology and Genetics of Tumours of the Breast and Female Genital Organs. Lyon: IARC Press; 2003. p. 245-9.

  2. Bitterman P, Chun B, Kurman RJ. The significance of epithelial differentiation in mixed mesodermal tumors of the uterus. A clinicopathologic and immunohistochemical study. Am J Surg Pathol 1990 Apr;14(4):317-28.

  3. McCluggage WG. Malignant biphasic uterine tumours: carcinosarcomas or metaplastic carcinomas? J Clin Pathol 2002 May;55(5):321-5.

  4. McCluggage WG. Uterine carcinosarcomas (malignant mixed Mullerian tumors) are metaplastic carcinomas. Int J Gynecol Cancer 2002 Nov;12(6):687-90.

  5. Sreenan JJ, Hart WR. Carcinosarcomas of the female genital tract. A pathologic study of 29 metastatic tumors: further evidence for the dominant role of the epithelial component and the conversion theory of histogenesis. Am J Surg Pathol 1995 Jun;19(6):666-74.

  6. Silverberg SG, Major FJ, Blessing JA, Fetter B, Askin FB, Liao SY, et al. Carcinosarcoma (malignant mixed mesodermal tumor) of the uterus. A Gynecologic Oncology Group pathologic study of 203 cases. Int J Gynecol Pathol 1990;9(1):1-19.

  7. Yamada SD, Burger RA, Brewster WR, Anton D, Kohler MF, Monk BJ. Pathologic variables and adjuvant therapy as predictors of recurrence and survival for patients with surgically evaluated carcinosarcoma of the uterus. Cancer 2000 Jun 15;88(12):2782-6.

  8. George E, Lillemoe TJ, Twiggs LB, Perrone T. Malignant mixed mullerian tumor versus high-grade endometrial carcinoma and aggressive variants of endometrial carcinoma: a comparative analysis of survival. Int J Gynecol Pathol 1995 Jan;14(1):39-44.

  9. Major FJ, Blessing JA, Silverberg SG, Morrow CP, Creasman WT, Currie JL, et al. Prognostic factors in early-stage uterine sarcoma. A Gynecologic Oncology Group study. Cancer 1993 Feb 15;71(4 Suppl):1702-9.

  10. Soslow RA, Bissonnette JP, Wilton A, Ferguson SE, Alektiar K, Duska LR, Oliva E. Clinicopathologic analysis of 187 high-grade endometrial carcinomas of different histologic subtypes: similar outcomes belie distinctive biologic differences, Am J Surg Pathol 2007 Jul;31(7):979-87.

  11. Alektiar KM, McKee A, Lin O, et al. Is there a difference in outcome between stage I-II endometrial cancer of papillary serous/clear cell and endometrioid FIGO Grade 3 cancer? Int J Radiat Oncol Biol Phys. 2002;54:79–85.

  12. Ferguson SE, Tornos C, Hummer A, Barakat RR, Soslow RA. Prognostic features of surgical stage I uterine carcinosarcoma. Am J Surg Pathol. 2007 Nov; 31(11):1653-61.

  13. Murray SK, Clement PB, Young RH. Endometrioid carcinomas of the uterine corpus with sex cord-like formations, hyalinization, and other unusual morphologic features: a report of 31 cases of a neoplasm that may be confused with carcinosarcoma and other uterine neoplasms. Am J Surg Pathol 2005 Feb;29(2):157-66.

  14. Tornos C, Silva EG, Ordonez NG, Gershenson DM, Young RH, Scully RE. Endometrioid carcinoma of the ovary with a prominent spindle-cell component, a source of diagnostic confusion. A report of 14 cases. Am J Surg Pathol. 1995 Dec;19(12):1343-53.

  15. Silva EG, Deavers MT, Bodurka DC, Malpica A Association of low-grade endometrioid carcinoma of the uterus and ovary with undifferentiated carcinoma: a new type of dedifferentiated carcinoma? Int J Gynecol Pathol. 2006 Jan;25(1):52-8

  16. Altrabulsi B, Malpica A, Deavers MT, Bodurka DC, Broaddus R, Silva EG. Undifferentiated carcinoma of the endometrium. Am J Surg Pathol. 2005 Oct;29(10):1316-21.

  17. Seidman JD, Chauhan S. Seidman JD, Chauhan S. Evaluation of the relationship between adenosarcoma and carcinosarcoma and a hypothesis of the histogenesis of uterine sarcomas. Int J Gynecol Pathol. 2003 Jan;22(1):75-82.

  18. Clement PB, Scully RE. Mullerian adenosarcoma of the uterus: a clinicopathologic analysis of 100 cases with a review of the literature. Hum Pathol. 1990;21:363–381.