Current topicPrognostic parameters of endometrial carcinoma
Section snippets
Histological type
Cell type has consistently been recognized as an important predictor of the biological behavior of endometrial carcinoma. Over the last 15 years, based on Bokhman’s clinical model for explaining the pathogenesis of endometrial carcinomas,7 these tumors have been classified into 2 types. Type I tumors (about 80%) are endometrioid carcinomas (Fig 1), frequently preceded by complex atypical hyperplasia and associated with estrogenic stimulation. Type I tumors occur predominantly in premenopausal
Histological grade
Whereas nonendometrioid carcinomas are considered high-grade tumors by definition and thus need not be graded, the grading of endometrioid carcinomas is prognostically important. These tumors span the spectrum from very well differentiated to almost completely undifferentiated, and probably merge at the higher end of the spectrum with undifferentiated carcinomas.31 The prognostic value of grading endometrial adenocarcinomas has been recognized for many years.6, 32 The 1988 International
FIGO stage
Although not perfect, FIGO stage is the single strongest prognostic parameter for women with endometrial carcinoma.37, 38, 39, 40 The 1971 FIGO clinical staging system (based on sounding of the uterus, fractional curettage, and pelvic examination) proved to be less accurate than histological evaluation of the hysterectomy specimen.37, 38, 39 Determination of the spread of tumor to the cervix (stage II) based on fractional curettage had an error rate of approximately 50%.37, 39 A decrease in
Intraendometrial carcinoma (stage IA)
Endometrial stromal invasion is manifested by various architectural and cytological features, some of which may be absent in individual cases, leaving the final decision to the overall evaluation of the lesion. In most cases, there is confluent arrangement of clearly malignant glands that may result in either a cribriform pattern or extensive papillary growth into the glandular lumina. Rarely, however, the invasive glands lack significant nuclear atypia, and the diagnosis of stromal invasion
Myometrial invasion
In low-stage endometrial carcinomas, myometrial invasion is an independent predictor of outcome.3, 4, 32, 36, 44, 45, 46 Recognition of myometrial invasion is straightforward when glands that are irregularly shaped, randomly distributed, and surrounded by a desmoplastic stroma infiltrate the myometrium (Fig 8). Occasionally, however, a carcinoma that has not invaded the myometrium exhibits an irregular endomyometrial junction in which rounded nests of tumor seem to protrude into the inner
Vascular invasion
The presence of tumor cells within endothelial-lined spaces is a strong predictor of recurrence and death from tumor, independent of histological grade or depth of myometrial invasion.57, 58, 59 As stated earlier, stromal retraction is a frequent artifact that may simulate vascular invasion. Immunoperoxidase staining of endothelial cells (factor VIII or CD31) may facilitate recognition of vascular channels.46 In addition, a perivascular lymphocytic infiltrate is often associated with vascular
Endometrial hyperplasia
Endometrioid carcinomas that are accompanied by atypical endometrial hyperplasia and exhibit various metaplasias, particularly ciliated and eosinophilic metaplasia, are associated with a favorable prognosis. These tumors are frequently low-grade carcinomas and lack myometrial invasion.60, 61 Alternatively, high-grade tumors are often accompanied by an atrophic endometrium.
Cervical involvement (stage II)
Cervical involvement in endometrial carcinoma, a finding present in 6% to 20% of cases, usually results from direct surface or stromal extension, although it may be secondary to implantation or lymphatic spread.39, 62, 63, 64 Implantation occurs in the denuded endocervix after fractional curettage in about 5% of cases.62, 64 As stated earlier, histological diagnosis of cervical involvement on the basis of the findings from endocervical curettage can be difficult.37, 39, 64 The diagnosis of
Serosal invasion, peritoneal cytology, and adnexal involvement (stage IIIA)
In the view of many gynecologic oncologists, the criteria used for classifying an endometrial carcinoma as stage III tumor (6% of cases) are somewhat artificial.65 The adverse prognostic significance of stage IIIA tumors (based on tumor involvement of the uterine serosa or positive peritoneal cytology) (Fig 15) has not been universally confirmed.69 Serosal invasion represents the extreme of deep myometrial invasion of stage I disease. It is quite different from adnexal spread, and according to
Lymph node metastasis (stage IIIC)
Endometrial carcinomas associated with pelvic and/or paraaortic lymph node metastasis are classified as stage IIIC tumors33 (Fig 16). Almost 33% of patients with positive pelvic lymph nodes have positive para-aortic nodes as well. Prognosis with positive pelvic lymph nodes seems more favorable than that with positive para-aortic lymph nodes.4 In fact, in a Gynecologic Oncology Group study, only 36% of patients with positive aortic nodes were free of tumor at 5 years, compared with 85% of those
Ploidy
Approximately 67% of endometrioid carcinomas are diploid, as evaluated by flow or static cytometry. In contrast, 55% of the nonendometrioid carcinomas (ie, serous, clear cell, or undifferentiated carcinomas) exhibit nondiploid DNA patterns. Diploid tumors are usually low-grade endometrioid carcinomas with only superficial invasion, and are associated with longer survival than aneuploid carcinomas.72, 73, 74, 75 Differences in disease-free survival for stage I tumors have been as significant as
