Incidence and Epidemiology
Epithelial ovarian cancer is the leading cause of death from gynecologic cancer in the United States. In 2004, 25,580 new cases were diagnosed and 16,090 women died from ovarian cancer. The disease accounts for 5% of all cancer deaths in women in the United States; more women die of this disease than from cervical and endometrial cancer combined.
The age-specific incidence of the common epithelial type of ovarian cancer increases progressively and peaks in the eighth decade. Epithelial tumors, unlike germ cell and stromal tumors, are uncommon before the age of 40. Epidemiologic studies suggest higher incidences in industrialized nations and an association with disordered ovarian function, including infertility, nulliparity, frequent miscarriages, and use of ovulation-inducing drugs such as clomiphene. Each pregnancy reduces the ovarian cancer risk by about 10%, and breast-feeding and ubal ligation also appear to reduce the risk. Oral contraceptives reduce the risk of ovarian cancer in patients with a familial history of cancer and in the general population. Many of these risk-reduction factors support the “incessant ovulation” hypothesis for ovarian cancer etiology, which implies that an aberrant repair process of the surface epithelium is central to ovarian cancer development. Estrogen replacement after menopause does not appear to increase the risk of ovarian cancer, although one study showed a modest increase in risk with >11 years of use.
Familial cases account for about 5% of all ovarian cancer, and a family history of ovarian cancer is a major risk factor. Compared to a lifetime risk of 1.6% in the general population, women with one affected first-degree relative have a 5% risk. In families with two or more affected first-degree relatives, the risk may exceed 50%. Three types of autosomal dominant familial cancer are recognized: (1) site-specific, in which only ovarian cancer is seen; (2) families with cancer of the ovary and breast; and (3) the Lynch type II cancer family syndrome with nonpolyposis colorectal cancer, endometrial cancer, and ovarian cancer.
Etiology and Genetics
In women with hereditary breast/ovarian cancer, two susceptibility loci have been identified: BRCA1, located on chromosome 17q12-21, and BRCA2, on 13q12-13. Both are tumor-suppressor genes, and their protein products act as inhibitors of tumor growth. Both genes are large, and numerous mutations have been described; most are frameshift or nonsense mutations, and 86% produce truncated protein products. The implications of the many other mutations including many missense mutations are not known. The cumulative risk of ovarian cancer with critical mutations of BRCA1 or -2 is 25%, compared to the lifetime risk of 50% for breast cancer for similar mutations. Men in such families have an increased risk of prostate cancer.
Cytogenetic analysis of sporadic epithelial ovarian cancers generally reveals complex karyotypic rearrangements. Structural abnormalities frequently appear on chromosomes 1 and 11, and loss of heterozygosity is common on 3q, 6q, 11q, 13q, and 17. Abnormalities of oncogenes are frequently found in ovarian cancer and include c-myc, H-ras, K-ras, and neu.
Ovarian tumors (usually not epithelial) are sometimes components of complex genetic syndromes. Peutz-Jeghers syndrome (mucocutaneous pigmentation and intestinal polyps) is associated with ovarian sex cord stromal tumors and Sertoli cell tumors in men. Patients with gonadal dysgenesis (46XY genotype or mosaic for Y-containing cell lines) develop gonadoblastomas, and women with nevoid basal cell carcinomas have an increased risk of ovarian fibromas.
Clinical Presentation and Differential Diagnosis
Most patients with ovarian cancer are first diagnosed when the disease has already spread beyond the true pelvis. The occurrence of abdominal pain, bloating, and urinary symptoms usually indicates advanced disease. Localized ovarian cancer is generally asymptomatic. However, progressive enlargement of a localized ovarian tumor can produce urinary frequency or constipation, and rarely torsion of an ovarian mass causes acute abdominal pain or a surgical abdomen. In contrast to cervical or endometrial cancer, vaginal bleeding or discharge is rarely seen with early ovarian cancer. The diagnosis of early disease usually occurs with palpation of an asymptomatic adnexal mass during routine pelvic examination. However, most ovarian enlargements discovered this way, especially in premenopausal women, are benign functional cysts that characteristically resolve over one to three menstrual cycles. Adnexal masses in premenarchal or postmenopausal women are more likely to be pathologic. A solid, irregular, fixed pelvic mass is usually ovarian cancer. Other causes of adnexal masses include pedunculated uterine fibroids, endometriosis, benign ovarian neoplasms, and inflammatory lesions of the bowel.
