Gestational trophoblastic disease: What have we learned in the past decade? Walter B. Jones, MD New York, New York Hydatidiform moles can be classified into two distinct entities: complete and partial. Complete moles are androgenetically derived, are characterized by a predominance of the 46,XX karyotype, and are associated with a significantly higher incidence of malignant sequelae than are partial moles. A subset of complete moles resulting from dispermy are heterozygous (46,XY) and appear to be associated with the greatest risk of malignancy. Free subunits of human chorionic gonadotropin, 13 core fragments, and acidic variants of human chorionic gonadotropin can now be readily measured by radioimmunoassays, fluoroimmunoassays, or isoelectric focusing techniques. Determination of these markers can be of value not only in monitoring response to therapy, but also as prognostic indicators. A subgroup of patients, traditionally classified as "high risk" or poor prognosis, who show poor responses to conventional combination chemotherapy have recently been designated as "ultra-high-risk" patients. They require more complex chemotherapy and possibly surgery and radiotherapy to achieve remission. The development of chemotherapy over the past decade, including the introduction of etoposide and cisplatin into current protocols and the treatment of patients with advanced gestational trophoblastic disease, are considered in this review. (AM J OBSTET GVNECOL 1990;162:1286-95.)

Key words: Hydatidiform mole, trophoblastic disease, chemotherapy One year ago when I began to think about a topic for this address, I remembered once reading that such a presentation should consider an area of concern to the audience about which the speaker is an expert. Because an expert is defined as a person with a high degree of knowledge or skill in a particular field (i.e., an authority), it occurred to me that whatever topic I settled on, in this the 125th anniversary celebration of this Society, could be considered pretentious. On the other hand, it seemed to me that the definition of experience-one's personal knowledge derived from participation or observation-might suggest a possible if not plausible subject. And so this presentation was decided upon. I would be remiss if at this time I failed to express my gratitude to Dr. John L. Lewis, Jr., who introduced me to the study of trophoblastic tumors when I became a Fellow at Memorial Hospital in New York in 1968. Indeed, he continues to provide expert advice. In fact, his Presidential address given before this Society 10 years ago, "Treatment of Metastatic Gestational Trophoblastic Neoplasms," was an important resource in my preparation for this article. This discussion will make no attempt to address all of the advances made in the study and treatment of From the Gynecology Service, Department of Surgery, Memorial Sloan-Kettering Cancer Center. Presidential address, presented to the New York Obstetrical Society, May 10, 1989. Reprint requests: Walter B. jones, MD, 1275 York Ave., New York, NY 10021. 611119593

1286

trophoblastic tumors in the past decade, as clinical successes over this time have been matched by numerous advances in the laboratory. I will focus instead on four areas in which new knowledge appears to have had an impact on current management strategies and probably on patient survival: the biology of hydatidiform moles, assays for human chorionic gonadotropin (hCG), the drugs etoposide (VP-16) and cisplatin, and approaches to the management of the high-risk patient. For purposes of organization, recent new insights about the biology of hydatidiform moles will be discussed first. We have learned that hydatidiform moles may be classified into two distinct entities (complete and partial moles) and that certain cytogenetic features of a subset of complete moles may be predictive of malignant sequelae. Because hCG functions as a specific tumor marker in patients with gestational trophoblastic disease, a brief review of some of the newer hCG assays that allow more precise measurement of the hormone and its subunits will then be presented. We have learned that normal levels of hCG in the serum l3-hCG assay may not always signify the absence of viable trophoblast and that the level of production of the free 13 subunit and certain variants of hCG may have prognostic significance. Following this, two new chemotherapeutic agents, etoposide (VP-16) and cisplatin, will be briefly highlighted. These agents have been integrated into current treatment protocols and have had a significant impact on prognosis. Finally, we have learned that there exists among patients traditionally classified as "highrisk" or poor prognosis, a subgroup of patients who

