Pathology of Ovarian Cancer

Jaime Prat, M.D., F.R.C.Path.

Professor of Pathology - Autonomous University of Barcelona.

Department of Pathology - Hospital de la Santa Creu i Sant Pau, Barcelona, Spain

Ovarian cancer is the fifth most common malignant tumor among women in the United States, with an annual incidence of 22,000 new cases. It affects predominantly postmenopausal women and is responsible for about 13,300 deaths per year. This high mortality rate is attributed to the lack of symptoms in most patients with early stages of disease. In approximately 70% of the patients the tumor has spread outside of the pelvis at the time of presentation. Extensive surgery is often insufficient to eliminate completely the intraabdominal tumor and response to chemotherapy is only partial in many of these patients1.

Etiology and pathogenesis

Classification

Surface epithelial-stromal tumors

Sex cord-stromal tumors

Germ cell tumors

Gonadoblastoma

Metastatic tumors

Pathologic factors affectin prognosis

FIGO staging of ovarian cancer

References

Return to the main page

 

Etiology and pathogenesis

The malignant transformation of the ovarian surface epithelium has been explained by the incessant ovulation theory, according to which ovulation disrupts the surface epithelium and is followed by reparative proliferation. The newly formed epithelial cells are capable of proliferating in response to additional trauma caused by subsequent ovulations. Continuous reparation may trigger mutations and activation of oncogens or elimination of tumor suppressor genes2.

The epidemiological data supporting this theory include the increased risk (1.5 to 3.2) associated with nulliparity, the decreased risk in women using oral contraceptives (0.4 to 0.7) as well as the increased incidence of ovarian cancer in patients who have received drugs that induce ovulation, but have not become pregnant.

Most ovarian carcinomas apparently arise from epithelial inclusion cysts which result from downgrowth of the surface epithelium following repair. It has also been postulated that elevated gonadotropin levels may contribute to malignant transformation. Oral contraceptives cause a reduction in pituitary gonadotropins.

The principal risk factor, however, is a family history of ovarian cancer, particularly when two or more first-degree relatives have been affected3. Three hereditary familial ovarian cancer syndromes have been described: the ovary specific syndrome; the breast-ovary syndrome; and the Lynch II syndrome, in which cancer may develop in the ovary, endometrium and colon (in absence of polyposis coli). Familial ovarian cancer accounts for 5-10% of the total. The responsible gene for the breast-ovary syndrome (BRCA 1) is located in the long arm of chromosome 17 at locus 17q12-q234.

From a histogenetic viewpoint, ovarian cancer may develop "de novo" or arise from pre-existing benign epithelial tumors5. In favor of the latter hypothesis is, a) the higher average age of the patients with carcinomas (approximately 12 years) as compared to that of patients with benign epithelial tumors; b) the finding of a benign epithelial component in a high number of ovarian carcinomas, specially mucinous carcinomas, which often exhibit focal, abrupt transition from benign to malignant epithelium.

The role of oncogens and tumor suppressor genes in ovarian carcinogenesis is not well known6. Amplification or overexpression of c-erb-B-2 has been found in approximately 30% of ovarian cancers and correlate with low survival7 ras mutations are frequent in mucinous ovarian tumors8-10. About 50-80% of ovarian carcinomas show overexpression of the tumor suppressor gene p53, which correlates with aneuploidy, high grade, advanced stage, and unfavorable prognosis11, 12.

Determinations of the serum tumor marker CA-125 and pelvic ultrasonography are highly specific screening methods.

Classification

Because of their remarkable diversity, ovarian tumors may be bewildering. Natural history and response to treatment vary considerably from one tumor group to another. The best therapeutic approach including chemotherapy and radiotherapy may be highly specific for a single type of neoplasm; accordingly, accurate histologic diagnosis is often a critical factor in achieving an optimum treatment response.

The following histogenetic classification is a simplified version of the one proposed by the ISGYP in 1993 and subsequently accepted by the WHO:

  1. Surface epithelial-stromal tumors
    1. Serous tumors
      1. Benign (cystadenoma)
      2. Of borderline malignancy
      3. Malignant (serous cystadenocarcinoma)
    2. Mucinous tumors, endocervical-like and intestinal-type
      1. Benign
      2. Of borderline malignancy
      3. Malignant
    3. Endometrioid tumors
      1. Benign
      2. Of borderline malignancy
      3. Malignant
      4. Epithelial-stromal
        1. Adenosarcoma
        2. Mesodermal (müllerian) mixed tumor
    4. Clear cell tumors
      1. Benign
      2. Of borderline malignancy
      3. Malignant
    5. Transitional cell tumors
      1. Brenner tumor
      2. Brenner tumor of borderline malignancy
      3. Malignant Brenner tumor
      4. Transitional cell carcinoma (non-Brenner-type)
    6. Undifferentiated carcinoma
  2. Sex cord-stromal tumors
    1. Granulosa-stromal cell tumors
      1. Granulosa cell tumors
      2. Tumors of the thecoma-fibroma group
    2. Sertoli-stromal cell tumors (androblastomas)
    3. Sex-cord tumor with annular tubules
    4. Gynandroblastoma
    5. Steroid (lipid) cell tumors
  3. Germ cell tumors
    1. Dysgerminoma
    2. Yolk sac tumor (endodermal sinus tumor)
    3. Teratoma
      1. Immature
      2. Mature (adult)
        1. Solid
        2. Cystic (dermoid cyst)
      3. Monodermal (e.g., struma ovarii, carcinoid)
    4. Mixed germ cell tumors
  4. Gonadoblastoma
  5. Tumors not specific to ovary
  6. Unclassified
  7. Metastatic tumors

The relative frequency with which different types of ovarian neoplasms occur is summarized in Table I. The proportions indicated are approximate and represent a synthesis of numerous reports.

Certain clinical or gross features may provide important diagnostic clues. One of the most important clinical features is the age of the patient. For example, within the surface epithelial category, borderline tumors are seen often in women in their 30s and less frequently in younger women, whereas invasive tumors in this category are rare in women under 40 years of age. On the other hand, germ cell tumors are almost never found in women over 50 years of age.

The stage and laterality of ovarian tumors also indicate their nature. For example, tumors in the sex cord-stromal category are almost always confined to a single ovary. On the other hand, approximately 65% of the metastatic tumors are bilateral.

Surface epithelial-stromal tumors

(common epithelial tumors)

Surface epithelial tumors form the most common group of ovarian neoplasms and include the majority (90%) of ovarian carcinomas (Table I). They originate from the celomic mesothelium that covers the ovary, which, after neoplastic transformation, seems to retain the capacity to recapitulate the epithelial components of the müllerian ducts. For example, the epithelium of serous tumors resembles that lining the tube whereas the cells that line mucinous cystadenomas resemble those of the endocervical mucosa. These tumors usually have a prominent cystic component with single or multiple loculations, a variable amount of fibrous stroma, and an epithelial lining that often is thrown into papillary tufts.

On the basis of histological features and clinical behavior, these tumors are divided into benign, borderline (low malignant potential) (LMP), and malignant categories. Benign cystic tumors are lined by a single layer of well-oriented columnar epithelial cells; papillary projections, if present, are supported by fibrovascular stromal stalks and covered by the same type of epithelium. Obviously malignant tumors have an anaplastic epithelial component that invades the underlying stroma in addition to forming the epithelial lining. The epithelial cells are usually several layers thick and have anaplastic nuclei, with a loss of polarity. The prognosis is poor; about 15% of patients survive 5 years.

Tumors in the borderline (LMP) category are characterized by epithelial proliferation greater than that seen in benign tumors of the same cell type but an absence of "destructive" invasion of the stroma13-16. Despite their lack of an invasive potential within the ovary, these tumors can implant on peritoneal surfaces and the implants often invade the underlying tissue; rarely, these tumors spread via lymphatics and, exceptionally, through blood vessels. The diagnosis, however, is based on pathological examination of the primary ovarian tumor itself without knowledge of whether spread beyond the ovary has taken place.

Although the behavior of borderline tumors is unpredictable in individual cases, as a group they have a much better prognosis than their malignant counterparts. Approximately 70% are Stage I, with spread within the pelvis (Stage II) or upper abdomen (Stage III) being present in the remainder. One-third of the tumors are bilateral.

The 5-year survival for borderline and malignant tumors confined to the ovary is 100% and 70%, whereas the 5-year survival for the same tumors involving the peritoneum is approximately 90% and 25%, respectively. Recurrences of borderline tumors typically appear after several years, if at all; the minority of tumors that have a malignant course tend to progress slowly.

