Introduction
This presentation briefly summarizes our current knowledge of the two most common types of non-endometrioid adenocarcinoma, uterine serous carcinoma (USC) and clear cell carcinoma (CC), emphasizing recent findings. The goal of the latter half of this discussion is to identify outstanding questions that could be addressed by members of our society.

Compared to endometrioid adenocarcinoma USC and CC: 1) portend a worse prognosis; 2) affect African Americans disproportionately and 3) likely differ in etiology and pathogenesis. Given, that these aggressive neoplasms are difficult to cure, early detection and prevention would be highly desirable, but this would require knowledge that we current lack.

Incidence data from the Surveillance, Epidemiology, and End Results Program (SEER) for the period 1992-1998 demonstrate that USC / CC (combined) are rare. The incidence of typical endometrial adenocarcinomas (mainly endometrioid) among white Hispanics and African Americans are similar, whereas the rate is about double for white Non-Hispanics. In contrast, the rate of USC/CC is about twice as high for African Americans compared to the other two groups.

Histologic Type	White Hispanic Incidence / 105 (C.I.)	African American Incidence / 105 (C.I.)	White Non-Hispanic
"Typical"	11.39 (10.80-12.00)	9.20 (8.66-9.77)	20.14 (19.79-20.49)
USC/CC	0.85 (0.69-1.03)	2.16 (1.90-2.45)	1.17 (1.09-1.26)

The following table compares rates and cumulative relative 5-year survival (survival of endometrial cancer patients to patients without endometrial cancer matched on age and race) for USC and CC separately. Survival data are not available for Hispanics.

Population	Serous Carcinoma Incidence (per 105) 5-Year Survival	Clear Cell Carcinoma Incidence (per 105) 5-Year Survival
African American	1.73 0.30	0.43 0.52
Whites	0.85 0.42	0.28 0.69

Incidence and survival patterns are similar for USC and CC among African Americans and whites; in both groups USC is more common and has a worse prognosis than CC. Several previous studies have shown that among African American women USC/CC accounts for a higher percentage of total endometrial adenocarcinomas than among whites, however analyses of percentages cannot distinguish whether this results from relatively fewer typical endometrial cancers among African Americans or a true excess of USC/CC. These demonstrate that the absolute incidence of USC/CC is significantly higher among African Americans (and the incidence of typical adenocarcinomas is lower). Typical endometrial cancers account for about 47% and USC/CC account for 18% of total mortality among African Americans, whereas among whites, the comparable figures are 64% and 12% respectively. Survival is worse for African Americans as compared to whites for all malignant tumors of the uterine corpus, including USC/CC. Only carcinosarcomas have a worse cumulative relative survival than USC and CC. The reasons for these racial differences are unknown.

Uterine Serous Carcinoma (USC)
The low incidence of USC has limited the collection of population-based data. Registry data provide useful descriptive data, but are limited by pathologic misclassification and lack of detailed clinical and epidemiologic data. One case-control study found that two established risk factors for endometrial cancer, obesity and use of exogenous estrogen, were not clear risk factors for USC. However, early age at menarche seemed to have been associated with increased risk for both tumor types and the relationships with protective factors (oral contraceptive use, parity and smoking) were similar. This analysis also found serum levels of all estrogenic hormones were lower among USC patients compared to controls, after adjusting for age and body mass index (differences for albumin-bound estradiol, estrone were significant). In another small retrospective case-control study, limited by access to consenting living subjects, women with USC tended to have a slightly higher non-obese weight compared to controls at 18 years of age (124 V.S. 117 lbs.), but similar weights at diagnosis. Among estrogen replacement users, USC was associated with a slightly longer mean duration of use. Overall, data favor a different etiology for USC and endometrioid adenocarcinomas, but a large well controlled epidemiologic study is needed to conclusively confirm and elucidate this finding.

The average age of women diagnosed with USC is higher than that of women diagnosed with endometrioid adenocarcinoma and nearly all are post-menopausal (median age ~65-70 years in most studies). Women generally present with abnormal bleeding and probably are more likely to present with an abnormal Pap smear or evidence of disseminated disease, especially ascites, than patients with usual histologic types of endometrial cancer.