Volume 162 Number 5

show poor responses to treatment with conventional combination chemotherapy. These "ultra-high-risk" patients require more complex chemotherapy and possibly surgery and radiotherapy to achieve remission. The biology of hydatidiform moles Hydatidiform mole, invasive mole, and choriocarcinoma are neoplasms that originate in the fetal chorion. According to Hertz, I the benign hydatidiform mole represents the beginning of a disease continuum, whereas the highly malignant choriocarcinoma represents the end of the spectrum. The term hydatid, first used by Aetius of Amida in the 6th century AD,2 might have been applicable six centuries later, when the Countess Margaret of Flanders is said to have brought forth at one birth 365 infants, of whom 182 were solemnly baptized John and 182 Elizabeth, and the odd one, adjudged a hermaphrodite, was buried without baptism.' However uncertain the origin of this story, it nevertheless suggests that interest in the sex of hydatidiform moles is more than a contemporary curiosity. Indeed, it has been known for many years, on the basis of the presence of sex chromatin, that most hydatidiform moles are female. In a collected series of hydatidiform mole sex chromatin studies, Lawler' reported a 90% overall frequency of chromatin-positive cases. The remaining 10% of cases were assumed to be male. Such findings have always constituted an enigma because one would expect a 50: 50 ratio of chromatinpositive to chromatin-negative cases. In 1976 Vassilakos and Kajii 5 classified hydatidiform moles for the first time into two distinct entities (complete and partial moles) and indicated that certain morphologic and cytogenetic features of moles may be predictive of subsequent proliferative sequelae. They observed that complete moles were characterized by the exclusive occurrence of the female karyotype, 46,XX. There was an absence of normal villi; the lack of an embryo, cord, or amniotic membranes; and the presence of marked trophoblastic hyperplasia and anaplasia. Partial moles were associated with a fetus, cord, and/or amniotic membrane; the absence of trophoblastic cell anaplasia; marked hyperplasia; and karyotypes that included trisomies, triploidy, and tetraploidy. Significantly, according to their findings on follow-up of the cases, only complete moles had the potential for malignant sequelae. In 1978 Szulman and Surti6 reported that hydatidiform moles could be divided into two distinct syndromes. They described the complete or classic mole without a discernable embryo or fetus and with a diploid karyotype, and the partial mole in which there was an ascertainable fetus and a triploid karyotype. Complete moles appeared clinically as missed abortions usually during the second trimester, whereas partial moles

Gestational trophoblastic disease

1287

presented in the first trimester or rarely as spontaneous abortions. Subsequent cytogenetic, morphologic, and clinical studies have firmly established that complete and partial moles may be distinguished not only with respect to the predominance of the 46,XX karyotype of complete moles, but also by the greater risk of malignancy after a complete mole. 7.6 Nevertheless, experience with patients requiring chemotherapy after evacuation of a partial mole emphasizes the need to monitor these patients closely, as their risk may not be insignificant. g · 12 The technique of fluorescent staining of chromosomes was developed in 1971 by Casperson et aI., 13 who demonstrated that by using quinacrine staining (Q banding), each chromosome in the human karyotype could be made to display a characteristic banding pattern that allowed its precise identification. Several other fluorescent banding techniques have been subsequently developed: Geisma staining (G banding) and staining the centromere region preferentially using Geisma stain (C banding) and Acridine orange (R banding). Although most bands are constant for all individuals, secondary constrictions on certain chromosomes and the centromeric regions vary in size and staining intensity from person to person. Studies of chromosomal banding polymorphisms by Kajii and Ohama l1 in 1977 revealed that complete moles had only paternal chromosomes and that all of their cases had a 46,XX karyotype (androgenesis). Their findings were subsequently confirmed by others, who again documented that a majority (92% to 96% of complete moles) had a 46,XX karyotype and approximately 4% to 8% had a 46,XY karyotype. 15. 16 According to Yamashita et al.,17 the mechanism of androgenesis involves duplication of the haploid genome (23,X) of a single sperm in the absence of maternal chromosomes due to either inactivation or absence of the egg nucleus. Because a cell must have at least one X chromosome to survive, the prevalence of the 46,XX karyotype can therefore be accounted for by the non viability of the YY conceptus. The subset of complete moles having a 46,XY karyotype has since been shown by Surti et al. 18 and others l9 . 22 to result from fertilization of the "empty ovum" by two spermatozoa (dispermy). Polymorphic allozymes are genetically determined variants that are identifiable by electrophoresis. These polymorphisms are inherited in a Mendelian fashion and can be confirmed by segregation analysis among different members of a family. Lawler et al!3. 21 and others 25 have used this technique to document the androgenetic origin of complete moles. An analysis of the inheritance of the allozymes glucose 6-phosphate dehydrogenase and esterase-D is an excellent illustration of this approach to determining the parentage of moles. 26 Human leukocyte antigen typing of molar and

1288

Jones

parental cells makes use of yet another polymorphic system to determine the genetic origin of hydatidiform moles."7 By far the most efficient method of determining the origin of moles is that of restriction fragment length polymorphisms. Restriction enzymes are site-specific endonucleases that recognize six or four base sequences and cut deoxyribonucleic acid (DNA) at that point. The technique developed by Southern in 1975, and known as the Southern blot, can reveal a loss or a gain of a restriction site in the DNAs of different individuals. 28 DNA polymorphisms have been used successfully in the diagnosis of many genetic disorders and more recently in determining the orgin of complete moles. For example, Wallace et aFY used restriction fragment length polymorphisms to study the parentage of three moles and found in all three cases that the mitochondrial DNA was, as expected, maternally derived. The chromosomes of the moles, on the other hand, were exclusively paternal in origin, proving that the moles arose from fertilization by one or more sperms with an anucleate ovum. In a similar study of six complete moles, Edwards et al. 3t ) once again found the mitochondrial DNA to be maternal in origin, with no contribution from the sperm, whereas the nuclear genome was shown to be exclusively paternal in five cases. The difference in the mechanism leading to the formation of complete moles has raised the question as to whether the risk of subsequent malignancy is influenced by the mole's zygosity. Although reports in the literature are conflicting, evidence suggests that a higher frequency of malignancy follows a heterozygous complete mole. The reported magnitude of the risk ranges between 20% and 60%, compared with an average of less than 20% for homozygous moles.IY.:lJ.:l:' For this reason, some authors have recommended prophylactic treatment of patients with heterozygous moles. It has been suggested that malignancy after a mole could be the result of a recessive mutation of genetic factors that control cell growth, and when homozygous for the mutation may escape normal control mechanisms. II The apparent high malignancy rate of dispermic heterozygous XY moles seems to be inconsistent with this view. Hopefully studies in the next decade will answer the question of whether the rate of malignancy is in fact higher after XY moles and provide some insight into the mechanism of malignant transformation. Recent demonstration of the myc and ras oncogene expression by in situ hybridization studies of hydatidiform moles may contribute new knowledge and answers to such questions in this fertile area of investigation."I. "5 Assay of human chorionic gonadotropin