Borderline tumors are best treated by removal of the uterus and adnexa as well as indicated biopsies for complete staging. However, in young women with stage I tumors, unilateral adnexectomy may be acceptable. The risk that another tumor will develop in a normal-appearing ovary in these circumstances is about 15% in a 5 to 7 year period17. Furthermore, the tumors that do recur consistently maintain their borderline clinical and pathologic characteristics, including an excellent prognosis.

Various tumor-associated antigens may circulate in the blood of women with borderline and malignant tumors of common epithelial type. CA-125 antigen is most commonly associated with serous and endometrioid tumors; mucinous tumors often produce carcinoembryonic antigen (CEA), as colon cancers do. Monitoring the appropriate antigen titers is very useful in identification of occult metastases18, monitoring of therapeutic response, and detection of asymptomatic recurrence at an early stage.

Serous tumors

Benign serous cysts and cystadenomas may form single or multiple loculations, lined by low columnar epithelium, which is sometimes ciliated and resembles tubal epithelium. The cyst fluid is watery or viscous and contains a variety of mucins; however, the epithelial cells that secrete the fluid do not have the characteristic vacuolated appearance of mucinous epithelium. Papillary processes are common and may be complex. The epithelial component of serous tumors, unlike other neoplasms of surface epithelium, may appear on the external surfaces. It is common to find tiny round laminated calcific concretions called "psammoma bodies" in the stroma of the papillary processes. Papillary and cystic tumors with a prominent or abundant fibrous stromal component are designated papillary adenofibromas and cystadenofibromas. Borderline serous tumors are often multilocular and have a more complex papillary pattern; fine papillae, closely packed, may resemble solid epithelial proliferation. Variable degrees of cell stratification, nuclear atypia, and mitotic activity are present.

Otherwise typical serous borderline tumors may contain foci of microinvasion19, 20. Such tumors have a prognosis similar to that of serous borderline tumors without this feature and conservation of the contralateral ovary and uterus may be acceptable.

Serous borderline tumors are associated in from about 30 percent of the cases with peritoneal implants which may vary greatly in their histologic appearance16, 21. It should be remembered that in these cases the designation of the tumor depends on the appearance of the primary neoplasm. The peritoneal implants have been classified into non-invasive and invasive categories with the former being further subdivided into epithelial and desmoplastic subtypes22. In the epithelial subtype of non-invasive implant papillary proliferations of atypical serous cells are present on the surface of the peritoneum. In contrast, the desmoplastic subtype of non-invasive implant is characterized by a predominant stromal reaction to the tumor which is layered upon serosal surfaces. Invasive implants are characterized by an irregular infiltration of a normal tissue, such as the omentum. They resemble histologically a serous adenocarcinoma; marked cytologic atypia may be present. Three recent studies have shown that the presence of invasion in implants is strongly related to progression of the disease16, 21, 22. Aneuploidy of the invasive implants seems to have an unfavorable prognostic significance16.

Tumors with a morphology identical to that of serous borderline tumors may arise as primary neoplasm of the peritoneum with no, or only minimal, involvement of the ovarian surfaces23. A tumor that shares many features with the serous borderline tumor is the rare serous psammocarcinoma24. This diagnosis is made when the following microscopic features are present: 1. Destructive invasion of ovarian stroma or vascular space invasion, or deep invasion of intraperitoneal viscera, 2. No more than moderate nuclear atypia, 3. No significant solid epithelial proliferation, 4. At least 75% of the papillae or nests are associated with or completely replaced by psammoma bodies. The clinical behavior of these tumors more closely resembled that of a borderline serous tumor than a serous carcinoma.

Mucinous tumors

Mucinous tumors are typically large unilocular or multilocular cystic masses. The epithelium that lines the cysts is composed of tall columnar cells with basal nuclei and prominent mucin vacuoles; it may resemble endocervical mucosa (endocervical-type). Most frequently, the epithelium appears like intestinal epithelium, including goblet cells, argentaffin cells, argyrophil cells and even Paneth's cells (intestinal-type). Rarely a mucinous intestinal tumor has produced enough gastrin to cause the Zollinger-Ellison syndrome. Since about 5% of mucinous tumors are associated with cystic teratomas, it has been suggested that some, at least, originate from germ cells; an intestinal metaplasia of these exotic cell types seems more likely. A larger number, however, may have endometrioid or serous elements, which supports their classification with surface epithelial tumors.

Mucinous borderline tumors are almost as common as serous borderline tumors and account for 40-50% of all mucinous malignant tumors25 and for 71% of those that are Stage I13. Approximately 85% of the cases are of the intestinal type26.

It is often difficult to decide whether a mucinous tumor is a borderline or a low-grade carcinoma. The main reason for this is the difficulty of determining stromal invasion in many cases of carcinoma since the stroma, instead of being desmoplastic, may resemble ovarian stroma and is often luteinized. The criteria for the diagnosis of carcinoma include: height of 4 or more epithelial cells26, finger-like projections of solid cellular masses without connective tissue supports or cribriform pattern13. Other investigators rely mostly on the nuclear features of the tumor cells25. Recently, separation of intestinal mucinous borderline tumors from well differentiated carcinomas has been achieved by quantitative nuclear morphologic analysis27.

Extra-ovarian spread of the intestinal-mucinous borderline tumors typically appears as mucinous ascites ("pseudomyxoma peritonei"). Patients with an ovarian borderline mucinous tumor and pseudomyxoma peritonei often prove to have a similar tumor of the appendix28, 29. The synchronous presentation, the histologic similarity, the frequency of bilaterality of the ovarian tumors, and the predominance of right-sided ovarian involvement, all point towards the probable metastatic nature of the ovarian tumors in such cases28, 29.

Borderline mucinous tumors of endocervical type have a distinctly better prognosis than the more common mucinous tumors of intestinal type, even in the presence of extraovarian spread30. They are frequently associated with endometriosis30.

Mucinous cystic tumors, benign, borderline or carcinomas, may contain one or more solid mural nodules composed occasionally of anaplastic carcinomas31, rarely of sarcoma32 and more often of a sarcoma-like proliferation which has been associated with an uneventful follow-up in all cases33-35.

Endometrioid tumors

Endometrioid carcinomas are so named because the histologic pattern closely resembles that of uterine endometrial adenocarcinomas. The distinction is most easily made in well-differentiated carcinomas. Less-differentiated lesions may have a typical endometrioid appearance only in focal areas or foci of squamous metaplasia as the only clues to their identity. They probably comprise between 15% and 20% of ovarian cancers. The benign counterpart is probably represented by cystic endometriosis of the ovary. Criteria for the diagnosis of a borderline endometrioid tumor are controversial36-38, though such lesions are rare.

Endometrioid carcinomas are often partly cystic with prominent solid areas; the cyst fluid is brown or bloody. Association with endometriosis is demonstrable in about one third of the cases, but its presence is not required for the diagnosis. Endometriosis is even more frequently encountered together with clear cells tumors. Unusual histologic appearances of endometrioid carcinomas that may cause confusion with sex cord-stromal tumors have been described39. Distinguishing endometrioid carcinoma from an occasional metastatic adenocarcinoma from the large intestine may be difficult.

The prognosis for well-differentiated carcinomas is good; about 60% of patients survive for 5 years, compared to 23% survival for poorly differentiated carcinomas.

In about one third of the patients there is a coexistent adenocarcinoma of the endometrium. It is generally accepted that both lesions are separate primary cancers because the survival rate in the presence of endometrial involvement is not appreciably lower40. Furthermore, the absence of extragenital tumor and the presence of endometrial hyperplasia, both favor a multifocal process. In contrast, simultaneous uterine and ovarian involvement by common epithelial carcinomas of other types such as papillary serous or clear cell carcinomas is usually associated poor prognosis40. In some cases, DNA flow cytometry may help distinguishing metastatic from synchronous tumors41.

Other endometrioid neoplasms such as stromal sarcoma42, 43, malignant mixed müllerian tumor44, and adenosarcoma have been reported rarely as primary tumors of the ovary. The histologic features and prognosis do not differ significantly from those of similar lesions occurring in the endometrium.

Clear cell tumors

The gross appearance of clear cell tumors is often a combination of solid and cystic components much like that of endometrioid carcinoma. The cyst is usually unilocular; the fluid is commonly brown, and the solid areas form nodular masses that protrude into the lumen.

