Gestational Trophoblastic Tumors
Gestational Trophoblastic Tumors
Gestational trophoblastic tumors (GTTs) encompass a spectrum of neoplastic disorders that arise from placental trophoblastic tissue after abnormal fertilization. GTTs are classified histologically into four distinct groups: hydatidiform mole (complete and partial), chorioadenoma destruens (invasive mole), choriocarcinoma, and placental site tumor [1,2]. Most commonly, GTT results in a hydatidiform “molar” pregnancy characterized by the lack of a fetus, trophoblastic hyperplasia, edematous chorionic villi, and a loss of normal villous blood vessels [3,4].
Most molar pregnancies spontaneously resolve after uterine evacuation with no further sequelae. However, at any time during or after gestation, malignant transformation may occur in approximately 10% to 20% of molar pregnancies. Nearly two thirds of these cases have an invasive mole confined to the uterus (chorioadenoma destruens), and in one third, choriocarcinoma characterized by distant metastatic spread develops [5,6]. Placental site tumors are rare neoplasms derived from intermediate trophoblast cells of the placenta, which are identified by cellular secretion of placental lactogen and small amounts of beta-human chorionic gonadotropin (hCG) .
In the United States, GTTs are uncommon and account for less than 1% of gynecologic malignancies; however, knowledge of the natural history and management of GTTs is important because of this tumor's potential for cure with appropriate therapy. More than 40 years ago, women with choriocarcinoma had a 95% mortality rate. Today, with the advent of effective chemotherapy and the development of a reliable tumor marker (beta-hCG), a cure rate of 90% to 95% is observed for choriocarcinoma.
Current studies continue to characterize GTT, and in recent years, much has been elucidated about the pathology, molecular biology, diagnosis, and treatment of this malignancy.
In the United States, a hydatidiform mole develops in approximately 1 in 1,000 to 2,000 pregnancies [8,9]. Molar pregnancies are reported in approximately 3,000 patients per year, and malignant transformation occurs in 6% to 19% of these cases [5,10]. Complete molar pregnancies occur in 1 in 40 molar pregnancies, 1 in 15,000 abortions, and 1 in 150,000 normal pregnancies . Overall, approximately 80% of cases of GTTs, are hydatidiform moles, 15% are chorioadenoma destruens, and 5% are choriocarcinomas. Choriocarcinoma is associated with an antecedent mole in 50% of cases, a history of abortion in 25%, term delivery in 20%, and ectopic pregnancy in 5%.
True estimates of the incidence of molar pregnancies are difficult to obtain because of considerable worldwide variation in the presentation and management of both normal and abnormal pregnancies. Early evaluations suggest a 5- to 15-fold greater incidence in the Far East and Southeast Asian countries than in the United States, with as many as 1 in 120 pregnancies in Taiwan being molar . More recent studies show that in most parts of the world, the incidence of hydatidiform mole is approximately 1 in 1,000, whereas Japan and Vietnam report an incidence as high as 1 in 500 .
Ethnic and racial differences also may contribute to the variable incidence of GTTs. In one study, African-American women in the United States were estimated to have a higher GTT incidence than white women , but the findings of another study did not support this estimation . Native Alaskans were found to have an incidence rate three- or fourfold that of white women . A Hawaiian study demonstrated a lower GTT rate in white and native Hawaiians than in Filipino and Japanese populations . In the United Arab Emirates, women born in the Persian Gulf region had a higher GTT incidence than that of women of Arab and Asian origin . Other investigators, however, have found racial differences to play a minor role in GTT. The observation that Malaysian, Indian, and Chinese populations within Kuala Lumpur have similar incidences of molar disease implies a lesser role for cultural or racial differences in the etiology of GTT .
Although the etiology of GTT is not well understood, the occurrence of this tumor has been associated with several factors: extremes of reproductive age (younger than 20 and older than 40 years), prior molar pregnancy, lower socioeconomic class, and particular ABO blood groups. Women older than age 40 have as much as afivefold increased risk of molar pregnancy [15,20]. In Singapore, the incidence of molar pregnancy in women older than 45 years was found to be 1 in 72 pregnancies . In general, for women younger than age 20, a relative risk of 1.5- to 2-fold has been reported [15,20,22]. Younger patients have an improved prognosis, with a long-term disease-free survival rate of 85.4%, compared with 77.8% for women older than age 40. Paternal age does not confer an increased risk of molar pregnancy .
A history of a previous hydatidiform mole is an established risk factor for GTT. Women with a previous molar pregnancy have up to a 10 times greater risk of developing a second molar pregnancy . Furthermore, women who have had a prior molar pregnancy have at least a 1,000-fold increased risk of choriocarcinoma, compared with women who have had a normal pregnancy. The New England Trophoblastic Disease Center demonstrated the increased risk of subsequent molar pregnancy to be 1%. Thus, a normal future reproductive outcome generally can be anticipated after such a pregnancy . Nevertheless, these women should be closely monitored with a first-trimester ultrasound, pathology review of the placenta, and beta-hCG measurements 6 weeks postpartum.
Lower socioeconomic status has been associated with a greater frequency of GTT. In the Philippines, women of lower socioeconomic standing have a rate of molar disease 10 times higher than that of affluent populations . Likewise, Bertini reported a higher incidence of GTT among Israeli women of poorer Middle Eastern and African heritage than that in women of European descent . Moreover, as the standard of living improved for Israeli women of Middle Eastern origin, the incidence of GTT declined. In the Western Hemisphere, the rate of molar pregnancy is 10 times higher in Mexicans than in other North Americans.
The relationship of GTT incidence to different geographic regions, cultures, and socioeconomic statuses suggests that diet and nutrition may contribute to the etiology of this disease. The results of most studies addressing deficiencies in animal protein, animal fat, and beta-carotene have been equivocal.
Parazinni and colleagues reported that low beta-carotene consumption was associated with GTT . Further studies are needed to delineate the dietary contributions to this disease. Cigarette smoking has no strong association with GTT .
The ABO blood groups of parents appear to be related to the development of choriocarcinoma. There is a particular risk for women in blood group A who are married to men in blood group O . Thus far, human lymphocyte antigen (HLA) studies have been inconclusive in clarifying the significant association of the ABO blood group with GTT.
Normal fertilization results from the union of a single sperm and egg, which is followed by rapid cell division and the creation of an embryo. Early embryonic differentiation gives rise to trophoblasts, specialized epithelial cells responsible for connecting the embryo to the uterus and for developing the placenta and villi. This is a remarkable event involving activated transcription factors, cytokines, hormone secretion, cell-adhesion molecules, and immunologic activity . The ability of normal trophoblasts to invade the endometrium is strikingly similar to the invasive behavior of cancer. However, in GTT, uncontrolled growth and invasion of trophoblasts occurs and is a result of chromosomal abnormalities and altered cell biology.
GTTs arise from the abnormal union of sperm and egg. This often occurs when a normal sperm fertilizes an ovum in which the female genetic material is extruded, followed by the duplication of paternal chromosomes and nondivision at the first blastomere mitosis, or alternatively, the sperm is diploid due to the absence of the second meiotic division, resulting in a totally chromosomal paternal zygote. This event results in abnormalities of the trophoblast  and probably in early embryo death . This aberrant fertilization creates specific genetic abnormalities and results in distinct pathologic characteristics. Based on these morphologic and cytogenetic features, Szulman and Surti divided hydatidiform moles into two unique syndromes: complete (classic) and partial (Table 1) [3,4].
|Karyotype||46XX (90%)||69XXY (90%)|
In addition, the roles of proto-oncogenes, tumor-suppressor genes, cytokines, and growth factors also are contributing to our understanding of GTT and tumor progression.
Hydatidiform moles, invasive moles, and choriocarcinomas have distinct morphologic features. Moles are described as partial or complete (classic) based on their morphologic, karyotypic, and clinical features [3,4]. Complete moles are distinguished by the complete absence of normal villi and by chromosomal material that is virtually always of paternal origin. Partial moles are characterized grossly by an admixture of normal and hydropic villi, a triploid karyotype, and the presence of both maternal and paternal chromosomal material.
Complete moles usually are detected during the second trimester and are identified by total hydatidiform enlargement of the villi, which are enveloped by hyperplastic and atypical trophoblasts . There is a notable absence of any embryonic or amniotic remnant. More than 90% of classic moles demonstrate a 46XX karyotype, which has been demonstrated to be of paternal origin by fluorescent banding polymorphic analysis . Approximately 20% of complete moles give rise to persistent trophoblastic disease .