Gestational trophoblastic neoplasms are unique III that they consistently produce heG when viable tumor is present. heG is a glycoprotein consisting of two dissimilar noncovalently linked polypeptide subunits with

May 1990 Am J Obstet Gynecol

attached carbohydrate side chains.""' 37 The subunits (designated u and 13) are essentially without biological activity as individual free subunits. The primary structure of u subunits of all glycoprotein hormones are nearly identical and consist of 89 to 92 amino acid residues in identical sequences. The 13 subunits, on the other hand, have similar yet distinctive amino acid contents and sequences and confer both immunologic and biologic specificities."8 It is noteworthy that 80% of the first 115 amino terminal residues of the heG 13 subunit are identical to those ofthe human leutinizing hormone (hLH) 13 subunit. This similarity in structure accounts for the well-known cross-reaction between heG and hLH in all biologic and nonspecific radioimmunoassays. The 13 subunit of heG, however, also differs by containing an additional 30 amino acid residues and a carboxy terminal sequence that is distinctive.'Y.1I In 1972, Vaitukitis et al. 42 developed a radioimmunoassay specific for heG and its 13 subunit in serum or plasma, which allowed clinicians to differentiate low levels of heG for pituitary hLH. Because this assay utilizes an antiserum generated against heG 13, it is commonly referred to as a 13 subunit assay. This designation, however, is misleading because it implies that only heG 13 is being measured, when in fact both heG and heG 13 are detected. Moreover, because the plasma half-life of heG is approximately 100 times longer than that of heG 13, the predominant molecular species measured is heG. The value of this assay in monitoring patients with gestational trophoblastic disease has been confirmed by many investigators:"·I' In the late 1970s many investigators 4fi .oJY developed new, highly sensitive urinary radioimmunoassays for heG that could detect heG when serum or plasma radioimmunoassays were negative. When controlled for low values of heG detectable in the urine of normal subjects, such assays clearly enabled the identification of persistent tumor activity and thus the need for further therapy. Also during this period, several investigators observed that the presence of free u and free 13 subunits in serum, urine, or tumor extracts were an expression of more severe malignancy. For example, Dawood et al. 50 reported in 1977 that u subunit levels increased while heG and heG 13 levels decreased or became undetectable in patients with choriocarcinoma in whom cerebral metastases eventually developed. In 1980 physicians at the Southeastern Trophoblastic Disease Center recommended the routine measurement of the u subunit in patients successfully treated to identify patients who require additional chemotherapy or more intensive follow-up.51 In the 1980s, as a result of rapidly expanding monoclonal antibody technology, radioimmunoassays have become readily available that distinguish free l3-subunit heG in the presence of intact heG. With these assays it has been possible to demonstrate that the ratio of free hCG 13 to total heG 13 is markedly higher in patients

Volume 162 Number 5

who develop high-risk metastatic gestational trophoblastic disease than in patients with hydatidiform moles or persistent trophoblastic disease in the low-risk category.52.53 Such findings suggest that excessive production of the free 13 subunit of hCG may be useful in identifying patients who should be assigned to the highrisk category. Urine from pregnant women contains several peptides related to hCG; the principle moieties include hCG, its free subunits, and a population of fragments of the hCG 13 subunit called 13 core molecules. These fragments are also present in the urine of patients with gestational trophoblastic disease. Potential applications of the 13 core radioimmunoassay in patients with gestational trophoblastic disease include monitoring response to therapy. In fact, the presence of these molecules in some patients who have undetectable serum and urine hCG may be the sole marker indicating persistent disease.54 In the last 2 to 3 years an intense amount of research activity has centered around the application of highly sensitive immunoassays for the 13 core fragment in the study of a variety of nontrophoblastic malignancies. Detectable levels of these fragments have been reported in patients with cervical cancer (70%), ovarian cancer (73%), and endometrial cancer (77%).55-57 For completeness, the recently available fluoroimmunoassays for hCG using a two-monoclonal antibody technique are very rapid and highly sensitive. These assays, in which one antibody is labeled with europium (whose fluorescence can be measured with very high sensitivity and low background by timeresolved fluorometry), have the distinct advantage of not requiring laboratory equipment for the handling of isotopes. 58. 59 Similarly, highly sensitive and specific "sandwich-type" enzyme immunoassays, in which two monoclonal antibodies are directed against different epitopes of the hCG molecule, are now available that utilize enzymes such as l3-galactosidase or alkaline phosphatase as labels instead of isotopes. 6o -62 Finally, it has been recently observed that patients with advanced choriocarcinoma contain in their sera unique acidic variants of hCG in levels that are significantly higher than those in pregnant women or patients with hydatidiform mole or invasive mole. In a group of five patients with choriocarcinoma reported by Yzaki et al.,63 four had isoelectric focusing analysis of their sera that showed high levels of these variants before treatment. None of these patients responded to chemotherapy and all died within 2 years after diagnosis. It appears that determining these variants may be useful not only in evaluating the efficacy of therapy but also as prognostic indicators. Etoposide and cisplatln New drugs and drug combinations have become available since the documentation three decades ago of the effectiveness of methotrexate and actinomycin D