Microscopically are characterized by masses of large epithelial cells with abundant clear cytoplasm, supported by delicate hyalinized stroma. The cytoplasm contains abundant glycogen. The most common pattern, however, is that of small tubules and cysts lined by a single layer of large hobnail cells; the cell nuclei are bulbous and protrude into the lumens of the tubules. Hyaline bodies are present in 25% of the cases45. The cytologic features, absence of lymphocytes and lower or negative serum values of alfa-fetoprotein, help to distinguish clear cell carcinoma from dysgerminoma and yolk sac tumor.

Less than 10% of clear cell tumors are bilateral. Benign and borderline clear cell tumors occur rarely, mainly as adenofibromas. Association with endometriosis is six times as great as with ovarian carcinoma in general46. Clear cell carcinomas of the ovary and identical neoplasms of the endometrium, cervix, and vagina are closely related to tumors of müllerian origin, particularly endometrioid tumors47. Endometrioid and clear cell components may occur together in ovarian and endometrial carcinomas, and clear cell tumors have been shown to arise in patients with endometriosis47. Survival rates are in an intermediate range; 37% of the patients survive 5 years.

Transitional cell tumors

The recently recognized transitional cell carcinomas can be divided into those in which there is malignant transformation of a pre-existing Brenner tumor and those in which the tumor is of a surface epithelial origin48, 49.

Brenner tumors account for 2 to 3% of all ovarian tumors50 and less than 2% of them are borderline (proliferative) or malignant50-52. Benign Brenner tumors are typically small and mostly solid; 10 to 25% appear as small, firm nodules in the wall of a mucinous cystadenoma. Borderline tumors are predominantly cystic with papillomatous masses protruding into the locules.

Microscopically, the benign tumor is composed of round nests of epithelial cells lying within an abundant fibrous stroma. The nests may be solid or have a central lumen that contains dense eosinophilic material or mucin. The tumor cells are polygonal and have clear cytoplasm and grooved nuclei. The cells lining the lumens often contain mucin. The stromal component may be focally hyalinized, and calcification is common.

The epithelial component of the borderline tumors resembles a Grade I papillary carcinoma of the urinary bladder. No invasion of the stroma is observed, and a benign component is also found in most cases.

In malignant Brenner tumors, the epithelial component has, at least in part, the typical features of higher grade transitional cell carcinoma or squamous cell carcinoma.

Transitional cell carcinoma of the ovary is rare48; it apparently responds much better to chemotherapy than most other surface epithelial carcinomas at all stages. A common problem is the distinction of these tumors from poorly differentiated carcinomas which typically exhibit central necrosis.

Mixed forms

Surface epithelial tumors often include a combination of the foregoing types. Unless two or more components are prominent, the tumor is classified after the predominant cell type.

Undifferentiated carcinoma

Undifferentiated carcinoma comprises about 6% of ovarian carcinomas. Clinically resembles serous carcinoma and about 50% are bilateral. Most are stage III tumors at presentation. Many are composed of small cells of nearly uniform size. Mitoses are numerous; nuclei are pleomorphic. Occasionally, undifferentiated carcinomas resemble superficially granulosa cell tumors. Because of the marked differences in prognosis, an accurate histologic diagnosis is important46.

A distinctive type of undifferentiated carcinoma that is associated with hypercalcemia in approximately two-thirds of the cases occurs predominantly in young women (average 22 years). It has been designated "small cell carcinoma"53, 54. The histogenesis of this aggressive tumor remains unknown54, 55. Hypercalcemia has been attributed to tumor production of parathyroid hormone-related substance56. All tumors have been diploid.

Sex cord-stromal tumors

The sex cord-stromal tumors arise from ovarian stromal cells specialized in steroid hormone production; they account for most functioning tumors with endocrine manifestations57. The designation of sex cord-stromal tumors if favored because the assumption that the embryonic sex cords and their derivatives are mesenchymal (stromal) derivatives (rather than celomic epithelium) remains unproved.

Granulosa cell tumors

Granulosa cell tumors are mostly estrogenic (75%) but rare cases, usually large cystic tumors, are associated with androgenic manifestations. The tumors are cystic and solid and characteristically contain blood clots.

Granulosa cells have uniform, oval, rounded or angulated pale nuclei, with a cleft in the nuclear membrane. Mitotic activity is not conspicuous. The cells are haphazardly distributed in masses with little intervening stroma. Characteristic rosette-like structures, or Call-Exner bodies, are almost always present; they are rounded masses of eosinophilic material surrounded by a circular row of typical granulosa cells and resemble the normal graafian follicle. The nuclei tend to lie adjacent to the inner rim of the space.

A variety of microscopic patterns occur; microfollicular and macrofollicular tumors resemble clusters of small or large graafian follicles. The descriptive terms trabecular, insular, and gyriform, are often applied without significant clinical correlations. The diffuse or "sarcomatoid" variety, characterized by somewhat spindle-shaped cells, may have a more aggressive behavior.

The juvenile type, chiefly prepubertal, has distinctive microscopic features such as rudimentary follicles, rarely grooved hyperchromatic nuclei and striking luteinization of both granulosa and thecal components58.

The most consistent indicator of aggressive behavior has been the presence of metastases or invasion of structures outside the ovary at the time of diagnosis. Also unfavorable but less significant factors are large tumor size, increasing age, poorly differentiated tumors with a high mitotic rate, and tumor rupture58, 59-61.

Juvenile granulosa cell tumors may produce precocious sexual development in prepuberal girls. In adult women, however, the most common symptom is uterine bleeding as a result of endometrial hyperplasia. Well-differentiated endometrial adenocarcinoma occurs in 10% to 25% of the cases in postmenopausal women. These adenocarcinomas usually have a remarkably good prognosis, possibly because most of them are highly estrogen dependent and therefore fail to progress when the source of estrogen is withdrawn.

Less than 5% of granulosa cell tumors are bilateral. Over 90% of the patients survived for 10 years, and some had residual tumor. Recurrences continue to appear as late as 30 years postoperatively. It is reasonable to conserve the opposite ovary and uterus of a young woman with a small tumor confined to one ovary.

Fibroma-Fibrosarcoma

Fibrous tumors of the ovary without clinical or morphologic evidence of endocrine activity are relatively common and account for 6% of all ovarian tumors. Densely collagenized fibrous tissue forms a monotonous histologic pattern occasionally broken by areas of calcification. In rare cases, large fibromas may be associated with benign ascites and pleural effusion, which disappear when the tumor is resected (Meig's syndrome). Cellular fibromas with only slight mitotic activity (3 or less MFs/10 HPFs) and mild to moderate nuclear atypia are almost always benign unless they are adherent or ruptured62. Fibrosarcomas typically exhibit higher mitotic rate and marked nuclear pleomorphism; they are extremely rare62.

Sertoli-Leydig cell tumors (Sertoli-stromal cell tumors)

These rare tumors occur in all age groups but they peak during reproductive years (average age 25 years). They are bilateral in only 3% of the cases. Although Sertoli-Leydig cell tumors often produce androgens and masculinize the patient, many are nonfunctioning, and some even have estrogenic effects63. Testosterone and a variety of androgenic precursors may be secreted in variable proportions. Their histologic appearance resembles that of the developing testis.

Five histologic categories are distinguished: a) Well-differentiated tumors that form tubular structures composed of Sertoli cells, separated by a fibrous stroma, and intermixed with nests of large round Leydig cells. b) Tumors of intermediate differentiation in which large eosinophilic Leydig cells are separated by a spindly stroma; abortive tubule formation resembles early sex cords of the embryonic testis. c) A less differentiated sarcomatoid variant, composed of spindle cells with vague trabecular arrangement. Leydig cells may or may not be present. d) Twenty per cent of the tumors contain heterologous elements such as neoplastic mucinous glands, cartilage, and skeletal muscle64. A fifth histologic subtype that simulates the rete of the ovary or testis has been designated retiform Sertoli-Leydig cell tumor65. About half the well-differentiated tumors, three fourths of intermediate tumors, and all sarcomatoid tumors are androgenic. Pure Sertoli cell tumors occur in younger women and children; most cause hyperestrinism, including isosexual precocious puberty in children, but few have caused virilization.

Almost all Sertoli-Leydig cell tumors are benign, with the rare malignant cases falling into the category of poorly differentiated tumors and those with mesenchymal heterologous elements64. Malignant behavior takes the form of intraabdominal spread, ordinarily without distant metastases.