Evaluation of patients with suspected ovarian cancer should include measurement of serum levels of the tumor marker CA-125. CA-125 determinants are glycoproteins with molecular masses from 220 to 1000 kDa, and a radioimmunoassay is used to determine circulating CA-125 antigen levels. Between 80 and 85% of patients with epithelial ovarian cancer have levels of CA-125 ≥ 35 U/mL. Other malignant tumors can also elevate CA-125 levels, including cancers of the endometrium, cervix, fallopian tubes, pancreas, breast, lung, and colon. Certain nonmalignant conditions that can produce moderate elevations of CA-125 levels include pregnancy, endometriosis, pelvic inflammatory disease, and uterine fibroids. About 1% of normal females have serum CA-125 levels >35 U/mL. However, in postmenopausal women with an asymptomatic pelvic mass and CA-125 levels ≥65 U/mL, the test has a sensitivity of 97% and a specificity of 78%.
In contrast to patients who present with advanced disease, patients with early ovarian cancers (stages I and II) are commonly curable with conventional therapy. Thus, effective screening procedures would improve the cure rate in this disease. Although pelvic examination can occasionally detect early disease, it is a relatively insensitive screening procedure. Transvaginal sonography is often useful, but significant false-positive results are noted, particularly in premenopausal women. In one study, 67 laparotomies were required to diagnose 1 primary ovarian cancer. Doppler flow imaging coupled with transvaginal ultrasound may improve accuracy and reduce the high rate of false positives.
CA-125 has been studied as a screening tool. Unfortunately, half of women with stages I and II ovarian cancer have CA-125 levels <65 U/mL. Attempts have been made to improve the sensitivity and specificity by combinations of procedures, commonly transvaginal ultrasound and CA-125 levels. In a screening study of 22,000 women, 42 had a positive screen and 11 had ovarian cancer (7 with advanced disease). In addition, eight women with a negative screen developed ovarian cancer. Thus, the false-positive rate would lead to a large number of unnecessary (i.e., negative) laparotomies if each positive screen resulted in a surgical exploration. The National Institutes of Health Consensus Conference recommended against screening for ovarian cancer among the general population without known risk factors for the disease. Although no evidence shows that screening saves lives, many physicians use annual pelvic examinations, transvaginal ultrasound, and CA-125 levels to screen women with a family history of ovarian cancer or breast/ovarian cancer syndromes.
In one study, proteomic spectra in the serum analyzed by an iterative searching algorithm were used to identify women with ovarian cancer. Preliminary studies have identified all 50 stage I patients with a sensitivity of 100%, a specificity of 95%, and a positive predictive value of 94%. The procedure can be automated, requires a pinprick of blood, and has many characteristics of an ideal screening test. However, difficulty in consistency of replicate samples, variation in spectroscopy equipment, and the tendency of the artificial intelligence algorithms to overfit the data makes conformation studies necessary before widespread application to screening is warranted.
Common epithelial tumors comprise most (85%) of the ovarian neoplasms. These may be benign (50%), frankly malignant (33%), or tumors of low malignant potential (16%) (tumors of borderline malignancy). Epithelial tumors of low malignant potential have the cytologic features of malignancy but do not invade the ovarian stroma. More than 75% of borderline malignancies present in early stage and generally occur in younger women. They have a much better natural history than their malignant counterpart.