Partial moles, in contrast, are more commonly accompanied by an identifiable embryo or amniotic membranes. These moles are described as partial because the hydatidiform changes in the villi tend to be focal. The hydropic villi usually are irregularly scalloped and have stromal hyperplastic inclusions . The villous capillaries appear to be functional, because they possess the same proportion of nucleated fetal erythrocytes as are found in the embryo. In partial moles, hydatidiform change occurs at a slower rate, and the proportion of relatively normal villi appears to correlate with the fetal survival rate. Maturation of mesenchymal elements is only minimally delayed, and there is a paucity of fibroblast karyorrhexis. Partial moles are usually aneuploid and most often exhibit an XXY karyotype, which is thought to be secondary to dispermic fertilization of the ovum with retention of the maternal genome. Approximately 2% of partial moles undergo malignant degeneration. Because of this sporadic malignant potential, follow-up and treatment of patients with partial moles are the same as for patients with complete moles.
Locally invasive moles have the same histologic features as complete moles and, in addition, are characterized by myometrial invasion without involvement of intervening endometrial stroma . Invasive moles are typically diagnosed approximately 6 months after molar evacuation. They tend to invade locally, causing hemorrhage and necrosis. Rarely, uterine perforation results. Hematogenous metastasis may occur, often to the lungs. Occasionally, metastatic deposits display hydropic villi, rather than the sheets of anaplastic cells that typify metastatic choriocarcinoma.
Choriocarcinomas have a unique histology that is distinct from that of the moles . The tumor is grossly red and granular and exhibits extensive necrosis and hemorrhage. Microscopically, the neoplasm is composed of a disordered array of syncytiotrophoblastic and cytotrophoblastic elements, frequent mitoses, and multinucleated giant cells. Vascular invasion occurs early, with resultant metastases to the lungs, vagina, brain, kidneys, liver, and gastrointestinal tract.
Placental site tumors are rare. They are derived from intermediate trophoblast cells of the placenta, which are identified by the secretion of placental lactogen and small amounts of beta-hCG . Occasionally, after a complete hydatidiform mole is removed, an unusual complication develops, characterized by a proliferation of intermediate trophoblast-forming nodules in the endometrium and myometrium . There are usually numerous nodules, which appear microscopically as cells with oval nuclei, and an abundant eosinophilic cytoplasm. No chorionic villi are seen. These tumors usually present as nodules confined to the endometrium and myometrium, produce a mild elevation in the hCG titer, do not respond well to chemotherapy, and may progress. In such instances, the lesions should be located and surgically removed to avoid unnecessary and ineffective chemotherapy.
Pathologic characteristics alone generally do not allow adequate discrimination of molar pregnancies. With the advent of cytogenetic techniques, such as chromosomal banding and restriction fragment-length polymorphism analysis of DNA, unique chromosomal patterns of molar pregnancies were discovered [33,38], allowing a recognizable distinction between complete and partial moles [3,39].
Using chromosomal banding, Kajii and Ohama first reported that complete moles contained only paternal chromosomes . Yamashita and colleagues confirmed this finding by showing that when paternal heterozygotes for the HLA locus give rise to a mole, the HLA expression of the molar tissue is homozygous . Approximately 85% to 92% of complete moles have a 46XX karyotype [33,41], which results from fertilization of an egg by a haploid sperm (23X) that undergoes duplication to create a diploid set of chromosomes. Why the loss of maternal DNA takes place is uncertain; it may involve extrusion of maternal chromosomes or fertilization of an empty egg. Regardless of the mechanism, the finding of maternal mitochondrial DNA suggests that moles result from an abnormal fertilization event .
Approximately 4% to 15% of complete moles have a 46XY karyotype [43,44], which results from dispermy, in which two spermatozoa (23X and 23Y) fertilize an empty ovum. There is no strong evidence that dispermic or Y chromosome-containing moles have greater malignant potential than the monospermic 46XX karyotype . Fisher and colleagues also found that nearly 5% of complete moles are heterozygous 46XX . A 46YY mole has not been reported because the X chromosome is probably required for survival.
The typical partial mole has a triploid karyotype (69 chromosomes), and both paternal and maternal chromosomes are present . The most common sex chromosome arrangement is XXY. The triploid genotype can result in two phenotypes. If the extra haploid chromosome is of paternal origin, a partial mole arises; if it is of maternal origin, a fetus develops .
Growth Factors and Oncogenes
The excess of paternal chromosomes in moles probably contributes to the induction of trophoblastic hyperplasia. The genomic imbalance may cause changes in the gene expression of growth factors located on the paternal allele . An insulin-like growth factor (IGF2) specifically located on the paternal allele may be inappropriately expressed in molar pregnancies, thus stimulating uncontrolled growth.
Both normal placentas and molar pregnancies contain paternal antigens; therefore, upon implantation, an immunologic response is initiated with infiltration of lymphocytes and macrophages and secretion of cytokines . The growth of choriocarcinoma may be related to the abundant expression of epidermal growth factor (EGF) receptor. Macrophage-derived cytokines—interleukin (IL-1-alpha, IL-1-beta), and tumor necrosis factor—can suppress cell growth and increase EGF receptor expression in choriocarcinoma cell lines, thus acting as paracrine mediators of cell growth .
The contribution of several oncogenes to the malignant transformation of GTT also has been examined. Growth regulation in the trophoblast recently has been found to be associated with expression of the transcription factor Mash-2 . Cheung et al have demonstrated increased expression of c-fms RNA in complete moles compared with that in normal placentas . In choriocarcinoma, increased expression of c-myc and ras RNA has been observed . At present, the significance of these findings is uncertain. Because trophoblasts are, by nature, rapidly dividing and invasive, increased expression of these oncogenes may be essential for normal cell function. Further studies are needed to elucidate these findings. Recently, expression of the c-erb B-2 oncogene product in persistent GTT was examined and found to have no significant contribution . Thus far, no gene mutations or rearrangements in GTT have been reported.
Progression of some tumors has been associated with the inactivation of tumor suppressor genes. The inactivation of p53 by mutation of the p53 gene has been observed in nearly 50% of patients with ovarian cancer . Expression of p53 in hydatidiform moles has recently been studied . Expression of p53 in moles was observed to be increased over that in normal trophoblasts. No p53 mutations were found. Persaud and colleagues further noted an overaccumulation of p53 protein in 50% of choriocarcinomas and 78% of hydatidiform moles but none in partial mole and normal placenta . Increased p53 expression may thus be an attempt to abrogate excessive trophoblastic proliferation in hydatidiform moles.
Several clinical features typify hydatidiform moles and metastatic trophoblastic disease.
The classic signs of a molar pregnancy include the absence of fetal heart sounds, physical evidence of a uterus that is larger than expected for gestational age, and vaginal bleeding. Although an intact fetus may coexist with a partial mole, this occurs in fewer than 1 in 100,000 pregnancies.
The most common presenting symptom of molar pregnancy is vaginal bleeding, reported in up to 97% of patients . Intrauterine clots may undergo oxidation and liquefaction, producing pathognomonic prune juice-like fluid. Rarely, spontaneous expulsion of grape-like villi will provide the diagnosis of hydatidiform mole. Prolonged or recurrent bleeding may result in iron-deficiency anemia. Symptoms of anemia occur in approximately 50% of patients at the time of diagnosis.
Abdominal pain may result from excessive uterine enlargement or prominent theca luteal cysts. An abdominopelvic examination may reveal a uterus larger than expected for the gestational date and with an irregular contour. Ovarian masses resulting from theca-luteal cysts may be palpable. Theca-luteal cysts, caused by hCG-induced hyperstimulation of both ovaries in about 50% of patients, may result in pelvic pressure or fullness. Usually, these cysts regress spontaneously after uterine evacuation; however, their rupture or tension can cause acute abdominal symptoms occasionally requiring surgery.
Early toxemia (hypertension, proteinuria, and edema) presenting during the first or second trimester is not uncommon in molar pregnancy. Toxemia, which was observed in 27% of patients at the New England Trophoblastic Disease Center , is thought to be precipitated by the release of large amounts of vasoactive substances from necrotic trophoblastic tissue. Very rarely, eclamptic convulsions may occur in this setting.
Hyperemesis gravidarum—protracted nausea and vomiting during pregnancy—is observed in approximately 10% of women with GTT. The mechanism is not well understood.