Historically, pathologic criteria for USC were formally codified in a seminal paper by Hendrickson et al, which identified this entity as a frequent case of recurrence in clinical stage I endometrial cancer. This paper helped clarify confusion regarding the pathologic classification and behavior of papillary adenocarcinomas, by emphasizing the distinction of USC from villoglandular endometrioid adenocarcinoma. The term "serous carcinoma" has gained favor ("papillary" has been dropped), emphasizing that it is cellular differentiation not papillary formation that defines tumors with this distinctive clinical phenotype. Numerous clinicopathologic studies have been published on USC in the last 20 years, for the most part demonstrating consistent results.

The morphologic diversity of USC has been described; tumors may occur in pure form, in association with benign polyps or combined with other histologic types of endometrial adenocarcinoma (clear cell, endometrioid). USC usually arises in an atrophic background; when endometrial hyperplasia is identified, the tumor often shows mixed endometrioid and serous differentiation. In such cases, the areas of endometrial hyperplasia may give rise to the endometrioid component, whereas serous differentiation may represent "tumor progression." However, it is also possible that these tumors arise de novo with dual differentiation – more studies are needed.

Architecturally, USC may contain papillae (of diverse appearances), glands or solid areas. Tumors characteristically show areas of well-formed glands or papillae (well differentiated architecture) associated with marked nuclear atypia (poor cytologic grade). Given the universally high nuclear grade and poor prognosis of USC, tumors may be considered "high-grade" without further specification. Nearly all tumors are associated with strong, diffuse immunoreactivity for p53 antigen, which reflects cytoplasmic accumulation of mutant p53, which has increased stability (as opposed to wild type p53 which is usual unstable and undetectable). Most USCs consist of large, pleomorphic bizarre cells with hyperchromatic smudgy chromatin and irregular outlines, but in some cases the nuclei are monotonously atypical. Mitoses including abnormal figures are frequent. The tumor cells may be cuboidal, polygonal or hobnail in shape. The cytoplasm may be eosinophilic or clear and when the latter pattern is prominent a diagnosis of mixed serous / clear cell adenocarcinoma is appropriate. Most tumors deeply invade the myometrium, usually in a contiguous pattern, but sometimes a deceptively dispersed pattern is found. Vascular invasion is frequent.

Recognition that most cases of USC are not associated with endometrial hyperplasia has prompted studies to identify the "precursor" or "earliest identifiable" manifestation of these tumors. Multiple groups have found that malignant replacement of the endometrial surface epithelium or the surface of an otherwise typical polyp represents a morphologic lesion that may be found early in the natural history of USC; this lesions has been designated as "endometrial intraepithelial carcinoma (EIC)," "carcinoma in situ," "uterine surface carcinoma,' and "noninvasive endometrial serous carcinoma." This proliferation of terms represents multiple efforts to identify the best terminology for a rather unique gynecologic lesion, which has been increasingly well characterized in recent years. The identification of EIC suggests that USC may develop directly from the endometrial surface of atrophic endometrium or the surface of benign polyps. In addition, these observations present clinical challenges: 1) what are the pathologic criteria for EIC and early USC; 2) what is the optimal management for these lesions and 3) what are their prognoses.

Several studies (including case reports and small series) have described cases of USC in which there was either no invasion identified in the uterus or invasion limited to the endometrial stroma or to the stroma of a polyp. A broad consensus has emerged that these patients require meticulous staging. Patients with any amount of extra-uterine disease have a poor prognosis. Although the prognosis is better for USC confined to the uterus, a proportion of these tumors also relapse fatally. Distinguishing EIC from minimally invasive USC, especially when the lesion occurs in a polyp may be challenging; however, both lesions require staging, the results of which direct subsequent management. The term minimal uterine serous carcinoma (MUSC) has been proposed for cases of EIC or invasive USC < 1cm, which do not display lymph or blood vessel invasion to emphasize their similar behavior. The mechanism of dissemination outside the uterus in cases of MUSC has been debated; multi-focal disease ("field effect"); transtubal expulsion and unrecognized invasion of the myometrium or vessels represent possibilities. In a few cases, results of immunohistochemical or clonality assays have suggested that the extra-uterine foci in such cases represent metastases. Careful examination of the endometrium to search for MUSC should be performed in patients have who presented with abdominal carcinomatosis to avoid a misdiagnosis of primary peritoneal carcinoma.