Steroid (lipid) cell tumors

Steroid cell tumors are a distinctive group of neoplasms that occur in the form of soft yellow or yellow-brown nodules. These tumors are divided into: Leydig stromal luteomas, (hilus) cell tumors, and steroid cell tumors not otherwise specified (NOS)66. The tumor cells are large, rounded or polyhedral cells resembling lutein cells, Leydig cells and adrenocortical cells. Although most of them contain abundant intracellular fat, some do not. Ultrastructure is consistent with ovarian stromal origin; cytoplasmic organelles resemble those of steroid-secreting cells. Stromal luteomas are located within the ovarian stroma, lack crystalloids of Reinke and are rarely androgenic. In contrast, Leydig (hilus) cell tumors are often androgenic and typically contain crystalloids of Reinke. Both tumors occur in postmenopausal women and are benign. Steroid cell tumors NOS occur in younger patients than the other two forms and may be malignant (25-40%). When functioning, these tumors are typically androgenic66. Malignant behavior is indicated by the presence of 2 MFs/HPF, necrosis, large size, and nuclear atypia66. A few have caused Cushing's syndrome.

Germ cell tumors

Germ cell tumors represent caricatures of normal embryogenesis. Like the embryo, they are capable of developing complex and highly differentiated tissues. The malignant potential of germ cell tumors reflects their degree of differentiation; however, maturation does not exclude malignancy, since mature tissues may eventually undergo malignant change.

The traditional concepts on the histogenesis of germ cell tumors have been recently challenged at the light of recent investigations67. According to these data, histogenesis may vary depending upon the site of origin of the tumor and the age of the patient. Although germ cell tumors of the ovary are morphologically similar to testicular germ cell tumors, they may not necessarily have an identical histogenesis67. An origin from pre-existing somatic neoplasms is supported by the occasional occurrence of extragonadal germ cell tumors, some of them in the female genital tract68, 69. Recent evidence has suggested that germinomas are not end stage tumors, as previously thought, but neoplasms capable of differentiation70, 71.

Morphologic overlapping between early developmental forms of germ cell tumors, such as seminoma, embryonal carcinoma and yolk sac tumor, supports the view that these tumors constitute a closely related group of neoplasms capable of extra- or intraembryonal differentiations. These new concepts have been represented in a tridimensional tetrahedron model of interrelationships between the different components.

Dysgerminoma

The most common malignant germ cell tumor of the ovary72 is morphologically identical to testicular seminoma. Grossly, these neoplasms are large, solid, encapsulated masses of soft, gray-white tissue. They are composed of large rounded or polygonal cells resembling primordial germ cells73. The tumor cells are distributed in large masses separated by delicate fibrous trabecula. The fibrous stroma is almost always infiltrated by lymphocytes and may contain sarcoid-like granulomas. The presence of calcification suggests an underlying gonadoblastoma. About 80% are stage I and 10% are grossly bilateral. The opposite ovary may contain microscopic foci of dysgerminoma even when it appears grossly normal (10%); if it is to be preserved, biopsy with frozen-section examination is desirable74, 75.

Dysgerminoma is an extremely radiosensitive tumor but, currently, cisplatin-based chemotherapy is the treatment of choice. The tumor is curable even in the presence of metastases. Thus, since most patients are young, unilateral salpingo-oophorectomy is desirable and sufficient when the opposite ovary is normal. The 5-year survival rate is between 70% and 90%. For pure stage I dysgerminoma, the 5-year survival rate approaches 95% for patients treated with unilateral salpingo-oophorectomy. The occasional dysgerminoma that contains syncytiotrophoblastic giant cells (about 3%) does not behave more aggressively than the ordinary dysgerminoma76. These tumors may produce human chorionic gonadotropin (hCG) which can serve as a tumor marker.

Yolk sac tumor (YST)

YST are almost as common as dysgerminoma under the age of 20 years. They consistently produce alpha-fetoprotein, which can be demonstrated in tissue sections by immunohistochemical techniques and in the patient's serum. This substance, which is normally produced in the yolk sac of the developing embryo, may serve as a tumor marker in evaluating the course of the patient after treatment77. The gross appearance may be similar to that of dysgerminoma except for the more extensive yellow and red areas of hemorrhage and necrosis and the common presence of cysts.

The typical microscopic feature of this rare neoplasm is the presence of isolated papillary projections with a central blood vessel and peripheral sleeve of embryonic epithelial cells. Cross sections of this structure once were erroneously compared with immature glomeruli. In fact, they closely resemble invaginations of yolk sac endoderm, as seen best in the rat placenta, forming the endodermal sinuses of Duval. Another distinctive feature is the presence of periodic acid-Schiff-positive, diastase-resistant hyaline globules partly composed of alpha-fetoprotein and alpha-1-antitrypsin. Some tumors contain multiple gland spaces with hourglass constrictions resembling yolk sac vesicles, a pattern designated as polyvesicular vitelline tumor78. Hepatic and enteric (glandular differentiation) may predominate in rare tumors79-81.

Before the use of combination chemotherapy, the prognosis was very poor; remissions have occurred in some patients treated postoperatively with multiple chemotherapeutic agents. For patients with stage I tumors, survival rates has approached 80%.

Embryonal carcinoma

Embryonal carcinoma is an uncommon ovarian germ cell tumor that has been confused with YST, which it resembles82. From a histogenetic viewpoint, embryonal carcinoma has been considered a pluripotent stem cell tumor capable of differentiating along different pathways. The patients are young, present with an abdominal mass, and consistently have positive pregnancy test as a result of hCG production by the tumor82. Premenarchal girls undergo precocious puberty. Histologically, it resembles testicular embryonal carcinoma: large primitive anaplastic cells form solid masses interspersed with glandlike cleft and scattered giant cells. Multinucleated giant cells of syncytiotrophoblastic type immunoreactive for hCG are common. Similarly, mononuclear embryonal cells stain for alpha-fetoprotein. Both substances can be used as tumor markers. Although these tumors are highly malignant, chemotherapy has resulted in some long-term survivors. Most tumors are unilateral and salpingo-oophorectomy is the recommended surgical procedure.

Polyembryoma

Some germ cell tumors contain large numbers of embryoid bodies, closely resembling early embryos, typically distributed in a primitive mesenchymal stroma83.

Choriocarcinoma

Nongestational primary choriocarcinoma of the ovary is rare and malignant. The histologic picture and clinical correlations are similar to those of gestational choriocarcinoma, except that the remarkable response to chemotherapy usually does not occur. Choriocarcinomatous differentiation, usually interpreted as "retrograde differentiation" has been reported in ovarian carcinomas69.

Teratomas

Teratomas are composed of recognizable tissues of ectodermal, mesodermal, and endodermal origin, in any combination. They are common and usually benign and inert but rarely produce remarkably bizarre and varied syndromes, reflecting the diverse potentials of the germ cells.

Benign cystic teratoma (dermoid cyst)

Cystic teratomas are composed of a great variety of mature somatic tissues. Most cysts are unilocular, and the tissue that forms the lining is usually skin. The desquamated keratin and secretions, notably from sebaceous glands, accumulate with masses of hair to fill the lumen of the cyst. This mixture is liquid at body temperature but solidifies when chilled. Other common components are salivary gland; bronchus; fat; smooth muscle; cartilage; bone; neural tissue, including ganglia, glia, and choroid plexus; retina; pancreas; thyroid; and teeth. Characteristically, a protuberance from the inner surface is the locus of growth of most of the hair and teeth. Uncommon tissues are skeletal and cardiac muscle, kidney, and liver.

Bilaterality occurs in 8% to 15% of cases. Cystic teratomas comprise 20% of all ovarian tumors in adults and 50% of all ovarian tumors in children. Most patients are between 20 and 40 years, but the tumors occur at all ages.

The pathogenesis of teratomas has always excited speculation because of their exotic composition. Cytogenetic analysis using chromosome-banding techniques, has indicated that ovarian teratomas are parthenogenetic tumors that must originate from a single germ cell after its first miotic division84.

The most serious complication of a mature cyst teratoma is malignant change. Almost any component may become malignant, but squamous cell carcinoma accounts for 90% of the cases. Sweat gland carcinoma, thyroid carcinoma, malignant melanoma, and various sarcomas, including osteosarcoma, occur rarely.

Solid teratomas (teratomas with abundant solid tissue and relatively small cysts) are nearly all malignant (as discussed subsequently), but a few benign solid teratomas have been reported85.