Hyperthyroidism is seen in approximately 7% of molar pregnancies. An elevation of triiodothyronine (T3) and thyroxine (T4) levels is observed more commonly than are the clinical manifestations of tachycardia, sweating, weight loss, and tremor. These hormonal elevations are presumed to be secondary to the structural similarity of hCG to thyroid-stimulating hormone (TSH); thus, markedly elevated hCG levels intrinsically stimulate thyroid activity . However, the findings surrounding the correlation of thyroid function to hCG level are conflicting and suggest that other substances elaborated from GTT are responsible for the hyperthyroidism . Rarely, patients may develop thyroid storm, which may be precipitated by surgical stress during molar evacuation. In such cases, the administration of beta-adrenergic blockers is prudent.
Because placental tissues are rich in thromboplastin-like substances, the extrinsic coagulation pathway is occasionally activated, resulting in the consumption of platelets and clotting factors. In rare instances, frank disseminated intravascular coagulation and microangiopathic hemolytic anemia may develop; these life-threatening emergencies usually resolve with molar evacuation.
Patients with partial mole do not have the same clinical features as those with complete mole; fewer than 10% of patients with the former have uterine enlargement. Goldstein and Berkowitz reviewed the cases of 81 patients with partial mole and found that none had prominent theca-luteal cysts, hyperthyroidism, or respiratory insufficiency and only 1 had toxemia . The diagnosis of partial mole was usually made after histologic review of curettage specimens.
Metastatic Trophoblastic Disease
Metastatic GTT is reported in 6% to 19% of patients after molar evacuation [5,10]. Metastases sometimes have an identical histology to that of molar disease, but the vast majority are choriocarcinomas. Metastatic spread is hematogenous. Because of its extensive vascular network, metastatic GTT often produces local, spontaneous bleeding. Berkowitz et al at the New England Trophoblastic Disease Center (NETDC) reported that the common metastatic sites of GTT are the lungs (80%); vagina (30%); pelvis (20%); liver (10%); brain (10%); and bowel, kidneys, and spleen (5% each) .
Pulmonary metastases are quite common (80% of patients with metastatic disease ) and occur when trophoblastic tissue enters the circulation via uterine venous sinuses. Most often this happens spontaneously, but it also may occur after molar evacuation. Because choriocarcinoma is a vascular tumor, hemoptysis is a frequent symptom of lung involvement. Other symptoms include chest pain, dyspnea, and cough. Pulmonary hypertension also may develop. In some cases, an asymptomatic lesion on a chest x-ray or computed tomographic (CT) scan may be the only sign of pulmonary involvement . The radiologic features may be protean or subtle and include alveolar, nodular, and miliary patterns . Pleural effusions may also be present. Pulmonary metastases can be extensive and can cause respiratory failure and death.
Right upper-quadrant pain has been observed when hepatic metastases stretch Glisson's capsule. Gastrointestinal lesions may result in severe hemorrhage or in perforation with peritonitis, both of which require emergency intervention. Vaginal examination may reveal bluish metastatic deposits; these and other metastatic sites should not undergo biopsy because severe uncontrolled bleeding may occur.
Central nervous system (CNS) involvement from metastatic GTT suggests widespread disease and has a poor prognosis. CNS metastases are clinically evident in 7% to 28% of patients with metastatic choriocarcinoma [22,30,56,62]. Bakri and colleagues reported a 17% incidence of patients with GTT metastatic to the brain . These patients had presenting neurologic symptoms of headache, hemiparesis, vomiting, dizziness, coma, grand mal seizure, visual disturbances, aphasia, and slurred speech. Some patients had multiple neurologic complaints, and a few were asymptomatic. All patients had concurrent pulmonary metastases. Cerebral metastases tend to respond favorably to both radiotherapy and chemotherapy.
Although the clinical presentation may suggest a diagnosis of GTT, certain laboratory studies, particularly a determination of the patient's beta-hCG level, and radiographic studies are needed to confirm this diagnosis.
The blood count will usually reveal anemia of mixed morphologic characteristics. If disseminated intravascular coagulation is present, thrombocytopenia, prolonged clotting times, and consumption of coagulation factors may be observed. Uncommonly, blood chemistry studies may reveal hepatic or renal abnormalities.
Thyroid function studies should be performed in all patients with a clinical history or physical examination suggestive of hyperthyroidism. Abnormal thyroid function, manifested as an elevated T4 level, is not uncommon in GTT. Metastatic deposits in the kidneys or gastrointestinal tract may reveal themselves by hematuria or hematochezia.
Tumor Markers: A well-characterized glycoprotein hormone secreted by the syncytiotrophoblast, hCG is essential to maintaining normal function of the corpus luteum during pregnancy . This hormone has an alpha-subunit identical to the alpha-subunit of the pituitary hormones and a beta-subunit (beta-hCG) that confers the hormone's unique biologic activity. The presence of hCG appears approximately 8 days after ovulation, and its concentration doubles every 2 to 4 days, until it peaks at 10 to 12 weeks of gestation; thereafter, beta-hCG levels decline steadily. Because all trophoblastic tumors secrete beta-hCG, this hormone serves as an excellent marker for tumor activity in the nonpregnant patient [65-67]. The beta-hCG level will always be elevated in a molar pregnancy and, therefore, should be measured in all suspected cases.
CA-125 may also have a role as a marker for GTT. In patients with complete hydatidiform mole, the CA-125 level was elevated; more significant was the association of the degree of CA-125 elevation with the development of persistent GTT .
Serial beta-hCG levels should be monitored during therapy to ensure adequate treatment. The level of beta-hCG is roughly proportional to the tumor burden and inversely proportional to therapeutic outcome. The approximately 10% to 20% of patients with hydatidiform mole who are not cured by local therapy or do not achieve a spontaneous remission can be identified by a rising or plateauing beta-hCG titer on serial determinations after the evacuation of a mole. These patients are considered to have persistent trophoblastic disease and require additional therapy.
Special Test: The ratio of beta-hCG in serum to beta-hCG in the cerebral spinal fluid (CSF) may contribute to the detection of brain metastases in GTT. A serum:CSF beta-hCG ratio of less than 60:1 is considered a positive predictor for brain metastases. Athanassiou et al compared brain CT scans with serum:CSF beta-hCG ratios for the detection of intracranial metastatic disease . Of 19 patients who underwent CT and measurement of serum:CSF beta-hCG ratios, 11 patients had positive brain CT scans and positive serum:CSF beta-hCG ratios, 7 had disease demonstrated by CT scan alone, and 1 patient had a positive ratio but a negative CT scan. No patient with both a negative CT scan and a negative ratio was found to have metastases on an isotope scan. Bagshawe and Harland reported similar results, with 29 of 33 patients who had CT-documented GTT brain involvement also having a positive serum:CSF beta-hCG ratio .
Because 70% to 80% of patients with metastatic GTT have lung involvement, a chest x-ray should always be performed [56,71]. Although this x-ray usually demonstrates nodular metastases, the patterns of metastatic disease can range from atelectatic areas to subtle pleural abnormalities. A CT scan often is helpful in evaluating these nonspecific areas.
Because it has been demonstrated that 97% to 100% of patients with CNS disease from choriocarcinoma have concomitant pulmonary metastases, a CNS workup in asymptomatic patients with normal chest x-rays is not routinely warranted . If the chest x-ray is abnormal, or if beta-hCG levels plateau or rise during treatment, a more thorough evaluation for metastatic disease is indicated. CT scans of the brain, abdomen, and pelvis should be performed to evaluate other likely sites of metastatic spread.
Historically, angiography and amniography were important procedures used in the diagnosis of intrauterine molar disease, but today, ultrasonography is the preferred diagnostic modality. Ultrasonography is a reliable, safe, economical, and relatively simple method for confirming the diagnosis of GTT. It is also useful in identifying embryonic remnants. More recently, transvaginal color-flow Doppler ultrasonography has become increasingly useful in the diagnosis and assessment of GTT .
Magnetic resonance imaging (MRI) may be useful in equivocal cases, particularly for evaluating the cerebellum and brain stem. MRI may replace CT because of the former's superior imaging characteristics of vascular metastases and improved definition of the brain stem and cerebellum, which are sites of occult metastases. The presence of intrauterine or ovarian disease also may be detected by MRI of the pelvis.
Staging systems attempt to define prognostic groups that can direct a rational therapeutic strategy. Such criteria depend on the tumor duration, volume, and location as well as on essential patient characteristics. Although several staging systems have been proposed, there is currently no uniform system for staging metastatic GTT. The World Health Organization (WHO) and the International Federation of Gynecology and Obstetrics (FIGO) have devised commonly used staging classifications [1,2].