The diagnosis of EIC may be challenging. Endometrial metaplasia likely represents the most perplexing differential diagnosis, especially when compounded by stromal breakdown with epithelial regeneration that results in nuclear degeneration and mitoses. Marked nuclear atypia, advanced age, strong diffuse p53 staining and a high Ki67 index favor a diagnosis of EIC; most metaplastic lesions occur in peri-menopausal women and nuclear atypia and p53 staining are generally less diffuse. The differential diagnosis of MUSC also includes endometrial hyperplasia in a polyp. Marked cellular pleomorphism, severe nuclear atypia and an atrophic background favor a diagnosis of EIC. Hyperplasia in a polyp is more common among peri-menopausal women and often involves the surrounding non-polypoid endometrium, where the diagnosis may be easier. An immunostaining pattern of p53 positive, ER and PR negative with high Ki67 index favors a diagnosis of EIC. Rarely the differential diagnosis of EIC/MUSC includes endocervical adenocarcinoma in situ or CIN3 with extension to the corpus; morphologic differences, stains (as noted above) and HPV testing may all be helpful. Finally, Liang and colleagues have suggested that lesser degrees of cellular atypia, which they have preliminarily termed, "endometrial glandular dysplasia," may represent a forerunner of EIC.

Surgical treatment for USC includes TAHBSO, omentectomy and lymph node dissection. To date, curative therapy for disseminated USC has not been described. Some groups have reported that whole abdominal radiation therapy is effective. Platinum containing regimens analogous to those used for serous carcinoma of the ovary have been tried, but some data has suggested that responses are not as good as for ovarian carcinoma. Interestingly, some studies have found that the molecular profiles of serous carcinomas of the endometrium and ovary also differ.

Clear Cell Carcinoma (CC)
Analysis of risk factors for CC has been limited by the rarity of this neoplasm. Scattered clinical descriptions of cases suggest that CC may not share identical risk factors with endometrioid adenocarcinoma, but compelling proof is lacking. Based on a retrospective clinicopathologic study of 21 CCs, one study divided cases of CC into two variants: "typical CC" (7 cases) and "CC with serous features" (14 cases). Both proposed types of CC were generally negative for hormone receptors and demonstrated a higher Ki67 index than endometrioid adenocarcinoma. Whereas typical CC was associated with endometrial hyperplasia, CC with serous features demonstrated a higher Ki67 index than typical CC, more frequent p53 abnormalities (by immunohistochemistry) and was more often associated with EIC. CC generally occurs at an older age than endometrioid adenocarcinoma, but the clinical presentation is similar. Based on small sample sizes, recent studies have found that the gene expression profiles for CC and USC revealed greater similarity than the comparison of either CC or USC with endometrioid adenocarcinoma.

Histologically, CCs may display tubulocystic, papillary or solid architecture. The cells may be cuboidal, or hobnail with clear or eosinophilic cytoplasm. Cytoplasmic clearing is attributable to glycogen accumulation. These tumors typically display marked nuclear atypia and several mitoses per 10 high power fields, although the atypical nuclei tend to be smaller and monotonous compared to USC. As noted above, many tumors classified as CC also show other histologic patterns, including USC, which may merge imperceptibly with classic appearing areas of CC. Many pathologists reserve the diagnosis of CC for tumors of mixed histology in which clear cell differentiation comprises at least 50% of the neoplasm. PAS positive diastase resistant hyaline globules and thin appearing watery mucin within gland lumina are characteristic when present. A thin fibrovascular network commonly traverses large regions of CC composed of solid sheets. The prognosis of CC is poor, with stage representing the best predictor of outcome. Generally, these tumors are not specifically graded and are simply considered high-grade. In contrast to USC, which often spreads intra-abdominally like ovarian carcinoma diffusely encasing viscera, CC may metastasize more similarly to typical endometrioid adenocarcinomas as solid masses.