Gestational trophoblastic disease

1289

in treating gestational trophoblastic disease.'''-66 Multiple agents with different spectrums of toxicity and different mechanisms of action are currently administered in an effort to diminish morbidity and reduce the patient's risk of developing a drug-resistant tumor. During the past decade two new chemotherapeutic agents, VP-16 and cisplatin, have been shown to be among the most active agents yet tested for the treatment of trophoblastic tumors. Podophyllotoxin is an extract derived from the roots of plants belonging to the genus Podophyllum. This agent was found as early as 1946 to have antitumor activity, but excessive toxicity prevented its clinical use. Modification of the molecule, however, yielded two compounds, VP-16-213 (etoposide) and VM-26 (tenoposide), which are less toxic and have cytotoxic activity against a variety of tumors. 67 Etoposide induces dose-dependent breaks in singlestranded DNA, an action thought to come about through the interaction with Topoisomerase II, which can reversibly cleave one or both DNA strands. fi8 69 Topoisomerase II forms a stabilized complex with DNA in the presence of etoposide, with the resulting strand breaks occurring in proportion to the etoposide concentration and does not involve binding of the drug to DNA.70 Renal clearance accounts for approximately 36% of an etoposide dose, with the fate of the remaining percentage unclear, although 16% of a dose is excreted in bile l and metabolism of the drug is thought to account for part of its elimination. Etoposide was introduced into clinical trials in the United States in the early 1970s. At the doses commonly used, the only frequent toxicity observed is myelosuppression, which is dose limiting. 67 Nausea and vomiting are mild and peripheral neuropathy is rare. Although mucositis and diarrhea are common, they are not usually severe. In 1977 Bagshawe'2 observed that within 2 weeks of its first use, it was clear that VP-16-213 was a powerful drug in gestational choriocarcinoma. According to Newlands and Bagshawe,73 the main drawback with etoposide is the marked but reversible alopecia that is induced in almost all patients. Cisplatin was discovered serendipitously by Dr. Barrett Rosenberg in the 1960s. He observed that an alternating current passed through a platinum electrode inhibited bacterial growth. 74 The cis isomer, cisplatin, found in the bacterial medium is now considered to be one of the most important new chemotherapeutic agents of the past decade. It is also one of the most toxic. Cisplatin, in contrast to its trans isomer, destabilizes DNA causing inter- and intrastrand cross-links inhibiting DNA replication." The most important doselimiting toxicity is renal dysfunction, although ototoxicity, neurotoxicity, and vomiting can be severe. 76 Several investigators, however, have shown that hyperhydration and the administration of hypertonic saline solution allows high doses of cisplatin to be given with-

1290 Jones

out a marked decline in renal function. 77 Phase I clinical trials of cisplatin were begun in the United States in 1971 and the drug was approved for the treatment of malignant disease in 1978. Its use in drug-resistant gestational choriocarcinoma was begun at Charing Cross Hospital in 1976, where the efficacy of cisplatin was most evident when the drug was used in combination chemotherapy regimens. In view of the side effects, however, some authors continue to limit the use of this drug to patients who have become resistant to all other cytotoxic chemotherapy. The role of VP-16 and cisplatin in current therapy will be considered later in this presentation. Classification and management of the high-risk patient