The FIGO staging system is a straightforward system based on anatomic criteria. In GTT, stage-0 tumor is a molar pregnancy limited to the uterine cavity; stage-I tumor is confined to the uterine body; and stage-II tumor includes local metastases to the pelvis and vagina. Stage-III tumor involves pulmonary metastases, whereas stage-IV tumor consists of distant metastatic disease. The WHO classification scheme (Table 2) is based on a scoring system to identify patients at high risk for treatment failure. With the WHO scoring system, patients are classified as being at low, middle, or high risk for treatment failure. A total score of up to 4 is considered low risk, 5 to 7 middle risk, and 8 or greater high risk.
|Age (yr)||39 or less||> 39|
|Antecedent pregnancy||Hydatidiform mole||Abortion||Term|
|Interval (mo)*||< 4||4-6||7-12||>12|
|hCG (IU/L)||< 1,000||1,000-|
|ABO groups |
(female × male)
|O × A or |
A × O
|B or AB|
|Largest tumor||< 3 cm||3-5 cm||> 5 cm|
|Site(s) of metastases||Lung||Spleen, kidney||GI tract, liver||Brain|
|Number of metastases||1-3||4-8||> 8|
|Prior chemotherapy||One drug||Two or more drugs|
|ª The total score for a patient is obtained by adding the individual scores of each prognostic factor: 4 = low risk; 5-7 = intermediate risk; > 7 = high risk.|
* Time between end of antecedent pregnancy and start of chemotherapy.
Adapted, with permission, from Sugarman SM, Javanagh JJ: Gestational trophoblastic tumors, in Pazdur R (ed): Medical Oncology: A Comprehensive Review, Huntington, NY, PRR Inc, 1993.
Several factors have been associated with a poor prognosis [1,22]. One important predictor of poor outcome is a large tumor volume at the time of diagnosis, manifesting as a high serum beta-hCG level (greater than 100,000 mIU/mL) or as bulky disease (mass greater than 5 cm in diameter). The number of metastatic sites is inversely related to prognosis. Patients with metastatic sites in the brain and liver have a worse outcome than patients with metastases to other sites. Women are at greater risk if the disease persists longer than 4 months after the antecedent pregnancy. Previous failure of chemotherapy, age older than 39 years, and maternal/paternal blood type also define a poor outcome.
Although the treatment strategy for GTT must be individualized for each patient, Figure 1 summarizes the general diagnostic and therapeutic approach used at The University of Texas M.D. Anderson Cancer Center . The stratification of risk groups enables physicians to direct an appropriate treatment strategy. Low-risk disease responds readily to single-agent chemotherapy and is virtually 100% curable. High-risk disease is not likely to be cured with single-agent therapy and therefore requires multidrug regimens.
For patients with complete or partial hydatidiform mole, evacuation of the mole by suction and sharp curettage should be performed . Oxytocics also are given to produce uterine involution and to control bleeding. However, these agents should be used judiciously as they may cause hyponatremia and fluid overload. A baseline chest x-ray and beta-hCG measurement should be obtained prior to surgery. If curettage is performed after 16 weeks' gestation, an increased incidence of uterine perforation, hemorrhage, and pulmonary complications arises. In this situation, cardiopulmonary functions should be closely monitored, with surgical and critical care facilities available .
If the patient is older than age 40 or desires sterilization, hysterectomy is a possible alternative. Approximately 25% of patients with molar pregnancy will have a prominent theca-luteal cyst; because these cysts will regress spontaneously, the uterine adnexa need not be removed during the hysterectomy unless adnexal metastases are seen. The safety and low toxicity of another alternative, a methotrexate/leucovorin regimen, make hysterectomy less appealing. Nevertheless, hysterectomy significantly decreases the risk of malignancy (3.5% vs 20% ) and is the preferred treatment in patients with placental site tumors because they are usually resistant to chemotherapy.
After molar evacuation, 80% of patients will need no further intervention [5,6,34]. However, these patients' weekly serum beta-hCG levels must be diligently monitored until they return to normal. Although normal beta-hCG levels typically return within 8 weeks of surgery, in a minority of patients, normal levels return at 14 to 16 weeks. Sometimes transient plateaus are observed before the beta-hCG level returns to baseline; however, an increased or prolonged plateau of beta-hCG titers implies persistent trophoblastic disease or metastatic spread and requires additional therapy.
Chemotherapy is indicated when there is a plateau or increase in beta-hCG levels on consecutive measurements, failure to reach normal titers by 16 weeks, or metastatic disease. Such patients are usually at low risk and will respond to single-agent chemotherapy. Methotrexate is the most commonly initiated single agent (Table 3). Other agents used successfully in this setting include dactinomycin (Cosmegen), fluorouracil, and etoposide (VePesid) [77,78]. Therapy is continued for one to two courses after a normal beta-hCG level is achieved.
|1 mg/kg (up to 70 mg) IM or IV days 1, 3, 5, 7||14 days|
|Folinic acid||0.1 mg/kg IM or IV days 2, 4, 6, 8|
|Methotrexate||0.4 mg/kg IM or IV daily for 5 days||14 days|
|Methotrexate||30 to 50 mg/m² IM||7 days|
|Dactinomycin||10 µg/kg (up to 0.5 mg) IV daily for 5 days||14 days|
|Dactinomycin||1.25 mg/m² IV single dose||14 days|
|ª Therapy based on SHO risk criteria (see Table 2).|
* Withhold treatment for marrow recovery if necessary.
Prophylactic Chemotherapy: The use of prophylactic chemotherapy after molar evacuation remains controversial. Although short courses of dactinomycin and methotrexate after molar evacuation reduce the incidence of persistent or recurrent disease in patients at high risk, the effectiveness of chemoprevention in patients at low risk has not been proven. In 1965, a prospective study of the use of chemotherapy at the time of molar evacuation was conducted under the auspices of the National Cancer Institute to determine whether chemoprevention could reduce or eliminate the malignant sequelae of molar disease . Of the 3 schedules evaluated, dactinomycin, 12 µg/kg/d intravenously for 5 days starting 3 days prior to molar evacuation, was found to be optimal in effectiveness and toxicity.
Goldstein randomized 200 patients to receive the same regimen or observation at the time of molar evacuation . He found no cases of metastatic trophoblastic disease in the treated group, whereas four cases developed in patients who did not receive chemoprevention. Two of the four cases were hydatidiform moles, one was choriocarcinoma, and the pathology in the fourth was undetermined . Because the disease remained treatable after progression or relapse, there was no overall survival benefit for the patients who received prophylactic chemotherapy.
In a prospective, randomized study of 71 patients, Kim et al found that chemoprophylaxis in patients with high-risk moles significantly reduced the incidence of persistent molar disease from 47% to 14%. However, among low-risk patients, there was no difference in the ability to achieve a complete response between the group receiving chemoprevention and the control group . More courses of chemotherapy (2.5 vs 1.4) were required to produce a complete response in the group receiving chemoprophylaxis, implying that it induced tumor resistance. It thus seems reasonable to administer chemoprevention only if it is doubtful whether the patient will return for serial beta-hCG determinations.
Follow-up: As mentioned previously, all patients with molar disease should obtain a baseline chest x-ray. Serum beta-hCG levels should be obtained every 1 to 2 weeks until the level is normal for 3 consecutive assays. Complete remission is defined by three consecutive normal beta-hCG levels. Once this has occurred, beta-hCG levels should be checked monthly for 12 months, every 4 months for the following year, and then yearly for 2 years.
Although the use of oral contraceptives during the surveillance period remains controversial, strict contraception is required, because pregnancy would obviate the usefulness of beta-hCG as a tumor marker. In general, once a 12-month surveillance establishes a disease-free status, conception is acceptable, although these women are always at high risk for future molar disease and will require close observation during future pregnancies . A pelvic ultrasound examination should be performed during the first trimester of all subsequent pregnancies to confirm that gestation is normal .
Occasionally, women who undergo removal of a complete hydatidiform mole may develop the unusual complication of intermediate trophoblastic disease . Such women usually present with vaginal bleeding and a slightly elevated hCG titer. Examination of the uterus reveals multiple nodules involving the endometrium and myometrium. These nodules do not readily respond to chemotherapy, and surgical intervention is therefore appropriate because progressive disease tends to develop.