CC is distinguished from USC by its characteristic architecture and increased tendency to display hyaline fibrosis, hyaline globules and thin mucus. Yolk sac tumors may also contain clear cells, but these tumors are extraordinarily unusual in the uterus, and unlike CC are immunoreactive for AFP. It is most critical to distinguish CC from low-grade tumors with clear cells, including secretory adenocarcinoma and endometrioid adenocarcinoma with glycogenated squamous metaplasia. Several distinguishing features are useful, notably the more severe nuclear atypia in CC, identification of architectural patterns and cell morphology characteristic of CC, the identification of sub-nuclear vacuoles in secretory carcinomas and evidence of clear cells merging with foci of definite squamous differentiation in endometrioid adenocarcinomas.

Data from selected studies of clear cell carcinoma are presented below.

Source	# Cases	% Total Cancers Classified as CC	Age	Survival (5 years)
				Stage I	Stage II
Photopulos '79	22	3	66	73	21
Webb '87	29	4	70-79	72	66
Abeler '91	97	5	65	59	27
Carcangiu '95	29 (Stage I, II)	NA	69	72	59

Note that the studies are small in size and retrospective, often spanning many years of case ascertainment. In clinical reports, CC usually accounts for < 5% of all endometrial adenocarcinomas; SEER data shows differences in incidence by race. Age of patients has been consistent across studies. Survival is probably stage related. Overall, survival outcomes in other studies: ~71% for 17 stage I and II patients (Malpica '95), 60% crude survival among 20 patients, 11 of 12 survivors stage I CCs (Kanbour-Shakir '91), 34% crude survival among 55 patients, all survivors stage I (Chirstopherson '82). There are scattered reports of survival for 5 years with later recurrence and death from tumor. In the future, as better treatment prolongs survival without necessarily achieving complete cures, it is likely that accurate case fatality estimates will require more than 5 years of follow-up.

New Directions
Expanding our understanding of USC and CC is important from an etiological, public health and clinical perspective. Given the rarity of these tumors, it is difficult for single institutions to develop meaningful studies of these entities; enlarged study sets that are highly enriched with consult material may be useful for addressing some questions, but are prone to biases. Accordingly, meetings of experts in gynecological pathology, such as this companion meeting, represent opportunities to air new concepts and hopefully foster productive collaborations. The following short list of topics is presented in this spirit. Certainly, many of these ideas, which are likely the subject of ongoing or planned work, could receive an added boost through further discussions.

Histopathologic Diagnosis

• Establish the reproducibility of the diagnosis of USC, EIC and CC; identify histologic features that lead to poor agreement and propose criteria for clarifying interpretation; address the concept of sub-typing CCs into "typical" CC and a variant with "serous features"
• Although most villoglandular and serous carcinomas containing papillae are easily distinguished, some examples showing borderline nuclear atypia may pose diagnostic problems
• The differential diagnosis of EIC and benign metaplasias may be difficult; further work in this area would be helpful (molecular, immunohistochemical, image analysis)
• The concept of endometrial glandular dysplasia requires additional study
• Refined criteria for distinguishing solid patterns of CC from squamous metaplasia in endometrioid adenocarcinoma may be helpful, especially for generalists
• Determine whether CCs are homogeneous or heterogeneous with regard to precursor lesions, behavior and etiology; the pathogenesis of USCs associated with other histologic patterns also could be clarified
• Consider protocols for histologic processing of uteri removed to treat USC or CC; should these differ from processing protocols for endometrioid adenocarcinomas and if so how?
• Define immunostaining patterns that are useful in the differential diagnosis of USC and CC and search for prognostic markers in early stage lesions (tissue microarrays enriched for USC and CC may be useful)
• Determine whether the expanded diagnostic criteria for USC that have been used in the last 5-10 years, has obviated the needed for a FIGO grading scheme for well differentiated endometrioid adenocarcinomas with marked nuclear atypia – are all of these tumors classified as USC currently?