That certain patients with metastatic disease were at high risk of treatment failure when treated with singleagent chemotherapy consisting of methotrexate or actinomycin D was reported by Ross et al. 7S in 1965. These patients had a long disease history and markedly elevated gonadotropin excretion. The site of metastases was subsequently noted to be an important risk factor. Over the years various classification systems have been developed to assess a patient's risk of treatment failure, but none has been universally applied. Because treatment results and treatment protocols (to be discussed later) are based on differing classifications, it may be useful at this time to review briefly some of the more commonly used systems. The clinical classification used by investigators at the Southeastern Trophoblastic Disease Center separates patients with metastatic disease into "good" and "poor" prognostic categories. 79 Anyone or more of the following factors assigns a patient to the "poor prognosis" group: (1) pretreatment hCG titer> 100,000 IU /24 hr in urine or >40,000 mIU/ml in serum; (2) period longer than 4 months since the antecedent pregnancy; (3) antecedent term pregnancy; (4) brain and/or liver metastases; and (5) prior chemotherapeutic failure. Other patients with metastatic disease are placed in the "good prognosis" category. Using similar criteria, the classification of patients with metastatic disease at Memorial Hospital separates patients into low-, moderate-, and high-risk groups.so In this classification, an antecedent term pregnancy in the absence of other high-risk factors does not place a patient in the highrisk group. A more comprehensive scoring system has been devised by Bagshawe,s' in which a variety of factors affecting prognosis are assigned numeric values. The totals are then used to divide patients into low- (less than 60), medium- (60 to 90), and high- (100 or above) risk groups. This classification, which is difficult to apply in some hospitals, can identify patients with a good prog-

May 1990

Am J Obstet Gyneco1

nosis who in other systems would be categorized as having a poor prognosis. At the same time, however, patients with nonmetastatic disease who are known to have an excellent prognosis may be classified, at least theoretically, into a higher risk category. More recently, the combination of anatomic staging and prognostic scoring used by the World Health Organization has gained in popularity.s2 Retrospective analysis of survival data using this scoring system has enabled the identification of a subset of patients within the traditional high-risk group who account for most deaths from gestational trophoblastic disease. These patients have a prognostic score of eight or greater and have been recently designated as ultra-high-risk patients. The score is obtained by adding individual scores for each prognostic factor: age, type of antecedent pregnancy, interval between the end of the antecedent pregnancy and the start of chemotherapy, hCG level, ABO groups (female X male), size of largest tumor, site of metastases, number of metastases, and prior chemotherapy. To decrease the incidence of drug resistance, the CHAMOMA protocol (consisting of cyclophosphamide, hydroxyurea, actinomycin D, methotrexate, oncovin, melphalan, and adriamycin) was devised by Bagshawe in the early 1970s for patients classified according to his scoring system as high-risk cases. The last two drugs in this 8- to 9-day schedule, melphalan and adriamycin, were found to be excessively myelosuppressive and were later replaced by cyclophosphamide and actinomycin D, resulting in the well known CHAMOCA regimen. With the use of CHAMOCA, a remission rate of 75% (24 of 32) was achieved in highrisk patients. 72 A retrospective analysis of high-risk patients treated at that institution with single-agent methotrexate and folinic acid alone as first therapy showed that only 29% of patients survived. The CHAMOCA regimen was also found to be effective in patients who had developed resistance after less complex combination therapy. This protocol and others to follow are being reviewed to illustrate the evolution and complexity of modern chemotherapy in the management of patients with advanced gestational trophoblastic disease. The importance of administering appropriate initial therapy based on a patient'S risk of developing a drug-resistant tumor was documented by Hammond et al. B3 in 1975, who reported that 70% of patients with a poor prognosis treated initially with combination chemotherapy achieved remission, compared with 14% of patients who received combination chemotherapy after resistance to single-agent therapy had developed. Similar results have been reported by others, with the result that combination chemotherapy has become the standard treatment of high-risk patients. s4 Experienced clinicians are now well aware of the fact that, although

Volume 162 Number 5

an initial response to single-agent chemotherapy may be observed in many high-risk patients, the response is not likely to be sustained and the use of combination chemotherapy after resistance has developed is not likely to be successful. Over the past decade, at least in the United States, the combination regimen consisting of methotrexate, actinomycin D, and an alkylating agent (MAC) has been the most widely used treatment for patients with highrisk or poor-prognosis gestational trophoblastic disease. Overall, this regimen has produced survival rates of 70% to 80%. Thus neither the MAC or the CHAMOCA regimen is adequate treatment for all high-risk patients. Further evidence that these regimens are less than optimal was recently documented in a Gynecologic Oncology Group study reported by Curry et al. B5 In this randomized comparison of methotrexate, actinomycin D, and chlorambucil (MAC) versus methotrexate, actinomycin D, cyclophosphamide, doxorubicin, melphalan, hydroxyurea, and vincristine (modified CHAMOMA) in poor-prognosis metastatic gestational disease, 16 (73%) of the 22 patients treated with MAC attained remission with their primary therapy, while 13 (65%) of the 20 patients receiving modified CHAMOMA did so. Overall remission rates using secondary therapies for the MAC and modified CHAMOMA regimens in this study were 95% and 70%, respectively. The MAC regimen, however, was clearly associated with less toxicity, with 9% of patients experiencing life-threatening hematologic toxicity compared with 44% in the modified CHAMOMA group. To reduce the incidence of primary treatment failure in high-risk patients, Berkowitz et al. 86 devised the MAC III regimen and achieved a 68.4% primary remission rate and a 84.2% overall remission rate. The MAC III regimen uses a higher dosage of methotrexate and actinomycin D and is considered to be associated with acceptable levels of hematologic and hepatic toxicity. Because this regimen induces remissions after shorter exposure to chemotherapy, it has been recommended as a suitable alternative to standard MAC for high-risk patients. Despite the significant improvement in survival achieved by the initial use of combination chemotherapy in high-risk or poor-prognosis patients, the relative lack of success in altering the outcome in many patients once drug resistance has occurred continues to be one of the major challenges to clinicians. A modification of the CHAMOMA protocol, in which the drugs in the original protocol were used in a different sequence with additional vincristine, was used at Memorial Hospital to treat eight patients who were resistant to triple therapy. All of the patients experienced an initial response, but only two patients achieved sustained remission. In 1980 Surwit and associates modified the Bagshawe