Finally, physicians should maintain an awareness of the psychological, social, and sexual effects of GTT on patients and their partners. Most of these women experience a significant level of distress associated with GTT . Many are likely to have fears of infertility despite physician reassurance. These women may also develop sexual dysfunction and lack of sexual desire, which may be attributed, in part, to a fear of future pregnancy. Psychosocial support should be provided, especially for patients undergoing treatment of metastatic disease.
Low-Risk Metastatic Disease
In more than 30 years of experience, single-agent chemotherapy with methotrexate has produced a high cure rate in patients with low-risk GTT . Likewise, methotrexate plus leucovorin induces remission in 90% of patients with low-risk metastatic disease and with low toxicity [86,87]. The use of dactinomycin in methotrexate-resistant patients increased the cure rate to more than 95%. Dactinomycin also has been used successfully as single-agent therapy for low-risk GTT [88,89]. Approximately 80% of patients experience toxic reactions with dactinomycin, the most common being nausea and vomiting. The toxicity of methotrexate is minimal and may result in mucositis and myelosuppression.
Suggested therapeutic regimens for low-risk GTT are outlined in Table 3 [80,84]. At The University of Texas M.D. Anderson Cancer Center, the regimen of methotrexate plus leucovorin is preferred because it obviates intravenous access problems, allows the therapeutic administration at home or work, minimizes the interruption of the patient's life, and is the least toxic of the regimens listed. These patients are usually treated for two to three courses after attaining a normal beta-hCG level.
If single-agent therapy with methotrexate or dactinomycin fails, or both, to achieve remission, multidrug chemotherapy must be attempted. This is necessary in nearly 40% of patients with low-risk metastatic GTT. Despite resistance to first-line chemotherapy, a cure rate of almost 100% is achieved with further combination chemotherapy .
High-Risk Metastatic Disease
Because single-agent therapy is inadequate for patients with high-risk GTT, intensive combination chemotherapy is mandatory. The most widely employed regimens include MAC (methotrexate, dactinomycin, and cyclophosphamide [Cytoxan, Neosar] or chlorambucil [Leukeran]), EMA-CO (etoposide, methotrexate, dactinomycin, cyclophosphamide, and vincristine [Oncovin]), and the modified Bagshawe regimen (Table 4).
|EMA-CO* (preferred regimen)|
|Course 1 (EMA)|
|Day 1||Etoposide||100 mg/m² IV over 30 min|
|Methotrexate||100 mg/m² IV bolus|
|Methotrexate||200 mg/m² IV as 12-h continuous infusion|
|Dactinomycin||0.5 mg IV bolus|
|Day 2||Etoposide||100 mg/m² IV over 30 min|
|Leucovorin||15 mg IV/IM/PO every 12 h for 4 doses, beginning 24 h after start of methotrexate|
|Dactinomycin||0.5 mg IV bolus|
|Course 2 (CO)|
|Day 8||Cyclophosphamide||600 mg/m² IV over 30 min|
|Vincristine||1 mg/m² (up to 2 mg) IV bolus|
|EHMMAC^ (alternative for middle-risk patients)|
|Etoposide||100 mg/m² IV daily for 5 days|
|Hydroxyurea||0.5 mg PO, repeat in 12 h, day 1|
|Methotrexate||50 mg IM, repeat every 48 h, days 2, 4, 6, 8|
|Dactinomycin||0.5 mg IV daily for 5 days|
|Vincristine||1.0 mg/m² IV days 1 and 3|
|Cyclophosphamide||400 mg/m² IV days 1 and 3|
|ª Therapy based on WHO risk criteria (see Table 2).|
* Repeat each regimen in sequence every 14 days as toxicity permits.
^ Courses should be given 7 to 10 days apart as toxicity permits. The sequence of courses is 1, 2, 3, 4, 1, 2, etc. Alternatively, regimen 4 may be held in reserve, to be used if one of the other regimens proves ineffective or toxic. The preferred sequence is the 1, 2, 3, 1, 2, 3, etc.
MAC has been the most widely used primary multidrug regimen and has produced cure rates ranging from 63% to 80% [91-94]. MAC dosages consist of methotrexate, 0.3 mg/kg intravenously (IV) or intramuscularly (IM); dactinomycin, 8 µg/kg IV; and cyclophosphamide, 3 mg/kg IV, or chlorambucil, 0.15 mg/kg PO. Each drug is given as a 5-day course for 3 to 4 courses, with a 9- to 14-day interval between each course. In middle- to high-risk GTT, MAC is most efficacious when used as initial chemotherapy (65% survival), as opposed to secondary chemotherapy (39% survival) following failed single-agent therapy.
The CHAMOCA regimen (hydroxyurea [Hydrea], dactinomycin, methotrexate with leucovorin rescue, cyclophosphamide, vincristine, and doxorubicin [Adriamycin, Rubex]), introduced by Bagshawe in the mid-1970s, produced a remission rate of 82%  but was inferior to MAC in terms of toxicity and effectiveness in a trial sponsored by the Gynecologic Oncology Group . The discovery that etoposide is an effective agent against trophoblastic disease led to the development of the EMA-CO regimen by Bagshawe, who reported a survival rate of 83% in patients with high-risk choriocarcinoma . This regimen has been confirmed to be highly effective at several centers, including the Brewer Trophoblastic Disease Center, where a 100% cure rate has been achieved over the past 5 years .
EMA-CO (Table 4) is the preferred regimen for high-risk GTT. We also utilize this regimen for patients with middle-risk GTT, as defined by the WHO criteria. EMA-CO is generally well tolerated, with no life-threatening toxic effects. Alopecia occurs universally, and anemia, neutropenia, and stomatitis are mild. Reproductive function is preserved in approximately 75% of patients.
Within hours of receiving chemotherapy, patients with a significant tumor burden are at risk of hemorrhage into tumors and surrounding tissues. Thus, any acute organ toxicity that begins shortly after the induction of chemotherapy should be considered as possibly related to this phenomenon; to minimize these sequelae, some researchers have advocated a reduction in dosage at the beginning of therapy in patients with large-volume disease.
Surgical resection of persistent sites of disease (including the uterus), with or without chemotherapy, is an important strategy for any patient with chemotherapy-resistant disease. Identifying the sites of active or inactive but radiologically apparent disease is a central problem in this strategy. It is hoped that improvements in monoclonal-antibody scanning techniques will help to overcome this problem.
Unfortunately, about 25% of women with high-risk metastatic disease become refractory to EMA-CO and fail to achieve a complete remission. Currently, there is no standard salvage chemotherapy regimen with EMA-CO failures. However, salvage regimens consisting of combinations of cisplatin (Platinol), etoposide, vinca alkaloids, and bleomycin (Blenoxane) have been administered.
Because of significant nephrotoxicity, cisplatin-containing regimens are withheld as primary therapy for GTT. Early studies show cisplatin-based regimens to be an effective salvage therapy in GTT. Cisplatin in combination with vincristine and methotrexate achieved a 33% response rate as salvage therapy . At The University of Texas M.D. Anderson Cancer Center, Gordon et al reported a sustained remission rate of 20% in 10 patients treated with PVB (cisplatin, vinblastine, and bleomycin) after failure to respond to MAC . A recent dose-intensive regimen, EMA-CE, utilizes cisplatin (100 mg/m²) and etoposide (200 mg/m²) combined with EMA, with favorable results .
Another alternative is to give cisplatin in the EMA-POMB regimen (cisplatin, vincristine, methotrexate, bleomycin). POMB is administered as vincristine, 1 mg/m² IV, and methotrexate, 300 mg/m² IV (day 1); bleomycin, 15 mg IV over 24 h by continuous infusion (CI), and folinic acid, 15 mg bid for four doses (day 2); bleomycin, 15 mg IV over 24 h CI (day 3); and cisplatin, 120 mg/m² IV (day 4).
A new PEBA regimen (cisplatin, etoposide, bleomycin, doxorubicin) was recently reported from China and was found to be effective in EMA-CO-resistant disease . A complete remission (CR) was achieved in 96% of the women, and 73% had a sustained CR that lasted at least 1 year. In a small study, ifosfamide (Ifex) alone and in combination with etoposide and cisplatin (VIP) showed promise as being an effective salvage drug in GTT .
Another consideration in the treatment of refractory GTT is the use of high-dose chemotherapy with autologous bone marrow transplantation. In this setting, Lotz et al treated five women with refractory GTT with high-dose ifosfamide, carboplatin (Paraplatin), and etoposide (ICE). Only one of the five women attained a durable CR (68+ months). The risk and benefits of high-dose chemotherapy in the treatment of GTT is still under investigation.