Steroid receptors
Most endometrioid carcinomas contain cells with both ERs and progesterone receptors (PRs) as a sign of differentiation. Currently, assessment of ERs and PRs is done mainly by immunohistochemistry. Characteristically, endometrioid carcinomas exhibit marked heterogeneity in ER and PR distribution. The presence and quantity of steroid receptors correlate with FIGO stage, histological grade, and survival.79, 80, 81, 82 There is also correlation between ER and PR and bcl-2, absence of mutant p53,
bcl-2
bcl-2 is a proto-oncogene that inhibits programmed cell death or apoptosis. The immunohistochemical expression of Bcl-2 protein varies during the menstrual cycle; it is highly expressed in the proliferative phase, with down-regulation during the secretory phase. High levels of Bcl-2 are found in endometrial hyperplasia, with decreased expression in adenocarcinoma. Loss of Bcl-2 has been associated with poorer prognosis, increased depth of invasion, higher FIGO stage, and aggressive cell types,83
c-erb-B2 (HER2/neu) and p53
Amplification or overexpression of the c-erb-B2 (HER2/neu) oncogene occurs in about 20% to 40% of endometrial carcinomas (Fig 17) and has been associated with other adverse prognostic factors, including advanced stage, higher grade, and worse overall survival.86, 87 However, some studies have found that c-erb-B2 is not independently associated with adverse prognostic factors, although it does seem to have an influence on overall survival.88, 89 Currently, c-erb-B2 is thought to be a parameter
Markers of proliferation
Mitotic count, S-phase fraction by flow cytometry, and proportion of proliferating cells by immunohistochemistry (PCNA, Ki-67, MIB-1) are the most commonly used methods of quantifying proliferation. Ki-67 and MIB-1 identify cells in most of the G1, S, G2, and M phases of the cell cycle.91, 92, 93 PCNA and Ki-67 expression in stage I endometrial carcinoma correlate with histological grade, depth of myometrial invasion, and risk of recurrence. Most endometrioid carcinomas express a low Ki-67
Microsatellite instability
Microsatellites are repetitive DNA sequences distributed widely throughout the genome. Because of their repetitive structure, microsatellites are particularly susceptible to replication errors, resulting in insertion or deletion mutations (ie, MSI).94 The genes responsible for MSI encode proteins involved in DNA mismatch repair, including MLH1, MLH6, MSH2, MSH3, and PMS2. Mutations of these genes alter the ability of the cells to repair replication errors. Although MSI was initially found in
PTEN
The tumor-suppressor gene PTEN (phosphatase and tensin homolog deleted from chromosome 10), also called MMAC1 (mutated in multiple advanced cancers) is mutated in 30% to 60% of endometrial carcinomas.98, 99 PTEN is located on chromosome 10q23.3 and encodes a phosphatidylinositol phosphatase. The product induces apoptosis and G1 cell cycle arrest through antagonizing the phosphatidylinositol 3′-kinase/Akt-mediated cell growth pathway.100 PTEN mutation occurs almost exclusively in endometrioid
Beta-catenin
Beta-catenin is involved in the maintenance of tissue architecture and cell polarity and also plays a role in transcriptional activation. Cytoplasmic levels of beta-catenin are regulated by the opposing actions of Wnt-1 and APC. An increase in Wnt-1 stimulation or a decrease in beta-catenin degradation results in nuclear accumulation of beta-catenin and transcriptional activation. Beta-catenin gene (CTNNB1) mutations have been reported in approximately 15% to 40% of low-grade endometrioid
K-ras
Mutations in codon 12 of K-ras occur in approximately 20% of endometrial carcinomas, mostly of the endometrioid type with MSI. In most series, the ras mutations have not been related to stage, grade, depth of myometrial invasion, or survival.110
Acknowledgements
The author thanks Alberto Gallardo, MD for providing clinicopathologic data and Julie A. Irving, MD for her assistance in preparing the manuscript.
References (110)
- et al.
Risk factors and recurrent patterns in stage I endometrial cancer
Am J Obstet Gynecol
(1985) - et al.
Relationship between surgical-pathological risk factors and outcome in clinical stage I and II carcinoma of the endometrium. A Gynecologic Oncology Group study
Gynecol Oncol
(1991) Two pathogenetic types of endometrial carcinoma
Gynecol Oncol
(1983)- et al.
Endometrial intraepithelial carcinomaA distinctive lesion specifically associated with tumors displaying serous differentiation
Hum Pathol
(1995) - et al.
Molecular pathology of endometrial hyperplasia and carcinoma
Hum Pathol
(2001) - et al.
Microsatellite instability in endometrial carcinomasClinicopathological correlations in a series of 42 cases
Hum Pathol
(1998) - et al.
PTEN mutations in endometrial carcinomas. A molecular and clinicopathological analysis of 38 cases
Hum Pathol
(2000) - et al.
Loss of heterozygosity in usual and special variant carcinomas of the endometrium
Hum Pathol
(1997) - et al.
Uterine papillary serous carcinomaA study of 108 cases with emphasis on the prognostic significance of associated endometrioid carcinoma, absence of invasion and concomitant ovarian carcinoma
Gynecol Oncol
(1992) - et al.
Serous papillary carcinoma of the endometriumA histopathological study of 22 cases
Gynecol Oncol
(1990)