Gestational trophoblastic disease

1291

CHOMOCA protocol (essentially by the incorporation of additional actinomycin D) and reported that five of six patients who had developed resistance to triple therapy achieved remission. However, in a subsequent report from the same institution by Weed et al.,"7 this protocol achieved a remission rate of only 56%. While other drugs and drug combinations have been investigated during the past decade, no clear therapeutic advance was documented until VP-I6 became available. Newlands and Bagshawe88 initially reported on two patients with drug-resistant choriocarcinoma who responded to VP-I6 given after a bleomycin infusion. After this, nine responses (defined as greater than log fall in serum hCG) and seven improvements (greater than 50% fall in serum hCG) were achieved in 26 drug-resistant patients treated with VP-I6 alone. Among a total group of 31 patients, there were II (35%) responses and nine (29%) improvements with VP-I6 alone or in combination with methotrexate and vincristine or after a bleomycin infusion. 89 These investigators integrated VP-I6 into the primary combination chemotherapy for patients in the high-risk group in 1979. When VP-I6 was given in a 5-day schedule alternating with CHAMOCA, remission was achieved in 10 of 12 (83%) high-risk patients. 72 During that time, VP-I6 was also introduced into a sequence of drugs devised for patients in the medium-risk category, which included hydroxyurea, methotrexate, 6mercaptopurine, and actinomycin D. Of 76 patients in this category reported on in 1986, all were alive and in remission. 90 Bagshawe arranged the five most effective and compatible drugs against trophoblastic tumors (VP-I6, methotrexate, actinomycin D, cyclophosphamide, and oncovin) into two groups, resulting in a split protocol with the acronym of EMACO. Although treatment is given weekly, EMACO requires only I day's hospitalization every 2 weeks and is significantly better tolerated than CHAMOCA. It also has the advantage of using a wide range of drugs in a short time without waiting for resistance to emerge. Of 65 high-risk patients treated with EMACO, survival rates reported in 1986 were 91 % in patients who had received no prior chemotherapy and 78% in patients who had received prior chemotherapy.9l When used as primary therapy in high-risk patients at the University of Milan, complete responses were achieved in 16 of 17 patients, and when given as second line therapy complete responses were achieved in 10 of 14 patients. 92 These achievements have resulted in the EMACO regimen currently being considered the regimen of choice in most high-risk patients and one of the most effective treatments for drug-resistant patients. The introduction of VP-I6 is therefore considered to be perhaps the most important advance in the treatment of trophoblastic neoplasms in the past 10

1292 Jones

years. The exceptional activity of this drug is apparent from the recent report by Wong et al.,"' in which eight of nine drug-resistant patients achieved complete remission after treatment with VP-16, methotrexate, and bleomycin. In addition, oral VP-16 given by these investigators to 37 patients with resistant or metastatic disease produced permanent remissions in all the patients and none required a change in therapy.94 Newlands first reported on the use of cisplatin in gestational choriocarcinoma in 1978. Of an initial group of 17 patients with drug-resistant tumors who received high-dose cisplatin (120 mg/m2), no activity was seen in two patients treated with cisplatin alone. However, in combination with vincristine and methotrexate, there were six responses, seven improvements, and three disease progressions."' Using this protocol, Newland subsequently reported 96 a 33% response rate in 24 drug-resistant patients. Cisplatin used in combination is therefore considered to be an active agent in the treatment of trophoblastic tumors. When used in combination with the POMB regimen (cisplatin, vincristine, methotrexate, and bleomycin), some patients with EMACO-resistant tumors were also shown to be responsive. Several authors have recently reported on patients with resistant disease who received vinblastine, cisplatin, and bleomycin (VPB), with sustained remission rates ranging from 20% to 50%.97.98 These responses suggest that cisplatin was the active agent in this regimen, in view of our experience with four patients treated with VAB who did not respond. The ultra-high-risk patient represents the 10% to 15% of patients traditionally classified as high-risk who die from progressive choriocarcinoma despite initial combination chemotherapy. The identification of these patients can be accomplished, as previously described, with the use of the scoring system adopted by the World Health Organization in 1983. As an example, a patient with brain metastasis presenting with an interval of greate than 12 months between the antecedent pregnancy and the start of chemotherapy would receive a score of eight. If, in addition, the initial hCG level was greater than 100,000 IU /L, the total score would be 12. In analyzing the Gynecologic Oncology Group randomized trial comparing MAC to modified CHAMOMA, Surwit"" pointed out that all patients in both arms of the study with a World Health Organization score of seven or less were cured. He further noted that the combined series of both MAC and modified CHAMOMA could identify the ultra-high-risk patient. These patients had a World Health Organization score of eight or greater and only 47% achieved primary remission. Similarly, clinicians at the New England Trophoblastic Disease Center retrospectively analyzed their data and found that only 43% of their highrisk patients with a World Health Organization score

