The prognosis of ovarian cancer can be correlated with numerous clinical and biologic factors. Tumor stage, grade, and size of metastatic disease after resection correlate best with outcome. As discussed below, among patients with low stage disease, tumor grade correlates with prognosis (ie, patients with stage I high-grade lesions have a shorter survival than those of low-grade). In patients with advanced stage disease, the size of residual disease after surgery correlates most clearly with survival.36 The rapidity with which disease regresses during chemotherapy also correlates with survival. A short apparent half-life of the serum tumor marker CA 125 has correlated with improved survival in more than a dozen studies. Normalization of CA 125 by the third course of chemotherapy has been associated with a favorable prognosis.

Several biologic factors have been correlated with prognosis in epithelial ovarian cancer. Using flow cytometry, Friedlander and others have shown that ovarian cancers are commonly aneuploid and that a correlation exists between FIGO stage and ploidy, (low-stage cancers tend to be diploid and high-stage tumors tend to be aneuploid). Patients with diploid tumors have a significantly longer median survival than those with aneuploid tumors: 5 years versus 1 year, respectively. Multivariate analyses have demonstrated that ploidy is an independent prognostic variable and one of the most significant predictors of survival. Flow cytometric analysis also provides data on the cell cycle, and the proliferation fraction (S phase) determined by this technique has correlated with prognosis in some studies. More recent reports utilizing comparative genomic hybridization suggest that copy number abnormalities for chromosome segments also have prognostic significance.

Most sporadic ovarian carcinomas evolve from a single clone of cells. When primary cancers and metastases have been compared, more than 90% share the same patterns of loss of heterozygosity, inactivation of the same X chromosome and, when present, the same mutations in p53. Ovarian cancers can metastasize by multiple routes. Like other epithelial neoplasms, ovarian cancers can spread through lymphatics to the level of the renal hylus and can also spread hematogenously. Most frequently, however, ovarian cancer spreads over the surface of the peritoneum, studding the serosal surface of the bowel and abdominal wall and ultimately producing intestinal obstruction. Ascites formation results from increased leakage of proteinaceous fluid from capillaries under the influence of VEGF/VPF produced by ovarian cancers and from inhibition of fluid outflow through diaphragmatic lymphatics that have been blocked by metastatic disease. Studies of the immunobiology of the peritoneal cavity suggest that it may function as an immunoprivileged site, with elevated levels of suppressive molecules and growth factors.

Epithelial ovarian cancers are thought to arise from a single layer of cells that covers the ovary or that lines cysts immediately beneath the ovarian surface. These cells are generally quiescent, but proliferate following ovulation to repair the defect created by rupture of a follicle. A distinctive profile of genetic and epigenetic alterations has been observed in early and late stage ovarian cancers. A number of oncogenes have been overexpressed and/or activated (

Table 118-1) and the function of several tumor suppressor genes has been lost (

Table 118-2). Among the oncogenes activated, particular attention has been given to the HER family of transmembrane tyrosine kinase growth factor receptors, including the epidermal growth factor receptor (EGFR) and c-erbB-2 (HER-2). EGFR is expressed by normal ovarian surface epithelial cells and EGFR expression is lost in approximately 30% of ovarian cancers, associated with an improved prognosis. HER-2 is overexpressed in approximately 15%to 30% of ovarian cancers and, in some studies, has been associated with a poor prognosis. The ras oncogene is mutated and activated in less than 20% of serous ovarian cancers, but is more frequently mutated in mucinous and borderline neoplasms. In addition, ras may be physiologically activated in a larger fraction of cancers. Myc is amplified and overexpressed in approximately 33% of cases. More frequent and consistent abnormalities have been found in the PI3 kinase signaling pathway. The p110 catalytic subunit of the PI3 kinase is amplified, overexpressed and activated in more than half of ovarian cancers. Downstream within the PI3 kinase signaling pathway, AKT is also amplified, overexpressed and activated in 20% of cases. PTEN is mutated in endometrioid ovarian cancers. Overall, abnormalities of the PI3 kinase pathway have been found in more than 70% of ovarian cancers.