Immature teratoma (malignant teratoma)

Malignant teratoma, the third most common ovarian germ cell tumor, usually presents in young adults and children (median age 18 years) as a unilateral solid mass with a heterogeneous on cut surface. Grossly evident dermoid cysts are present in 25% of the cases86. Histology is also extremely variable; the immature tissue is predominantly neuroectodermal tissue in the form of rosettes and tubules and cellular foci of glia with numerous mitoses. Islands of immature cartilage, bone, and glandular structures are distributed through a poorly differentiated stroma. Bilateral involvement in patients with stage I malignant teratoma is extremely rare, but a dermoid cyst is present in the contralateral ovary in 10% of the cases86.

The relative amount of primitive neuroepithelial tissue is an important factor in grading and determining the prognosis. Grade 1 has been applied to tumors with rare foci of immature neural tissue occupying less than one low power field (LPF) per slide, grade 2 to tumors with moderate quantities of immature neural tissue filling more than one but 3 or fewer LPF per slide, and grade 3 to tumors with large amounts of immature neural tissue occupying 4 or more LPF per slide. Patients with relatively mature (grade 1) malignant teratomas have a good prognosis, whereas for those with immature teratomas (grade 3) the prognosis is poor87-90. Alpha-fetoprotein, once believed to be a marker exclusive for YST, is also secreted by the immature endodermal tissues in immature teratomas.

About one third of immature teratomas are found in Stages II or III at the time of laparotomy and are usually associated with peritoneal implants. The neural component, even in the implants, may mature, leaving well-differentiated glial vestiges on the peritoneal surfaces. Mature glial implants are innocuous and not an indication for radical treatment87. Therefore the grading of metastases, once they have occurred, is also prognostically important88. In contrast to the dismal results obtained by surgery or radiation in the past, combination chemotherapy has produced remarkable effect resulting in over 70% survival90. Chemotherapy induces maturation of the implants.

Specialized teratomas

Rare teratomas composed exclusively of thyroid tissue are usually benign but may function and even cause thyrotoxicosis. Carcinoid tumors with the insular pattern typical of midgut derivatives91 and trabecular carcinoids of the foregut and hindgut type also occur as primary ovarian tumors92. The latter type may be mixed with thyroid tissue (strumal carcinoid)93. Both are nearly always unilateral and benign; insular carcinoids, especially if large, may cause the carcinoid syndrome. On the other hand, intestinal carcinoids metastatic to the ovary are usually bilateral and have a poor prognosis. It is especially important to distinguish them from granulosa cell tumors and Sertoli-Leydig cell tumors, as well as from primary ovarian carcinoids94. Several primitive neuroectodermal tumors, similar to neoplasms of the central nervous system (medulloepithelioma, neuroblastoma, ependymoblastoma and medulloblastoma) have been described in the ovary. Prognosis is poor95, 96.

Mixed forms

Germ cell tumors occur in various combinations.

Malignant mixed germ cell tumors (stage I) have a poor prognosis if more than one third of the tumor consists of endodermal sinus tumor, choriocarcinoma, or grade 3 teratoma. Tumors that contain less than one third of these components or contain combinations of dysgerminoma, embryonal carcinoma, or grade 1 or 2 teratoma have a good prognosis. Patients with tumors less than 10 cm in diameter are more likely to survive regardless of tumor composition97.

Gonadoblastoma

Gonadoblastoma is a rare tumor that may arise in a dysgenetic gonad. The patients are usually phenotypic females, but nearly all are genotypic males (that is, have a Y chromosome). The tumor contains both immature germ cells and sex cord-stromal cells, which resemble granulosa or Sertoli cells, arranged in small islands and intermixed with rounded pink hyaline bodies. Leydig cells or lutein cells are distributed through the intervening stroma in about two thirds of the cases and are responsible for the endocrine manifestations. Small calcifications may be extensive and detected by X-rays. Most gonadoblastomas are benign, but dysgerminomas and other malignant germ cell tumors develop occasionally form them98.

Metastatic tumors

Metastases to the ovary are common. Approximately 10% of all malignant tumors involving the ovary are actually metastatic, but their nature is frequently unsuspected by gynecologists and even by pathologists, particularly with the tumors are cystic99, 100. Metastases often mimic primary ovarian tumors and may even be the first manifestation of disease. In most cases, the primary site is the breast, colon or stomach, but the original tumor may be small and difficult to detect, even when the ovarian metastases are large.

The routes of spread vary, but lymphatic and hematogenous dissemination to the ovaries is the most common mechanism99, 100. Direct extension is also frequent, particularly for tumors arising from adjacent organs such as the uterus, fallopian tube and sigmoid colon. Transtubal spread of uterine cancers may also occur. Tumors from abdominal organs, such as the appendix, may also reach the ovary by the transcelomic route. Embolic spread results in the development of multiple intraovarian tumor nodules, and in these cases, vascular invasion is frequently found microscopically in the ovarian hilum99, 100.

The most helpful gross and microscopic clues in the diagnosis of metastatic tumors to the ovary are: Bilaterality, the finding of small (often microscopic) and superficial tumor nodules, extensive extraovarian dissemination, unusual patterns of spread, infrequent histologic features, vascular invasion, and prominent desmoplastic reaction99, 100.

Colonic adenocarcinomas probably account for most metastatic tumors to the ovary that are microscopically confused with endometrioid carcinoma99, 100. Frequently, the ovarian and primary tumors are discovered synchronously or the intestinal tumors have been resected months or years previously without the pathologist's knowledge100. Occasionally, the colonic primary tumor is discovered several months after resection of the ovarian tumor. About 75% of them arise in the rectum or sigmoid colon99, 100. Features favoring a metastasis are: Large cysts filled with necrotic material, abundant eosinophilic debris ("dirty necrosis"), focal segmental necrosis of the glands, occasional presence of goblet cells, and absence of müllerian features (squamous differentiation, or association with endometriosis)99, 100. Strong immunoreactivity for CEA favors colonic origin.

The ovarian stroma may be stimulated to secrete estrogenic, androgenic or progestogenic hormones in the presence of metastatic carcinoma, particularly those from the gastrointestinal tract. In these ovarian tumors with functioning stroma, hCG or hCG-like substances produced by the neoplastic cells may be responsible101, 102

Krukenberg tumor is almost invariably secondary from a primary gastric or, less frequently, intestinal adenocarcinoma. Grossly, the tumors are bilateral in about 60% of the cases, and typically large, solid, and multinodular. The basis for the selective enhancement of growth is unknown. Most patients are premenopausal, about a decade younger than those with primary ovarian adenocarcinoma; the phenomenon may be hormone dependent. The term "Krukenberg's tumor" is usually reserved for this typical clinicopathologic presentation when the tumor cells have large eccentric mucin vacuoles (signet ring cells) and the primary site is the stomach. Stromal proliferation may resemble a fibromatous tumor (fibroma, thecoma or sclerosing stromal tumor)100. A misdiagnosis of Krukenberg tumor can be made in rare cases of primary mucinous carcinoid tumor of the ovary. Chromogranin, synaptophysin, and Grimelius stains may be of help. The tubular variant, commonly associated with stromal luteinization, can be confused with sex cord-stromal tumors, particularly with Sertoli-Leydig cell tumors. PAS-diastase and mucicarmine stains are of great value103. Metastatic tumors from the pancreas, biliary tract104 and appendix28, 29 simulate primary ovarian mucinous tumors.

Breast carcinoma is the most common metastatic tumor to the ovaries, but only occasionally presents itself as a primary ovarian tumor100. It may exhibit a endometrioid-like appearance, an insular pattern simulating a carcinoid tumor or even a diffuse pattern mimicking granulocytic sarcoma or lymphoma. Positive immunostaining for EMA, S-100 and E2R protein and negativity for CA 12.5 and CA 19.9 can be useful.

Metastatic renal cell carcinomas and melanomas may be interpreted as clear cell carcinomas and granulosa cell or lipid cell tumors 100, 105.

Lymphomas and leukemias frequently involve the ovaries late in the course of the disease (25%), but rarely present themselves initially as ovarian tumors (0.2%)106, 107. The tumor cells are often distributed in single-file rows, producing a histologic picture reminiscent of metastatic lobular carcinoma of the breast. Large cell lymphomas may simulate several primary tumors, such as dysgerminoma. The appearance of the cell nuclei is very important. Immunohistochemistry for lymphoid markers and placental alkaline phosphatase are helpful. In some cases, carcinoid tumors, granulosa cell tumors or small cell carcinoma can resemble lymphomas.