At the time of diagnosis, pulmonary metastases can be extensive and may cause respiratory failure and death . Kelly et al identified several prognostic factors for early respiratory death in patients presenting with GTT and dyspnea . These factors included cyanosis, pulmonary hypertension, anemia, tachycardia, extensive lung opacification, and a high WHO prognostic score. Bakri and colleagues described similar associations with early respiratory failure in 75 patients with pulmonary metastases . Their clinical and radiographic findings included dyspnea, anemia, clinical pulmonary hypertension, cyanosis, more than 50% lung opacification, mediastinal involvement, and bilateral pleural effusion. Patients with extensive lung opacification (particularly when associated with anemia), pulmonary hypertension, or cyanosis were at high risk for respiratory failure.
In patients with extensive pulmonary metastases, reduced doses of initial chemotherapy (eg, 50%) have been suggested to diminish the risk of respiratory failure . However, reduction of the initial chemotherapy dose did not uniformly protect against pulmonary failure and death . All 19 patients who required mechanical ventilation for early respiratory failure died [105,106]. Because of the increased risk of pulmonary decompensation, women with extensive pulmonary metastases should be observed in an intensive care setting during induction chemotherapy. In a rare experience at our institution, induction chemotherapy in a young woman with extensive pulmonary metastases resulted in an embolic phenomenon, leading to severe CNS morbidity (multiple cerebrovascular accidents), leading, in turn, to sudden brain death, despite the patient's having been free of CNS disease prior to therapy. Early detection and diligent follow-up in patients with molar disease are mandatory to improve survival.
Brain metastases pose a significant threat to the survival of patients with GTT, especially if the metastases appear while the patient is receiving chemotherapy. Although cerebral disease is observed clinically in only 7% to 28% of patients with choriocarcinoma [22,30,56,62], postmortem examinations demonstrate CNS involvement in as many as 40% of cases . This subset represents a significant fraction of patients who die of the disease.
Several investigators have examined the use of multimodality therapy in this group of patients. Weed et al reported a survival rate of 50% in patients with CNS involvement by choriocarcinoma who received multimodality therapy; disease-free periods of 12 to 120 months were achieved . In six of seven patients who survived, the interval from diagnosis to remission averaged 5.5 months. In the seven patients who died, the average duration of CNS involvement was 17 months. The addition of whole-brain irradiation and a change in the chemotherapeutic regimen to agents that cross the blood-brain barrier were successful in eradicating the tumor in three of seven patients who developed brain lesions while receiving therapy or who had a recurrence.
In 1983, Athanassiou et al reported results of a 23-year experience with choriocarcinoma involving the CNS at Charing Cross Hospital . Overall, 8.8% of 782 patients with choriocarcinoma who received chemotherapy had CNS metastases. Of these patients, 48% presented with CNS disease prior to treatment. Although 49% of patients who presented with CNS metastases enjoyed long-term survival, only 6% of the patients who developed CNS disease while on therapy survived. After 1974, the overall survival rate was 80% in patients who presented with CNS disease and 25% in patients who developed CNS disease after initiation of treatment. The improved outcome was attributed to early detection of CNS disease by measurement of the serum:CSF beta-hCG ratio, CT scans of the brain, CNS prophylaxis in patients identified as being at high risk for developing brain metastases, and the combination of systemic and intrathecal therapy used for CNS metastases. Radiotherapy did not appear to benefit patients whose disease was resistant to chemotherapy.
In 1987, Yordan et al reported a retrospective analysis of 70 cases of GTT involving the CNS . Of the 70 patients, half died before therapy was initiated. Of the remaining patients, 24% of those given chemotherapy alone survived, and 50% of patients given concurrent chemotherapy plus whole-brain irradiation achieved long-term remission. In the chemotherapy-plus-radiation group, none of the deaths was attributed to CNS disease.
Two different doses of radiation therapy were analyzed for effectiveness and toxicity. Among 10 patients receiving 3,000 cGy or more, there were no CNS disease recurrences. Of the seven patients receiving 2,500 cGy or less, one developed CNS relapse and achieved remission after further treatment. Resolution of CNS disease was documented by brain scan or autopsy in 50% of patients receiving 3,000 cGy or more but in only one of seven patients receiving 2,500 cGy or less. Yordan et al concluded that at the time of diagnosis of CNS disease, irradiation of the brain with 3,000 cGy given over 10 fractions should be initiated simultaneously with the start of chemotherapy.
The introduction of MRI has led to the diagnoses of even smaller CNS lesions than found in the past. Accordingly, the optimal integration of irradiation and chemotherapy remains to be defined. Likewise, select patients who have isolated CNS lesions and are refractory to chemotherapy should be evaluated by a neurosurgeon for possible resection.
Although the majority of patients with metastatic GTT will be cured, there remains a subset of patients who demonstrate persistent or recurrent disease after aggressive multiagent chemotherapy. These patients require special strategies. It is hoped that new therapeutic agents, such as the emerging class of topoisomerase-I inhibitors (camptothecin and its derivatives), angiogenesis inhibitors (TNP-470), and microtubule agents (paclitaxel [Taxol] and docetaxel [Taxotere]), will have activity in this disease. Strategies involving dose intensification with the use of growth factors (granulocyte colony-stimulating factor [filgrastim, Neupogen] and granulocyte-macrophage colony-stimulating factor [sargramostim, Leukine]) are reasonable approaches for patients with high-risk disease who do not respond quickly to conventional treatment. Monoclonal antibody-based diagnostic and therapeutic modalities are ideally suited for the identification and treatment of metastatic disease, because the tumor uniformly expresses a unique paternal antigen and has a rich vascular supply. It is hoped that emerging insights into the molecular mechanisms of GTT will lead to genetically based therapeutic strategies.
1. World Health Organization Scientific Group: Gestational Trophoblastic Disease. Technical Report Series No 692. Geneva, World Health Organization, 1983.
2. Lurain JR: Gestational trophoblastic tumor. Semin Surg Oncol 6:347–353, 1990.
3. Szulman AE, Surti U: The syndromes of hydatidiform mole: I. Cytogenic and morphologic correlations. Am J Obstet Gynecol 131:665–671, 1978.
4. Szulman AE, Surti U: The syndromes of hydatidiform mole: II. Morphologic evolution of the complete and partial mole. Am J Obstet Gynecol 132:20–27, 1978.
5. Lurain JR, Brewer JI, Torok EE, et al: Natural history of hydatidiform mole after primary evacuation. Am J Obstet Gynecol 145:591–595, 1983.
6. Miller JM, Surwit EA, Hammond CB: Choriocarcinoma following term pregnancy. Obstet Gynecol 53:207–212, 1979.
7. Driscoll SG: Placental site chorioma: The neoplasm of the implantation site trophoblast. J Reprod Med 29:821, 1984.
8. Goldstein DP: Worldwide controversies in gestational trophoblastic neoplasms. Int J Gynaecol Obstet 15:207–215, 1977.
9. Buckley JD: The epidemiology of molar pregnancy and choriocarcinoma. Clin Obstet Gynecol 27:153, 1984.
10. Bagshawe KD, Golding PR, Orr AM: Choriocarcinoma after hydatidiform mole: Studies related to the effectiveness of follow-up practice after hydatidiform mole. Br Med J 2:733, 1969.
11. Mazur MT, Kurman RJ: Choriocarcinoma and placental site trophoblastic tumor, in Szulman AE, Buchsbaum HJ (eds): Gestational Trophoblastic Disease, pp 45–68. New York, Springer-Verlag, 1987.
12. Wei PY, Ouyang PC: Trophoblastic diseases in Taiwan. Am J Obstet Gynecol 85:844, 1963.
13. Palmer JR: Advances in the epidemiology of gestational trophoblastic disease. J Reprod Med 39:155–162, 1994.
14. Yen S, MacMahon BB: Epidemiologic factors of trophoblastic disease. Am J Obstet Gynecol 101:126, 1968.
15. Hayashi K, Bracken MB, Freeman DH, et al: Hydatidiform mole in the United States (1970–1977): A statistical and theoretical analysis. Am J Epidemiol 115:67–77, 1982.
16. Martin PM: High frequency of hydatidiform mole in native Alaskans. Int J Gynaecol Obstet 15:395, 1978.
17. Matsuura J, Chui D, Jacobs PA, et al: Complete hydatidiform mole in Hawaii: An epidemiological study. Genet Epidemiol 1:171, 1984.