May 1990 Am J Obstet Gynecol

of eight or greater survived, whereas 100% of patients with a prognostic score of seven or less survived. loo A striking example of the significance of this prognostic score can be seen in Lurain's studylOl of the causes of treatment failure at the John I. Brewer Trophoblastic Disease Center from 1962 to 1986, in which 96% of patients who died had a prognostic score of eight or greater, with an average score of 13. The value of such analyses lies in the fact that experienced clinicians no longer recommend conventional combination chemotherapy for this group of patients. The optimal therapy for patients with high prognostic scores has obviously not yet been determined. It does seem clear, however, that VP-16 and possibly cisplatin should be included early in the chemotherapy of any multiagent regimen designed for this group of patients. Hopefully the number of drug-resistant fatal cases can then be reduced. Comment

New knowledge about the genetics of trophoblastic neoplasms may have far-reaching implications relative to the control mechanisms of neoplastic cell growth in general. Modern biochemical monitoring of patients with trophoblastic disease requires sophisticated technology that suggests the need for consultations at centers where such assays are available. Etoposide and cisplatin have become valuable new drugs for treating gestational trophoblastic disease. Chemotherapy protocols in the future will ideally incorporate additional new cytotoxic agents that are as effective but less toxic than those currently available. In reviewing the 10 years from 1968 to 1978, Dr. John L. Lewis, Jr. noted that no center had made more important contributions to our understanding and treatment of gestational choriocarcinoma than the unit at Charing Cross Hospital, London under the direction of Dr. Kenneth Bagshawe. From 1978 to 1988, significant improvement in the prognosis for patients with trophoblastic tumors has come about largely as a result of the innovative chemotherapy protocols developed by these investigators. In closing, therefore, a timely quote from Dr. Bagshawe seems appropriate. In 1987, in discussing the evolution and complexities of modern treatment of gestational choriocarcinoma, he wrote: "It is an all too frequent observation that whether a young woman with choriocarcinoma survives can still depend on the general experience and training of the doctors she encounters early in her illness."72

REFERENCES I. Hertz R. Biological aspects of gestational neoplasms derived from trophoblast. Ann NY Acad Sci 1971; 172:27987. 2. Ober WB, Fass OS. The early history of choriocarcinoma.] Hist Med 1961;16:49-73.