The ovaries, like the facial bones and orbital tissue, are common sites of involvement in Burkitt's lymphoma, including the American variety. In rare cases, myelogenous leukemia appears initially as an ovarian tumor (so-called granulocytic sarcoma). Only a few tumors contain sufficient myeloperoxidase to have a green color on gross inspection (chloroma). Chloroacetate esterase, muramidase and F VIII immunostains are helpful.

Pathologic factors affecting prognosis

The foregoing discussion has emphasized that each different type of ovarian neoplasm is in fact a separate disease. There are, however, some general characteristics that also affect the outcome of treatment.

Established prognostic factors include FIGO stage, histological type, tumor grade; and clinical/surgical parameters including residual disease after debunking surgery, the presence or absence of ascites, performance status, and age108-110. Of these, FIGO stage is the most important independent prognostic factor. Five-year survival for Stage I tumors (limited to the ovary) is 60-80%; 45 to 60% for Stage II (spread to the pelvis); 15% for Stage III (intraabdominal or lymph node metastases); and less than 10% for Stage IV disease (distant or parenchymal liver metastases). Survival rates vary somewhat among the histologic types of epithelial-stromal tumors, for instance serous cystadenocarcinomas have a lower survival rate than endometrioid cystadenocarcinomas. This difference is most likely dependent on grade, as most endometrioid cystadenocarcinomas are well or moderately differentiated, whereas serous tumors are more often poorly differentiated. Tumor grade is related to survival rates, although this is not an independent prognosticator when controlled for stage. As these pathological and clinical parameters present an incomplete profile of the behavior of ovarian carcinomas, the search for prognostic factors continues.

Ploidy is an independent prognostic factor, with a decreased rate and length of survival in aneuploid ovarian carcinomas111. Ploidy is also related to stage, as 60 to 80% of Stage III and IV tumors are aneuploid, compared to 45 to 60% of Stage I and II tumors. Analyses of cells in the S-phase fraction, and other proliferation markers generally parallel those for ploidy; an S-phase fraction above 10% is a marker of aggressivesness112.

OC-125 is a monoclonal antibody that recognizes a glycoprotein, CA-125, present on more than 80% of nonmucinous ovarian tumors113. Serum OC-125 is elevated above 35 U/ml in more than 80% of ovarian carcinoma patients, and is most often elevated in higher stage disease.

Studies on oncogenes and anti-oncogenes in ovarian carcinoma have begun to outline their role in carcinogenesis and have shown several that are of prognostic significance, although this work is still considered investigational.

FIGO staging of ovarian cancer

Stage I

Growth limited to the ovaries.

Stage Ia: Growth limited to one ovary; no ascites. No tumor on the external surface; capsule intact.

Stage Ib: Growth limited to both ovaries; no ascites. No tumor on the external surface; capsules intact.

Stage Ic: Tumor either stage Ia or stage Ib but with tumor on the surface of one or both ovaries; or with capsule ruptured; or with ascites present containing malignant cells; or with positive peritoneal washings.

Stage II

Growth involving one or both ovaries with pelvic extension

Stage IIa: Extension and/or metastases to the uterus and/or tubes.

Stage IIb: Extension to other pelvic tissues.

Stage IIc: Tumor either stage IIa or IIb but with tumor on the surface of one or both ovaries; or with capsule or capsules ruptured; or with ascites present containing malignant cells; or with positive peritoneal washings.

Stage III

Tumor involving one or both ovaries, with peritoneal implants outside the pelvis and/or positive retroperitoneal or inguinal nodes. Superficial liver metastasis qualifies as Stage III.

Stage IIIa: Tumor grossly limited to true pelvis, with negative nodes but with histologically confirmed microscopic seeding of abdominal peritoneal surfaces.

Stage IIIb: Tumor of one or both ovaries with histologically confirmed implants of abdominal peritoneal surfaces, none exceeding 2 cm in diameter. Nodes are negative.

Stage IIIc: Abdominal implants greater than 2 cm in diameter and/or positive retroperitoneal or inguinal nodes.

Stage IV

Growth involving one or both ovaries with distant metastases. If pleural effusion is present, there must be positive cytologic findings to allot a case to stage IV. Parenchymal liver metastases equals stage IV.

Special category

Unexplored cases that are believed to be ovarian carcinoma.

Stage is based on findings at clinical examination and surgical exploration. The final histology after surgery is to be considered in the staging, as well as cytology as far as effusions are concerned.

Ascites is peritoneal effusion that, in the opinion of the surgeon, is pathologic and/or clearly exceeds normal amounts.

REFERENCES

1. Cannistra SA. Cancer of the ovary. N Engl J Med 329:1550, 1993.

2. Hamilton TC. Ovarian cancer, Part I: Biology. In: Ozols RF. Ed. Current problems in cancer. 16:3, 1992.

3. Lynch HT, Lynch JF. Hereditary ovarian carcinoma. Hematol Oncol Clin North Am 6:783, 1992.

4. Narod SA, Feunteun J, Lynch HT et al. Familial breast-ovarian cancer locus at chromosome 17q12-q23. Lancet 338:82, 1991.

5. Bell DA, Scully RE. Early de novo ovarian carcinoma. A study of 14 cases. Cancer 73:1859, 1994.

6. Berchuck A, Kohler MF, Bast RC Jr. Oncogenes in ovarian cancer. In: Ozols RF, Ed. Hematology/Oncology Clinics of North America. Philadelphia 1992, WB Saunders, 6:813.

7. Berchuck A, Kamel A, Whitaker R et al. Overexpression of HER-2/neu is associated with poor survival in advanced epithelial ovarian cancer. Cancer Res 50:4087, 1990.

8. Enomoto T, Weghorst CM, Inoue M et al. K-ras activation occurs frequently in mucinous adenocarcinomas and rarely in other common epithelial tumors of the human ovary. Am J Pathol 139:777, 1991.

9. Ichikawa Y, Nishida M, Suzuki H et al. Mutation of K-ras protooncogene is associated with histological subtypes in human mucinous ovarian tumors. Cancer Res 54:33, 1994.

10. Prat J, Cuatrecasas M, Matías-Guiu X et al. K-ras mutations in ovarian mucinous tumors (Abstr) Mod Pathol 7:94A, 1994.

11. Kohler MF, Kernes B-J M, Humphrey PA, et al. Mutation and overexpression of p53 in early-stage ovarian cancer. Obstet Gynecol 81:643, 1993.

12. Frank TS, Bartos RE, Haefner HK et al. Loos of heterozygosity and overexpression of the p53 gene in ovarian carcinoma. Mod Pathol 7:3, 1994.

13. Hart WR. Ovarian epithelial tumors of borderline malignancy (carcinomas of low malignant potential). Human Pathol 8:541, 1977.

14. Russell P. The pathological assessment of ovarian neoplasms. II: The proliferating "epithelial" tumours. Pathology 11:251, 1979.

15. Bostwick DG, Tazelaar HD, Ballon SC, Hendrickson MR, Kempson RL. Ovarian epithelial tumors of borderline malignancy. A clinical an pathologic study of 109 cases. Cancer 58:2052, 1986.

16. De Nictolis M, Montironi R, Tommasoni S, Carinelli S, Ojeda B, Matías-Guiu X, Prat J. Serous borderline tumors of the ovary: A clinicopathologic, immunohistochemical and quantitative study of 44 cases. Cancer 70:152, 1992.

17. Tazelaar HD, Bostwick DG, Ballon SC, Hendrickson MR, Kempson RL. Conservative treatment of borderline ovarian tumors. Obstet Gynecol 66:417, 1985.

18. Finkler NJ, Kopnick SJ, Griffiths CT, Knapp RG. Elevated CA-125 serum levels in epithelial ovarian cancer metastatic to retroperitoneal lymph nodes. Gynecol Oncol 29:356, 1988.

19. Tavassoli FA. Serous tumor of low malignant potential with early stromal invasion (serous LMP with microinvasion). Modern Pathology 1:407, 1988.

20. Bell DA, Scully Re. Ovarian serous borderline tumors with stromal microinvasion: A reported of 21 cases. Hum Pathol 21:397, 1990.

21. McCaughey WTE, Kirk ME, Lester W, Dardick I. Peritoneal epithelial lesions associated with proliferative serous tumors of ovary. Histopathology 8:195, 1984.

22. Bell DA, Weinstock MA, Scully RE. Peritoneal implants of serous borderline tumors: Histologic features and prognosis. Cancer 62:2212, 1988.