Molecular Analysis

• Develop a clinical assay with high negative predictive value for extra-uterine disease among women with pathologic stage I tumors, permitting conservative treatment.
• Using existing knowledge and emerging data from expression arrays, develop a PCR based assay for peritoneal washes or circulating cells in blood that predicts recurrences

• Molecular profiling (RNA, DNA methylation, proteomics, etc.) of USC and CC to identify alterations in pathways that might be exploited for diagnosis and treatment Determine whether p53 mutation is truly an early event in the development of USC; explore the significance of p53 abnormalities in typical endometrial adenocarcinoma – does this imply that the tumor should now be classified as a mixed USC
• Further characterizations of surface epithelium, especially: benign atrophic, metaplastic, and surface lining of endometrial polyps; explore the apparent association between polyps and CC and USC – is there an increased risk and if so, is this simply because there are more about the surface cells of polyps which predisposes to USC and CC. Are the surface of these polyps components of an endometrial hyperplasia that has regressed but left behind mutated cells or are the cells of polyps exposed to more genetic stresses or carcinogens? How does genetic damage accumulate in endometrial surface epithelium and does this related to the development of p53 mutations?
• Exploration of epithelial-stromal interactions

• Analysis of risk factors for USC and CC in a sufficiently large data set with expert pathology review and access to other biological samples (blood, urine, germline DNA); confirm whether protective factors for all types of endometrial cancer are similar. Explore whether exogenous hormones or anti-inflammatory drugs protect against USC and possibly CC.
• Explain the increased incidence of USC, CC and other aggressive tumors among African American women in the U.S. (and the lower incidence of endometrioid adenocarcinomas)
• Determine whether tamoxifen is associated with increased risk for USC, CC