18. Graham IH, Fajardo AM: The incidence and morphology of hydatidiform mole in Abu Dhabi, United Arab Emirates. Br J Obstet Gynaecol 95:391, 1988.
19. Joint Project for the Study of Choriocarcinoma and Hydatidiform Mole in Asia: Geographic variation in the occurrence of hydatidiform mole and choriocarcinoma. Ann N Y Acad Sci 80:174, 1959.
20. Bagshawe KD, Dent J, Webb J: Hydatidiform mole in England and Wales, 1973–1983. Lancet II:673, 1986.
21. Teoh ES, Dagwood MY, Ratnam SS: Epidemiology of hydatidiform mole in Singapore. Am J Obstet Gynecol 110:415, 1971.
22. Bagshawe KD: Risk and prognostic factors in trophoblastic neoplasia. Cancer 38:1373–1385, 1976.
23. Berkowitz RS, Bernstein MR, Laborde O, et al: Subsequent pregnancy experience with gestational trophoblastic disease: New England Trophoblastic Disease Center, 1965–1992. J Reprod Med 39:228–232, 1994.
24. Acosta-Sison H: Statistical study of chorioepithelioma in the Philippine General Hospital. Am J Obstet Gynecol 58:125, 1949.
25. Bertini B: Epidemiology of hydatidiform mole in Israel: A study based on 113 patients. Int J Gynaecol Obstet 11:55, 1973.
26. Parazinni F, LaVecchia C, Mangili G, et al: Dietary factors and risk of trophoblastic disease. Am J Obstet Gynecol 158:93, 1988.
27. Berkowitz RS, Cramer DW, Bernstein MR, et al: Risk factors for complete molar pregnancy from a case-control study. Am J Obstet Gynecol 152:1016, 1985.
28. Bagshawe KD, Rawlings G, Pike MC, et al: The ABO group in trophoblastic neoplasia. Lancet I:553, 1971.
29. Cross JC, Werb Z, Fisher SJ: Implantation and the placenta: Key pieces of the development puzzle. Science 266:1508–1518, 1994.
30. Park WW, Lees JC: Choriocarcinoma: A general review with analysis of 516 cases. Arch Pathol 49:73–104, 205–241, 1950.
31. Hertig AT, Edmonds HW: Genetics of hydatidiform mole. Arch Pathol 30:260, 1940.
32. Driscoll SG: Trophoblastic growths: Morphologic aspects and taxonomy. J Reprod Med 26:2181, 1981.
33. Kajii T, Ohama K: Androgenetic origin of hydatidiform mole. Nature 268:633–655, 1977.
34. Morrow CP: Postmolar trophoblastic disease: Diagnosis, management, and prognosis. Clin Obstet Gynecol 27:211, 1984.
35. Lurain JR, Brewer JI: Invasive mole. Semin Oncol 9:174–180, 1982.
36. Kurman RJ, Main CS, Chen HC: Intermediate trophoblast: A distinctive form of trophoblast with specific morphological, biochemical, and functional features. Placenta 5:349–370, 1984.
37. Silva EG, Tornos C, Lage J, et al: Multiple nodules of intermediate trophoblast following hydatidiform moles. Int J Gynecol Pathol 12:324–332, 1993.
38. Lawler S, Fisher RA, Dent J: A prospective genetic study of complete and partial hydatidiform moles. Am J Obstet Gynecol 164:1270, 1991.
39. Jacobs PA, Wilson CM, Sprenkle JA, et al: Mechanisms of origin of complete hydatidiform moles. Nature 286:714–716, 1980.
40. Yamashita K, Wake N, Araki T, et al: Human lymphocyte antigen expression in hydatidiform mole: Androgenesis following fertilization by a haploid sperm. Am J Obstet Gynecol 135:597–600, 1979.
41. Wake N, Fujino T, Hoshi S, et al: The propensity to malignancy of dispermic heterozygous mole. Placenta 8:319, 1987.
42. Wallace DC, Surti U, Adams CW, et al: Complete moles have paternal chromosomes but maternal mitochondrial DNA. Hum Genet 61:145 1982.
43. Ohama K, Kajii T, Okamoto E, et al: Dispermic origin of XY hydatidiform moles. Nature 292:551–552, 1981.
44. Fisher RA, Povey S, Jeffries AJ, et al: Frequency of heterozygous complete hydatidiform moles, estimated by locus-specific minisatellite and Y chromosome-specific probes. Hum Genet 82:259, 1989.
45. Mutter GL, Pomponio R, Berkowitz RS, et al: Sex chromosome composition of complete hydatidiform moles: Relationship to metastasis. Am J Obstet Gynecol 168:1547, 1993.
46. McFadden DE, Kalousek DK: Two different phenotypes of fetuses with chromosomal triploidy: Correlation with parental origin of the extra haploid set. Am J Med Genet 38:535, 1991.
47. Roberts DJ, Mutter GL: Advances in the molecular biology of gestational trophoblastic disease. J Reprod Med 39:201–208, 1994.
48. Steller MA, Mok SC, Yeh J: Effects of cytokines on epidermal growth factor receptor expression by malignant trophoblast cells in vitro. J Reprod Med 39:209–216, 1994.
49. Guillemot F, Nagy A, Auerbach A, et al: Essential role of Mash-2 in extraembryonic development. Nature 371:333–336, 1994.
50. Cheung AN, Srivastava G, Pittaluga S, et al: Expression of c-myc and c-fms oncogenes in hydatidiform mole and normal human placenta. J Clin Pathol 46:204, 1993.
51. Sarkar S, Kacinski BM, Kohorn EI, et al: Demonstration of myc and ras oncogene expression by hybridization in situ in hydatidiform mole and in the BeWo chariocarcinoma cell line. Am J Obstet Gynecol 154:390, 1986.
52. Cameron B, Gown AM, Tamini HK: Expression of c-erb B2 oncogene product in persistent gestational trophoblastic disease. Am J Obstet Gynecol 170:1616–1621, 1994.
53. Sheridan E, Hancock BW, Goyns MH: High incidence of mutations of the p53 gene detected in ovarian tumours by the use of chemical mismatch cleavage. Cancer Lett 68:83, 1993.
54. Cheung AN, Srivastava G, Chung LP, et al: Expression of the p53 gene in trophoblastic cells in hydatidiform moles and normal human placentas. J Reprod Med 39:223–227, 1994.
55. Persaud V, Ganjei P, Nadji M, et al: Cell proliferation activity and mutation of p53 suppressor gene in human gestational trophoblastic disease. West Indian Med J 42:142–143, 1993.
56. Berkowitz RS, Goldstein DP: Pathogenesis of gestational trophoblastic neoplasms. Pathol Annu 11:391–411, 1981.
57. Yuen BH, Carron W, Benedet JL, et al: Plasma beta-subunit human chorionic gonadotropin assay in molar pregnancy and choriocarcinoma. Am J Obstet Gynecol 127:711–712, 1977.
58. Amir SM, Osathanondh R, Berkowitz RS, et al: Human chorionic gonadotropin and thyroid function in patients with hydatidiform mole. Am J Obstet Gynecol 150:723, 1984.
59. Goldstein DP, Berkowitz RS: Current management of complete and partial molar pregnancy. J Reprod Med 39:139–146, 1994.
60. Mutch D, Soper JT, Baker ME, et al: Role of computed axial tomography of the chest in staging patients with nonmetastatic gestational trophoblastic disease. Obstet Gynecol 68:348–352, 1986.
61. Kumar J, Ilancheran A, Ratnam SS: Pulmonary metastases in gestational trophoblastic disease: A review of 97 cases. Br J Obstet Gynaecol 95:70, 1988.
62. Stilp TJ, Bucy PC, Brewer JI: Cure of metastatic choriocarcinoma of the brain. JAMA 221:276–279, 1972.
63. Bakri Y, Berkowitz RS, Goldstein DP, et al: Brain metastases of gestational trophoblastic tumor. J Reprod Med 39:179–183, 1994.
64. Pastorfide GB, Goldstein DP, Kosasa TS: The use of a radioimmunoassay specific for human chorionic gonadotropin in patients with molar pregnancy and gestational trophoblastic disease. Am J Obstet Gynecol 120:1025–1028, 1974.
65. Vaitukaitis JL, Braunstein GD, Ross GT: A radioimmunoassay which specifically measures human chorionic gonadotropin in the presence of human luteinizing hormone. Am J Obstet Gynecol 113:751–758, 1972.