23. Bell DA, Scully RE. Serous borderline tumors of the peritoneum. Am J Surg Pathol 14:230, 1990.

24. Gilks CB, Bell DA, Scully RE. Serous psammocarcinoma of the ovary and peritoneum. Int J Gynecol Pathol 9:110, 1990.

25. Bell DA, Rutgers JL, Scully RE. Ovarian epithelial tumors of borderline malignancy. Progress in reproductive and urinary tract pathology (Damjanov I, Cohen AH, Mills SE, Young RH, Eds) Vol 1. p.1. Philadelphia, 1989. Field and Wood, Inc.

26. Hart WR, Norris HJ. Borderline and malignant mucinous tumors of the ovary. Histologic criteria and clinical behavior. Cancer 31:1031, 1973.

27. De Nictolis M, Montironi R, Tommasoni S et al. Benign, borderline and well-differentiated malignant intestinal mucinous tumors of the ovary: a clinicopathologic, histochemical, immunohistochemical and nuclear quantitative study of 57 cases. Int J Gynecol Pathol 13:10, 1994.

28. Young RH, Gilks CB, Scully RE. Mucinous tumors of the appendix associated with mucinous tumors of the ovary and pseudomyxoma peritonei: A clinicopathological analysis of 22 cases supporting an origin in the appendix. Am J Surg Pathol 15:415, 1991.

29. Cuatrecasas M, Matias-Guiu X, Prat J. Synchronous mucinous tumors of the appendix and the ovary associated with pseudomyxoma peritonei: A clinicopathologic study of six cases with comparative analysis of c-Ki-ras mutations. Am J Surg Pathol 1996; 20:739-746.

30. Rutgers JL, Scully RE. Ovarian mullerian mucinous papillary cystadenomas of borderline malignancy. A clinicopathologic analysis of 30 cases. Cancer 61:340, 1988.

31. Prat J, Young RH, Scully RE. Ovarian mucinous tumors with foci of anaplastic carcinoma. Cancer 50:300, 1982.

32. Prat J, Scully RE. Sarcomas in ovarian mucinous tumors. A report of two cases. Cancer 44:1327, 1979.

33. Prat J, Scully RE. Ovarian mucinous tumors with sarcoma-like mural nodules. A report of seven cases. Cancer 44:1332, 1979.

34. Matías-Guiu X, Aranda I, Prat J. Immunohistochemical study of sarcoma-like mural nodules in a mucinous cystadenocarcinoma of the ovary. Virchow Archiv A 419:89, 1991.

35. Baergen RN, Rutgers JL. Mural nodules in common epithelial tumors of the ovary. Int J Gynecol Pathol 13:62, 1994.

36. Roth LM, Czernobilsky B, Langley FA. Ovarian endometrioid adenofibromatous and cystadenofibromatous tumors: Benign, proliferating, and malignant. Cancer 48:1838, 1981.

37. Bell DA, Scully RE. Atypical and borderline endometrioid adenofibromas of the ovary. Am J Surg Pathol 9:205, 1985.

38. Snyder RR, Norris HJ, Tavassoli F. Endometrioid proliferative and low malignant potential tumors of the ovary. A clinicopathologic study of 46 cases. Am J Surg Pathol 12:661, 1988.

39. Young RH, Prat J, Scully RE. Ovarian endometrioid carcinomas resembling sex cord-stromal tumors. A clinicopathologic analysis of 13 cases. Am J Surg Pathol 6:513, 1982.

40. Eifel P, Hendrickson M, Ross J et al. Simultaneous presentation of carcinoma involving the ovary and the uterine corpus. Cancer 50:163, 1982.

41. Prat J, Matías-Guiu X, Barreto J. Simultaneous carcinoma involving the endometrium and the ovary. A clinicopathological, immunohistochemical and DNA flow cytometric study of 18 cases. Cancer 1991; 68:2455-2459.

42. Young RH, Prat J, Scully RE. Endometrioid stromal sarcomas of the ovary. A clinico-pathologic analysis of 23 cases. Cancer 53:1143, 1984.

43. Chang KL, Crabtree GS, Lim-Tan SK. Primary extrauterine endometrial stromal neoplasms. A clinicopathologic study of 20 cases and a review of the literature. Int J Gynecol Pathol 12:282, 1993.

44. Terada KY, Johnson TL, Hopkins M, Roberts JA. Clinicopathologic features of ovarian mixed mesodermal tumors and carcinosarcomas. Gynecologic Oncology 32:228, 1989.

45. Klemi PJ, Meurmann L, Gronroos M et al. Clear cell (mesonephroid) tumor of the ovary with characteristics resembling endodermal sinus tumor. Int J Gynecol Pathol 1:95, 1982.

46. Scully RE. Recent progress in ovarian cancer. Hum Pathol 1:73, 1970.

47. Scully RE, Barlow JF. "Mesonephroma" of ovary. The tumor of müllerian nature related to the endometroid carcinoma. Cancer 20:1405, 1967.

48. Austin RM, Norris HJ. Malignant Brenner tumor and transitional cell carcinoma of the ovary. A comparison. Int J Gynecol Pathol 6:29, 1987.

49. Robey SS, Silva EG, Gershenson DM et al. Transitional cell carcinoma in high-grade high-stage ovarian carcinoma. Cancer 63:839, 1989.

50. Trebeck CE, Friedlander ML, Rusell P et al. Brenner tumors of the ovary: A study of the histology, immunohistochemistry and cellular DNA content in benign, borderline and malignant ovarian tumors. Pathology 19:241, 1987.

51. Hallgrimson J, Scully RE. Borderline and malignant Brenner tumours of the ovary. A report of 15 cases. Acta Pathol Microbiol Scand (A) 80 (suppl) 233:56, 1972.

52. Miles PA, Norris HJ. Proliferative and malignant Brenner tumors of the ovary. Cancer 30:174, 1972.

53. Dickersin GR, Kline IW, Scully RE. Small cell carcinoma of the ovary with hypercalcemia: A report of eleven cases. Cancer 49:188, 1982.

54. Young RH, Oliva E, Scully RE. Small cell carcinoma of the ovary, hypercalcemic type. A clinicopathological analysis of 150 cases. Am J Surg Pathol 18:1102, 1994.

55. Aguirre GR, Thor AD, Scully RE. Ovarian small cell carcinoma. Histogenetic considerations based on immunohistochemical and other findings. Am J Clin Pathol 92:140, 1989.

56. Matías-Guiu X, Prat J, Young RH, Capen CC, De Lellis RA, Scully RE. Human parathyroid hormone-related protein (PTHrp) in ovarian small cell carcinomas: An immunohistochemical study. Cancer 73:1878, 1994.

57. Young RH, Scully RE. Ovarian sex cord-stromal tumors. Recent advances and current status. Clin Obstet Gynaecol 11:93, 1984.

58. Young RH, Dickersin GR, Scully RE. Juvenile granulosa cell tumor of the ovary. A clinicopathologic analysis of 125 cases. Am J Surg Pathol 8:575, 1984.

59. Stenwig JT, Hazekamp JT, Beecham JB. Granulosa cell tumors of the ovary. A clinicopathological study of 118 cases with long term follow-up. Gynecol Oncol 7:136, 1979.

60. Fox H, Agrawal K, Langley FA. A clinicopathologic study of 92 cases of granulosa cell tumor of the ovary with special reference to the factors influencing prognosis. Cancer 35:231, 1975.

61. Bjorkholm E, Silfersward C. Prognostic factors in granulosa cell tumors. Gynecol Oncol 11:261, 1981.

62. Prat J, Scully RE. Cellular fibromas and fibrosarcomas of the ovary: a comparative clinicopathological analysis of 17 cases. Cancer. 1981; 47:2663-2670.

63. Young RH, Scully RE. Ovarian Sertoli-Leydig cell tumors. A clinicopathological analysis of 207 cases. Am J Surg Pathol 9:543, 1985.

64. Prat J, Young RH, Scully RE. Ovarian Sertoli-Leydig cell tumors with heterologous elements: (2) With cartilaginous and skeletal muscle components. A clinico-pathologic analysis of twelve cases. Cancer 50:2465, 1982.

65. Young RH, Scully RE. Ovarian Sertoli-Leydig cell tumors with a retiform pattern: A problem in histopathologic diagnosis. A report of 25 cases. Am J Surg Pathol 77:755, 1983.

66. Hayes MC, Scully RE. Ovarian steroid cell tumor (not otherwise specified): A clinicopathological analysis of 63 cases. Am J Surg Pathol 11:835, 1987.