There are five major subtypes of common epithelial tumors: serous (50%), mucinous (25%), endometroid (15%), clear cell (5%), and Brenner tumors (1%), the latter derived from the urothelium. Benign common epithelial tumors are almost always serous or mucinous and develop in women ages 20 to 60. They are frequently large (20 to 30 cm), bilateral, and cystic.
Malignant epithelial tumors are usually seen in women over 40. They present as solid masses, with areas of necrosis and hemorrhage. Masses >10 to 15 cm have usually already spread into the intraabdominal space. Spread eventually results in intraabdominal carcinomatosis, which leads to bowel and renal obstruction and cachexia.
Although most ovarian tumors are epithelial, two other important ovarian tumor types exist — stromal and germ cell tumors. These tumors are distinct in their cell of origin but also have different clinical presentations and natural histories and are often managed differently.
Metastasis to the ovary can occur from breast, colon, gastric, and pancreatic cancers, and the Krukenberg tumor was classically described as bilateral ovarian masses from metastatic mucin-secreting gastrointestinal cancers.
Staging and Prognostic Factors
Laparotomy is often the primary procedure used to establish the diagnosis. Less invasive studies useful in defining the extent of spread include chest x-rays, abdominal computed tomography scans, and abdominal and pelvic sonography. If the woman has specific gastrointestinal symptoms, a barium enema or gastrointestinal series can be performed. Symptoms of bladder or renal dysfunction can be evaluated by cystoscopy or intravenous pyelography.
A careful staging laparotomy will establish the stage and extent of disease and allow for the cytoreduction of tumor masses in patients with advanced disease. Proper laparotomy requires a vertical incision of sufficient length to ensure adequate examination of the abdominal contents. The presence, amount, and cytology of any ascites fluid should be noted. The primary tumor should be evaluated for rupture, excrescences, or dense adherence. Careful visual and manual inspection of the diaphragm and peritoneal surfaces is required. In addition to total abdominal hysterectomy and bilateral salpingo-oophorectomy, a partial omentectomy should be performed and the paracolic gutters inspected. Pelvic lymph nodes as well as para-aortic nodes in the region of the renal hilus should be biopsied. Since this surgical procedure defines stage, establishes prognosis, and determines the necessity for subsequent therapy, it should be performed by a surgeon with special expertise in ovarian cancer staging. Studies have shown that patients operated upon by gynecologic oncologists were properly staged 97% of the time, compared to 52% and 35% of cases staged by obstetricians/gynecologists and general surgeons, respectively. At the end of staging, 23% of women have stage I disease (cancer confined to the ovary or ovaries); 13% have stage II (disease confined to the true pelvis); 47% have stage III (disease spread into but confined to the abdomen); and 16% have stage IV disease (spread outside the pelvis and abdomen). The 5-year survival correlates with stage of disease: stage I - 90%, stage II - 70%, stage III - 15 to 20%, and stage IV - 1 to 5% (
Prognosis in ovarian cancer is dependent not only upon stage but on the extent of residual disease and histologic grade. Patients presenting with advanced disease but left without significant residual disease after surgery have a median survival of 39 months, compared to 17 months for those with suboptimal tumor resection.
Prognosis of epithelial tumors is also highly influenced by histologic grade but less so by histologic type. Although grading systems differ among pathologists, all grading systems show a better prognosis for well- or moderately differentiated tumors and a poorer prognosis for poorly differentiated histologies. Typical 5-year survivals for patients with all stages of disease are: well-differentiated - 88%, moderately differentiated - 58%, poorly differentiated - 27%.
The prognostic significance of pre- and postoperative CA-125 levels is uncertain. Serum levels generally reflect volume of disease, and high levels usually indicate unresectability and a poorer survival. Postoperative levels, if elevated, usually indicate residual disease. The rate of decline of CA-125 levels during initial therapy or the absolute level after one to three cycles of chemotherapy correlates with prognosis but is not sufficiently accurate to guide individual treatment decisions. Even when the CA-125 level falls to normal after surgery or chemotherapy, “second-look” laparotomy identifies residual disease in 60% of women.