66. Wehmann RE, Ayala AR, Birken S: Improved monitoring of gestational trophoblastic neoplasia using a highly sensitive assay for urinary human chorionic gonadotropin. Am J Obstet Gynecol 140:753–757, 1981.
67. Kenimer JG, Hershman JM, Higgins HP: The thyrotropin in hydatidiform moles is human chorionic gonadotropin. J Clin Endocrinol Metab 40:482–491, 1975.
68. Koonongs PP, Schalerth JB: CA125: A marker for persistent gestational trophoblastic disease? Gynecol Oncol 49:240–242, 1993.
69. Athanassiou A, Begent RHL, Newlands ES, et al: Central nervous system metastases of choriocarcinoma: Twenty-three years' experience at Charing Cross Hospital. Cancer 52:1728–1735, 1983.
70. Bagshawe KD, Harland J: Immunodiagnosis and monitoring of gonadotropin-producing metastases in the central nervous system. Cancer 39:112, 1976.
71. Bagshawe KD: Choriocarcinoma: The Clinical Biology of the Trophoblast and Its Tumors. London, Edward Arnold, 1969.
72. Carter J, Fowler J, Carlson J, et al: Transvaginal color flow Doppler sonography in the assessment of gestational trophoblastic disease. J Ultrasound Med 12:595–599, 1993.
73. Dubuc-Lissoir J, Sweizig S, Schlaerth JB, et al: Metastatic gestational trophoblastic disease: A comparison of prognostic classification systems. Gynecol Oncol 45:40–45, 1992.
74. Hammond CB, Weed JC Jr, Currie JL: The role of operation in the current therapy of gestational neoplastic disease. Am J Obstet Gynecol 136:844–858, 1980.
75. Soper JT: Surgical therapy for gestational trophoblastic disease. J Reprod Med 39:168–174, 1994.
76. Curry SL, Hammond CB, Tyrey L, et al: Hydatidiform mole: Diagnosis, management, and long-term follow-up of 347 patients. Obstet Gynecol 45:1, 1975.
77. Song HZ, Wu PH: Reevaluation of 5-fluorouracil as a single therapeutic agent for gestational trophoblastic neoplasms. J Obstet Gynecol 150:69–75, 1984.
78. Wong LC, Choo YC, Ma HK: Primary oral etoposide therapy in gestational trophoblastic disease. Cancer 58:14–17, 1986.
79. Goldstein DP: Prophylactic chemotherapy of molar pregnancy. Obstet Gynecol 38:817–822, 1971.
80. Goldstein DP: Prevention of gestational trophoblastic disease by use of actinomycin D in molar pregnancies. Obstet Gynecol 43:475–479, 1974.
81. Kim DS, Hyung M, Kyung TK, et al: Effects of prophylactic chemotherapy for persistent trophoblastic disease in patients with complete hydatidiform mole. Obstet Gynecol 67:690–694, 1986.
82. Sand PK, Lurain JR, Brewer JI: Repeat gestational trophoblastic disease. Obstet Gynecol 63:140–144, 1984.
83. Walden PAM, Bagshawe KD: Pregnancies after chemotherapy for gestational trophoblastic tumors. Lancet II:1241, 1979.
84. Wenzel LB, Berkowitz RS, Robinson S, et al: Psychological, social, and sexual effects of gestational trophoblastic disease on patients and their partners. J Reprod Med 39:163–167, 1994.
85. Lurain JR: Chemotherapy of gestational trophoblastic disease, in Deppe G (ed): Chemotherapy of Gynecologic Cancer, pp 273–301. New York, Alan R Liss, 1990.
86. Berkowitz RS, Goldstein DP, Bernstein MR: Ten years' experience with methotrexate and folinic acid as primary therapy for gestational trophoblastic disease. Gynecol Oncol 23:111–118, 1986.
87. Gleeson NC, Finan MA, Fiorica JV, et al: Nonmetastatic gestational trophoblastic disease: Weekly methotrexate compared with 8-day methotrexate-folinic acid. Eur J Gynaecol Oncol 14:461–465, 1993.
88. Schlaerth JB, Morrow CP, Nalick RH, et al: Single-dose actinomycin D in the treatment of postmolar trophoblastic disease. Gynecol Oncol 19:53–56, 1984.
89. Twiggs LB: Pulse actinomycin D scheduling in nonmetastatic gestational trophoblastic neoplasia: Cost-effective chemotherapy. Gynecol Oncol 16:190–195, 1983.
90. Jones WB: Management of low-risk metastatic gestational trophoblastic disease. J Reprod Med 26:213–217, 1981.
91. Hammond CB, Borchert LG, Tyrey L, et al: Treatment of metastatic trophoblastic disease: Good and poor prognosis. J Obstet Gynecol 115:451–457, 1973.
92. Lurain JR, Brewer JI: Treatment of high-risk gestational trophoblastic disease with methotrexate, actinomycin D, and cyclophosphamide chemotherapy. Obstet Gynecol 65:830–836, 1985.
93. Gordon AN, Gershenson DM, Copeland LJ, et al: High-risk metastatic gestational trophoblastic disease. Obstet Gynecol 65:550–556, 1985.
94. Berkowitz RS, Goldstein DP, Bernstein MR: Modified triple chemotherapy in the management of high-risk metastatic gestational trophoblastic tumors. Gynecol Oncol 19:173–181, 1984.
95. Begent RHJ, Bagshawe KD: The management of high-risk choriocarcinoma. Semin Oncol 9:198–203, 1982.
96. Curry SL, Blessing JA, Disaia PJ, et al: A prospective randomized comparison of methotrexate, actinomycin D, and chlorambucil (MAC) versus modified Bagshawe regimen in `poor-prognosis' gestational trophoblastic disease. Obstet Gynecol 73:357–362, 1989.
97. Bagshawe KD: Treatment of high-risk choriocarcinoma. J Reprod Med 29:813–820, 1984.
98. Newlands ES: New chemotherapeutic agents in the management of gestational trophoblastic disease. Semin Oncol 9:230, 1982.
99. Gordon AN, Kavanagh JJ, Gershenson DM, et al: Cisplatin, vincristine, and bleomycin combination therapy in resistant gestational trophoblastic disease. Cancer 58:1407–1410, 1986.
100. Surwit EA, Childers JM: High-risk metastatic gestational trophoblastic disease: A new dose-intensive, multiagent chemotherapeutic regimen. J Reprod Med 36:45–48, 1991.
101. Li-Pai C, Shu-Mo C, Jian-Xuan F, et al: PEBA regimen (cisplatin, etoposide, bleomycin, and Adriamycin) in the treatment of drug-resistant choriocarcinoma. Gynecol Oncol 56:231–234, 1995.
102. Sutton GP, Soper JT, Blessing JA, et al: Ifosfamide alone and in combination in the treatment of refractory malignant gestational trophoblastic disease. Am J Obstet Gynecol 167:489–495, 1992.
103. Lotz JP, Andre T, Donsimoni R, et al: High dose chemotherapy with ifosfamide, carboplatin, and etoposide combined with autologous bone marrow transplantation for the treatment of poor-prognosis germ cell tumors and metastatic trophoblastic disease in adults. Cancer 75:874–885, 1995.
104. Dubeshter B, Berkowitz RS, Goldstein DP, et al: Analysis of treatment failure in high- risk metastatic gestational trophoblastic diesease. Gynecol Oncol 29:199, 1988.
105. Kelly MP, Rustin GJS, Ivory C, et al: Respiratory failure due to choriocarcinoma: A study of 103 dyspneic patients. Gynecol Oncol 38:149, 1990.
106. Bakri YN, Berkowitz RS, Khan J, et al: Pulmonary metastases of gestational trophoblastic tumor: Risk factors for early respiratory failure. J Reprod Med 38:175–178, 1994.
107. Rustin GJS, Bagshawe KD: Gestational trophoblastic tumours. CRC Crit Rev Oncol Haematol 3:103, 1985.
108. Gilbert HA, Kagan AR: Incidence, detection, and evaluation without histologic confirmation, in Weiss L (ed): Fundamental Aspects of Metastases, pp 385–405. Amsterdam, North Holland, 1976.
109. Weed JC, Hammond CB: Cerebral metastatic choriocarcinoma: Intensive therapy and prognosis. Obstet Gynecol 55:89–94, 1980.
110. Yordan EL Jr, Schlaerth JB, Gaddis O, et al: Radiation therapy in the management of gestational choriocarcinoma metastatic to the central nervous system. Obstet Gynecol 69:627–630, 1987.