67. Damjanov I, Pathobiology of human germ cell neoplasia. Recent Results in Cancer Research 123:1, 1991.

68. Rutgers JL, Young RH, Scully RE. Ovarian yolk sac tumor arising from endometrioid carcinoma. Hum Pathol 18:1296, 1987.

69. Oliva E, Andrada E, Pezzica E, Prat J. Ovarian carcinomas with choriocarcinomatous differentiation. Cancer 72:2441, 1993.

70. Oosterhuis JW, Castedo SMMJ, Dejong B, et al. Ploidy of primary germ cell tumors of the testis; pathogenesis and clinical relevance. LI 50:14, 1989.

71. Damjanov I. Is seminoma a relative or a precursor of embryonal carcinoma? Lab Invest 60:1, 1989.

72. Scully RE. Tumors of the Ovary and Maldeveloped Gonads. In Atlas of Tumor Pathology, 2nd series, Fascicle No. 16. Armed Forces Institute of Pathology, Washington DC.

73. Gondos B. Comparative studies of normal and neoplastic ovarian germ cells 2. Ultrastructure and pathogenesis of dysgerminoma. Int J Gynecol Pathol 6:124, 1987.

74. De Palo G, Lattuada A, Kenda R et al. Germ cell tumors of the ovary: the experience of the national cancer institute of Milan. I. Dysgerminoma. Int J Radiat Oncol Biol Phys 13:853, 1987.

75. Bjorkholm E, Lundell M, Gyftodimos A, Silfversward C. Dysgerminoma: The Radiumhemmet series 1927-1984. Cancer 65:38, 1990.

76. Zaloudek CJ, Tavassoli FA, Norris HJ. Dysgerminoma with syncytiotrophoblastic giant cells. A histologically and clinically distinctive subtype of dysgerminoma. Am J Surg Pathol 5:361, 1981.

77. Kurman RJ, Norris HJ. Endodermal sinus tumor of the ovary. A clinical and pathologic analysis of 71 cases. Cancer 38:2404, 1976.

78. Nogales FF Jr. Matilla A, Nogales-Ortiz F et al. Yolk sac tumors with pure and mixed polyvesicular vitelline patterns. Hum Pathol 9:553, 1978.

79. Prat J, Bhan AK, Dickersin GR, Robboy SJ, Scully RE. Hepatoid yolk sac tumor of the ovary (endodermal sinus tumor with hepatoid differentiation). A light microscopical, ultrastructural and immunohistochemical study of seven cases. Cancer 50:2355, 1982.

80. Clement PB, Young RH, Scully RE. Endometrioid-like yolk sac tumor of the ovary. A clinicopathological analysis of eight cases. Am J Surg Pathol 11:767, 1987.

81. Nakashima N, Fukatsu T, Nagasaki T et al. The frequency and histology of hepatic tissue in germ cell tumors. Am J Surg Pathol 11:682, 1987.

82. Kurman RJ, Norris HJ. Embryonal carcinoma of the ovary. A clinicopathologic entity distinct from endodermal sinus tumor resembling embryonal carcinoma of the adult testis. Cancer 38:2420, 1976.

83. Prat J, Matías-Guiu X, Scully RE. Hepatic yolk sac differentiation in an ovarian polyembryoma. Surg Pathol (USA) 2:147, 1989.

84. Lindner D, McCaw BK, Hecht F. Parthenogenic origin of benign ovarian teratomas. N Engl J Med 292:63, 1975.

85. Thurlbeck WM, Scully RE. Solid teratoma of the ovary. A clinicopathological analysis of 9 cases. Cancer 13:804, 1960.

86. Yanai-Inbar I, Scully RE. Relations of ovarian dermoid cysts and immature teratomas: an analysis of 350 cases of immature teratoma and 10 cases of dermoid cyst with microscopic foci of immature tissue. Int J Gynecol Pathol 6:203, 1987.

87. Robboy SJ, Scully RE. Ovarian teratoma with glial implants on the peritoneum. Hum Pathol 1:643, 1970.

88. Kurman RJ, Norris HJ. Malignant germ cell tumors of the ovary. Hum Pathol 8:551, 1977.

89. Norris HJ, Zirkin HJ, Benson WL. Immature (malignant) teratoma of the ovary. A clinical and pathologic study of 58 cases. Cancer 37:2359, 1976.

90. Gershenson DM, Del Junco G, Silva EG et al. Immature teratoma of the ovary. Obstet Gynecol 68:624, 1986.

91. Robboy SJ, Norris HJ, Scully RE. Insular carcinoid primary in the ovary. A clinicopathologic analysis of 48 cases. Cancer 36:404, 1975.

92. Robboy SJ, Scully RE, Norris HJ. Primary trabecular carcinoid of the ovary. Obstet Gynecol 49:202, 1977.

93. Robboy SJ, Scully RE. Strumal carcinoid of the ovary: an analysis of 50 cases of a distinctive tumor composed of thyroid tissue and carcinoid. Cancer 46:2019, 1980.

94. Robboy SJ, Scully RE, Norris HJ. Carcinoid metastatic to the ovary. A clinicopathologic analysis of 35 cases. Cancer 33:798, 1974.

95. Aguirre P, Scully RE. Malignant neuroectodermal tumor of the ovary, a distinctive form of monodermal teratoma. Report of five cases. Am J Surg Pathol 6:283, 1982.

96. Kleinman GM, Young RH, Scully RE. Primary ovarian neuroectodermal tumors. A report of 25 cases. Am J Surg Pathol 17:764, 1993.

97. Kurman RJ, Norris HJ. Malignant mixed germ cell tumor of the ovary. A clinical and pathologic analysis of 30 cases. Obstet and Gynecol 48:579, 1976.

98. Scully RE. Gonadoblastoma a review of 74 cases. Cancer 25:1340, 1970.

99. Lash RH, Hart WR. Intestinal adenocarcinomas metastatic to the ovaries: A clinicopathologic evaluation of 22 cases. Am J Surg Pathol 11:114, 1987.

100. Young RH, Scully RE. Metastatic tumors of the ovary. In: Kurman RJ, ed. Blaustein's Pathology of the Female Genital Tract, ed 3. New York, Springer-Verlag, 1994.

101. Young RH, Scully RE. Sex cord-stromal, steroid cell and other ovarian tumors with endocrine, paraendocrine and paraneoplastic manifestation. In Kurman RJ, et. Blaustein's Pathology of the Female Genital Tract, ed 4. New York: Springer-Verlag, 1994.

102. Matías-Guiu X, Prat J. Ovarian tumors with functioning stroma: an immunohistochemical study of 100 cases with hCG mono and polyclonal antibodies. Cancer. 65:2001, 1990.

103. Bullon A, Arseneau J, Prat J, et al. Tubular Krukenberg tumor. A problem in histopathologic diagnosis. Am J Surg Pathol 5:225, 1981.

104. Young RH, Hart WR: Metastases from carcinoma of the pancreas simulating primary mucinous tumors of the ovary: A report of seven cases. Am J Surg Pathol 13:748,1989.

105. Young RH, Scully RE. Malignant melanoma metastatic to the ovary: A clinicopathologic analysis of 20 cases. Am J Surg Pathol 15:849, 1991.

106. Osborne BM, Robboy SJ. Lymphomas or leukemia presenting as ovarian tumors. An analysis of 42 cases. Cancer 52:1933, 1983.

107. Monterosso V, Jaffe ES, Merino MJ, Medeiros LJ. Malignant lymphomas involving the ovaries. A clinicopathologic analysis of 39 cases. Am J Surg Pathol 17:154, 1993.

108. De Souza PL, Friedlander ML. Prognostic factors in ovarian cancer. Hem /Onc Clinics N Am 6:761, 1992.

109. Saksela E. Prognostic markers in epithelial ovarian cancer. Int J Gyn Pathol 12:156, 1994.

110. Hendrickson MR, Longacre TA, Kempson RL. Clinicopathology of malignant surface epithelial neoplasms of the ovary, in Hendrickson MR, ed. Pathology: State of the Art Reviews, Vol 1, No.2, 1993. Hanley & Belfus, Inc. Philadelphia 367, 1993.

111. Iversen O-E. Prognostic value of the flow cytometric DNA index in human ovarian carcinoma. Cancer 61:971, 1988.

112. Henriksen R, Stranh P, Backstrom T et al. KI-67 immunostaining and DNA flow cytometry as prognostic factors in epithelial ovarian cancers. Anticancer Res 14:603, 1994.

113. Jacobs I, Bast RC. The CA 125 tumour-associated antigen: a review of the literature. Human Reprod 4:1, 1989.