Selected References

Serous Carcinoma

1. Hendrickson M, Ross J, Eifel P, et al. Uterine papillary serous carcinoma: a highly malignant form of endometrial adenocarcinoma. Am J Surg Pathol 1982;6:93-108.
2. Carcangiu ML, Chambers JT. Uterine papillary serous carcinoma: a study of 108 cases with emphasis on the prognostic significance of associated endometrioid adenocarcinoma, absence of invasion, and concomitant ovarian carcinoma. Gynecol Oncol 1992;47:298-305.
3. Sherman ME, Bitterman P, Rosenshein NB, et al. Uterine serous carcinoma: a morphologically diverse neoplasm with unifying clinicopathologic features. Am J Surg Pathol 1992;16:600-10.
4. Ambros RA, Sherman ME, Zahn CM, et al. Endometrial intraepithelial carcinoma: a distinctive lesion specifically associated with tumors displaying serous differentiation. Hum Pathol 1995:26:1260-7.
5. Liang SX, Parkash V, Zheng W. Endometrial glandular dysplasia, a preneoplastic lesion of uterine surface carcinoma/endometrial intraepithelial carcinoma: Part I, morphological studies.(USCAP 2003, submitted abstract).
6. Tashiro H, Isacson C, Levine R, et al. p53 gene mutations are common in uterine serous carcinoma and occur early in their pathogenesis. Am J Pathol 1997;15):177-85.
7. Lee KR, Belinson JL. Recurrence in noninvasive endometrial carcinoma. Am J Surg Pathol 1991;15:965-73.
8. Aquino-Parsons C, Lim P, Wong F, et al. Papillary serous and clear cell carcinoma limited to endometrial curettings in FIGO Stage 1a and 1b endometrial adenocarcinoma: treatment implications. Gynecol Oncol 1998;71:83-6.
9. Carcangiu ML, Tan LK, Chambers JT. Stage IA uterine serous carcinoma. A study of 13 cases. Am J Surg Pathol 1997;21:1507-14.
10. Cirisano FD, Robboy SJ, Dodge RK, et al. The outcome of stage I-II clinically and surgically staged papillary serous and clear cell endometrial cancers when compared with endometrioid carcinoma. Gynecol Oncol 2000;77:55-65.
11. Cirisano FD, Robboy SJ, Dodge RK, et al. Epidemiologic and surgicopathologic findings of papillary serous and clear cell endometrial cancers when compared to endometrioid carcinoma. Gynecol Oncol 1999;74:385-94.
12. Gehrig PA, Groben PA, Fowler WC. Noninvasive papillary serous carcinoma of the endometrium. Obstet Gynecol 2001;97:153-7.
13. Goff BA, Kato D, Schmidt RA, et al. Uterine papillary serous carcinoma: patterns of metastatic spread. Gynecol Oncol 1994;54:264-8.
14. Lim P, AL Kushi A, Gilks B, et al. Early stage uterine papillary serous carcinoma of the endometrium. Effect of adjuvant whole abdominal radiotherapy and pathologic parameters on outcome. Cancer 2001;91:752-7.
15. Soslow RA, Pirog E, Isacson C. Endometrial intraepithelial carcinoma with associated peritoneal carcinomatosis. Am J Surg Pathol 2000;24:726-32.
16. Wheeler DT, Bell KA, Kurman RJ, et al. Minimal uterine serous carcinoma: Diagnosis and clinicopathologic correlation. Am J Surg Pathol 2000;24:797-806.
17. Yan ZJ, Parkash V, Zheng WX, et al. Minimal uterine serous carcinoma with extrauterine tumor of identical morphology: A pathological study of 13 cases. (USCAP 2003, submitted abstract).
18. Kupryjanczyk J, Thor AD, Beauchamp R, et al. Ovarian, peritoneal, and endometrial serous carcinoma: clonal origin of multifocal disease. Mod Pathol 1996;9:166-73.
19. Risinger JI, Maxwell GL, Chandramouli GVR, et al. Microarray analysis reveals distinct gene expression profiles among different histologic types of endometrial cancer.
20. Sherman ME. Theories of endometrial carcinogenesis: a multidisciplinary approach. Mod Pathol 2000;13:295-308.

21. Abeler VM, Kjorstad KE. Clear cell carcinoma of the endometrium: A histopathological and clinical study o 97 cases. Gynecol Oncol 1991;40:207-17.
22. Malpica A, Tornos C, Burke TW, et al. Low-stage clear-cell carcinoma of the endometrium. Am J Surg Pathol 1995;19:769-774.
23. Christopherson WM, Alberhasky RC, Connelly PJ. Carcinoma of the endometrium: I. A clinicopathologic study of clear-cell carcinoma and secretory carcinoma. Cancer 1982;49:1511-23.
24. Kanbour-Shakir A, Tobon H. Primary clear cell carcinoma of the endometrium: A clinicopathologic study of 20 cases.
25. Kurman RJ, Scully RE. Clear cell carcinoma of the endometrium. An analysis of 21 cases. Cancer 1976;37:872-82.
26. Silverberg SG, De Giorgi LS. Clear cell carcinoma of the endometrium. Clinical, pathologic, and ultrastructural findings. Cancer 1973;31:1127-40.
27. Photopulos GJ, Carney CN, Edelman DA, et al. Clear cell carcinoma of the endometrium. Cancer 1979;43:1448-56.
28. Lax SF, Pizer ES, Ronnett BM. Clear cell carcinoma of the endometrium is characterized by a distinctive profile of p53, Ki-67, estrogen, and progesterone receptor expression. Hum Pathol 1998;29:551-8.
29. Webb GA, Lagios MD. Clear cell carcinoma of the endometrium. Am J Obstet Gynecol 1987;156:1486-91.
30. Carcangiu ML, Chambers JT. Early pathologic stage clear cell carcinoma and uterine papillary serous carcinoma of the endometrium: Comparison of clinicopathologic features and survival. Int J Gynecol Pathol 1995;14:30-8.