Recently, a murine model has been developed that mimics human ovarian cancer and that permits the introduction of multiple genes using an avian retroviral gene delivery technique. When the ovarian target cells were derived from transgenic mice deficient for p53, the addition of any two of the oncogenes c-myc, k-ras and AKT were sufficient to induce ovarian tumor formation when infected cells were injected at intraperitoneal, subcutaneous or ovarian sites. Another model has been developed from normal human ovarian surface epithelial cells that have also been immortalized with telomerase and with viral T antigen which neutralizes p53 and RB function. Introduction of activated human H-ras or K-ras genes transforms these immortalized cells and permits them to grow in immunosuppressed mice with a nodular or papillary histology that resembles human ovarian cancer.

A number of tumor suppressor genes have been studied in ovarian cancers. The RB and VHL genes can be downregulated in a fraction of ovarian cancers, but their function remains intact. P53 is mutated and consequently overexpressed in 50& to 60% of advanced ovarian cancers, but in only 15% of stage I disease. p53 Is rarely mutated in borderline cancers, but when this occurs, it is associated with a poor prognosis. The pattern of p53 abnormalities is most consistent with spontaneous mutation rather than the activity of chemical carcinogens. Novel tumor suppressor genes have been discovered in recent years. ARHI and Lot-1 are imprinted genes with only one functioning allele in adult cells. As the “first hit” has already occurred during imprinting, loss of function can occur with a single genetic or epigenetic event. Loss of the functional allele of ARHI can occur through loss of heterozygosity, methylation and transcriptional downregulation. ARHI is homologous to the oncogenes ras and rap, but has the opposite function, inhibiting cell proliferation, motility and invasion. Overexpression of ARHI induces apoptosis and may prove useful, as may other suppressor genes, for gene therapy alone or in combination with standard cytotoxic agents.

Several autocrine or paracrine factors affect the growth of normal and transformed ovarian epithelial cells including epidermal growth factor (EGF) transforming growth factor (TGF)-α, TGF-β, tumor necrosis factor (TNF)-α, interleukin-1 (IL-1), interleukin-6 (IL-6), macrophage colony stimulating factor (M-CSF, CSF-1) and lysophosphatidic acid (LPA). Different ovarian cancers exhibit abnormal production of these factors, altered receptor function and/or altered signaling. A fraction of ovarian cancers have lost expression of EGFR and consequently the ability to be stimulated by EGF and TGF-α, and this loss is associated with an improved prognosis. In contrast to normal epithelium, the proliferation of many ovarian cancer cells from different patients may lose the ability to express TGF-βeta and may be only partially inhibited by exogenous TGF-βeta. Thus, the inhibitory effect that TGF-βeta exerts in normal epithelium may be lost during transformation. Interestingly, TGF-βeta can produce apoptosis in malignant cells, but not in normal cells. TNF can stimulate, fail to affect or inhibit growth of different cell lines. As ovarian cancers transform, receptors for M-CSF (fms) and for LPA (edg-7) are upregulated resulting in greater proliferation, invasiveness, and resistance to apoptosis in the presence of the ligands.

Several additional molecular alterations contribute to the ability of ovarian cancers to metastasize. Alteration in integrins and expression of altered CD44 promote adhesion to mesothelial cells lining the peritoneum, Upregulation of proteases enhances invasion. Several angiogenic factors are secreted including IL-8, β FGF and VEGF. The VEGF molecule is also known as vascular permeability factor that increases leakage of proteinacious material from capillaries and the formation of ascites.

Upregulation and aberrant glycosylation of extracellular mucins have provided targets for therapy and markers for monitoring disease. MUC-1 is a mucin expressed by more than 80% of ovarian cancers. In transformed cells, aberrant glycosylation exposes peptide determinants recognized by murine monoclonal antibodies that have been used for serotherapy. CA 125 is also a mucin (MUC-16) associated with cells that line the coelomic cavity during embryonic development. CA125 is shed from 80% of epithelial ovarian cancers and can be measured using the murine monoclonal antibody OC125. Regression and progression of disease tend to correlate well with falling or rising CA125 levels. The precise function of the glycoprotein is unknown.