Genetic and biologic factors may influence prognosis. Increased tumor levels of p53 are associated with a worse prognosis in advanced disease. Epidermal growth factor receptors in ovarian cancer are associated with a high risk of progression, but the increased expression of HER-2/neu has given conflicting prognostic results, and expression of Mdr-1 has not been of prognostic value. HER-2/neu is highly expressed in 20% of ovarian cancers, and responses have been seen to trastuzumab in this subset of patients.
The selection of therapy for patients with epithelial ovarian cancer depends upon the stage, extent of residual tumor, and histologic grade. In general, patients are considered in three separate treatment groups: (1) those with early (stages I and II) ovarian cancer and microscopic or no residual disease; (2) patients with advanced (stage III) disease but minimal residual tumor (<1 cm) after initial surgery; and (3) patients with bulky residual tumor and advanced (stage III or IV) disease.
Patients with stage I disease, no residual tumor, and well or moderately differentiated tumors need no adjuvant therapy after definitive surgery, and 5-year survival exceeds 95%. For all other patients with early disease and those stage I patients with poor prognosis histologic grade, adjuvant therapy is probably warranted, and single-agent cisplatin or platinum-containing drug combinations improve survival by 8% (82% vs 76%, p =. 08).
For the patients with advanced (stage III) disease but with limited or no residual disease after definitive cytoreductive surgery (about half of all stage III patients), the primary therapy is platinum-based combination chemotherapy. Approximately 70% of women respond to initial combination chemotherapy, and 40 to 50% have a complete regression of disease. Only about half of these patients are free of disease if surgically restaged. Although a variety of combinations are active, a randomized prospective trial of paclitaxel and cisplatin compared to paclitaxel and carboplatin in patients with optimally resected advanced disease demonstrated equivalent results (median time to progression 20.7 months vs 19.4 months, median survival 57.4 months vs 48.7 months) but with significantly reduced toxicity using carboplatin. This regimen of paclitaxel, 175 mg/m2 by 3-h infusion, and carboplatin, dosed to an AUC (area under the curve) of 7.5 is the treatment of choice for patients with previously untreated advanced-stage disease.
Patients with advanced disease (stages III and IV) and bulky residual tumor are generally treated with a paclitaxel-platinum combination regimen as well and, while the overall prognosis is poorer, 5-year survival may reach 10 to 15%.
Historically, patients who had an excellent initial response to chemotherapy and no clinical evidence of disease have had a second-look laparotomy. For patients with stage I ovarian cancer or for germ cell tumors, the operation rarely detects residual tumor and has been largely abandoned. Even for those with stages II and III epithelial tumors, the second-look surgical procedure itself does not prolong overall survival. Its routine use cannot be recommended. Maintenance therapy (12 cycles of paclitaxel every 28 days) may extend progression-free survival among patients who achieve a complete response; an effect on overall survival has not yet been shown.
Patients with advanced disease whose disease recurs after initial treatment are usually not curable but may benefit significantly from limited surgery to relieve intestinal obstruction, localized radiation therapy to relieve pressure or pain from mass lesions or metastasis, or palliative chemotherapy. The selection of chemotherapy for palliation depends upon the initial regimen and evidence of drug resistance. Patients who have a complete regression of disease that lasts ≥6 months often respond to reinduction with the same agents. Patients relapsing within the first 6 months of initial therapy rarely do. Chemotherapeutic agents with >15% response rates in patients relapsing after initial combination chemotherapy include gemcitabine, topotecan, liposomal doxorubicin, and vinorelbine. Intraperitoneal chemotherapy (usually cisplatin) may be used if a small residual volume (<1 cm3) of tumor exists. Progestational agents and antiestrogens produce responses in 5 to 15% of patients and have minimal side effects.
Patients with tumor of low malignant potential, even with advanced-stage disease, have longer survivals when managed with surgery alone. The added value of radiation and chemotherapy has not been shown.
Revision date: July 5, 2011
Last revised: by Andrew G. Epstein, M.D.