Gestational trophoblastic disease

Volume 162 Number 5

3. Rather Lj. Pare A. The Countess Margaret, multiple births and hydatidiform moles. Bull NY Acad Med 1971;47:508-15. 4. Lawler SD. Genetic studies on hydatidiform mules. In: Patillo RA, Hussa RO, eds. Advances in experimental medicine and biology. New York: Plenum, 1984; 176: 147-61. 5. Vassilakos P, Kajii T. Hydatidiform mole: two entities. Lancet 1976; 1:259. 6. Szulman AE, Surti U. The syndrumes of hydatidiform mole. I. Cytogenetic and morpholugic correlations. A:M J OBSTET GYNECOL 1978;131:655-71. 7. Vassilakos P, Riotton G, Kajii T. Hydatidifurm mole: twu entities-a morphologic and cytogenetic study with some clinical considerations. AM j OBSTET GY~ECOL 1977; 127: 167-70. 8. Szulman AE, Surti U. The syndromes of hydatidiform mole. II. MOI'phologic evolution of the complete and partial mole. AM j OBSTET GYNECOL 1978; 132:20-7. 9. Szulman AE, Surti U, Berman M. Patient with partial mole requiring chemotherapy. Lancet 1978;2: 1099. 10. Berkowitz RS, Goldstein DP, Marean AR, Bernstein MR. Proliferative sequelae after evacuation of partial hydatidiform mole. Lancet 1979;2:804-5. 11. Szulman AE, Ma HK, Wong LC, Hsu C. Residual trophoblastic disease in association with partial hydatidiform mole. Obstet Gynecol 1981;57:392-4. 12. Hunore LH. Recurrent partial hydatidiform mole: report of a case. AM j OBSTET GYJ\ECOL 1987; 165:922-4. 13. Caspersson T, Lomakka G, Zech L. The 24 fluorescence patterns of human metaphase chromosomes distinguishing characters and variability. Hereditas 1971 ;67: 89-102. 14. Kajii T, Ohama K. Androgenetic origin of hydatidiform mole. Nature 1977;268:633-4. 15. Wake N, Takagi T, Sasaki M. Androgenesis as a cause of hydatidiform mole. JNCI 1978;60:51-7. 16. jacobs PA, Hassold Tj, Matsuyama AM, Newlands 1M. Chromosome constitution of gestational trophoblastic disease. Lancet 1978;2:49. 17. Yamashita K, Wake N, Araki T, Ichinoe K, Kuroda M. Human lymphocyte antigen expression in hydatidiform mole: androgenesis following fertilization by a haploid sperm. AM .I OBSTET GYNECOL 1979; 135:597-600. 18. Surti U, Szulman AE, O'Brien S. Complete (classic) hydatidiform mole with 46,XY karyotype of paternal origin. Hum Genet 1979;51:153-5. 19. Surti U, Szulman AE, O'Brien S. Dispermic origin and clinical outcome of three complete hydatidiform moles with 46,XY karyotype. AM j OBSTET GY:';ECOl. 1982; 144:84-7. 20. Pattillo RA, Sasaki S, Katayama KP, Roesler M, Mattingly RF. Genesis of 46,XY hydatidiform mole. AM j OBSTET GYJ\ECOL 1981;141:104-5. 21. Fisher RA, Sheppard DM, Lawler SD. Two patients with complete hydatidiform mole with 46,XY karyotype. Br j Obstet Gynaecol 1984;91 :690-3. 22. Kajii T, Kurashige H, Ohama K, Uchino F. XY and XX complete moles: clinical and morphologic correlations. AM .I OBSTET GYj\;ECOl. 1984; 150:57-64. 23. Lawler SD, Pickthall Vj, Fisher RA, et al. Genetic studies on complete and partial hydatidiform moles. Lancet 1979;2:580. 24. Lawler SD, Povey S, Fisher RA, Pickthail VJ. Genetic studies on hydatidiform moles. II. The origin of complete moles. Ann Hum Genet 1982;46:209-22. 25. jacobs PA, Wilson CM, Sprenkle jA, Rosenshein NB, Migeon BR. Mechanism of origin of complete hydatidiform moles. Nature 1980;286:714-6. 26. Surti U. Genetic concepts and techniques. In: Szulman AE, Buchsbaum Hj, eds. Gestational trophoblastic disease. New York: Springer-Verlag, 1987:111-21. 27. Yamashita K, Wake N, Araki T, Ichinoe K, Kuroda M.

28. 29. 30.

31.

32. 33. 34.

35.

36. 37. 38.

39.

40. 41.

42.

43.

44.

45.

46.

47.

1293

A further HLA study of hydatidiform moles. Gynecol Oncol 1981; II :23-8. Southern EM. Application of DNA analysis to mapping the human genome. Cytogenet Cell Genet 1982;32: 52-7. Wallace DC, Surti U, Adams CW, Szulman AE. Complete moles have paternal chromosomes but maternal mitochondrial DNA. Hum Genet 1982;61:145-7. Edwards YH,jeremiah Sj, McMillan SL, Povey S, Fisher RA, Lawler SD. Complete hydatidiform moles combine maternal mitochondria with a paternal nuclear genome. Ann Hum Genet 1984;48:119-27. Vejerslev LO, Mogensen B, Olsen S. Hydatidiform mole in a Danish material. Morphology and cytogenetics: a preliminary communication. Acta Radiol Oncol 1984; 23:97-102. Wake N, Seki 1; Fujita H, et al. Malignant potential of homozygous and heterozygous complete moles. Cancer Res 1984;44: 1226-30. Davis JR. Cytogenetic and clinical features of placental moles: synopsis and update. Cancer Genet Cytogenet 1986; 19: 123-7. Sarkar S, Kacinski BM, Kohorn EI, et al. Demonstration of myc and ras oncogene expression by hybridization in situ hydatidiform mole and in the BeWo choriocarcinoma cell line. AM j OBSTET GYj\;ECOL 1986;154: 390-3. Baker VV, Shingleton HM, Hatch KD, Miller DM. Selective inhibition of c-myc expression by the ribonucleic acid synthesis inhibitor mithramycin. A~t .I OBSTET GyNECOL 1988;158:762-7. Swaminathan N, Bahl OP. Dissociation and recombination of the subunits of human chorionic gonadotropin. Biophys Res Commun 1970;40:422-7. Morgan FJ, Canfield RE. Nature of the subunits of chorionic gonadotropin. Endocrinology 1971 ;88: 1045-53. Vaitukaitis jL. Glycoprotein hormones and their subunits: immunological and biological characterization. In: McKerns KW, ed. Structure and function of the gonadotropins. New Yurk: Plenum, 1978:339-60. Bahl OP, Carlsen RB, Bellisario R, Swaminathan N. Human chorionic gonadotropins: amino acid sequence of the alpha and beta subunits. Biochem Biophys Res Commun 1972;48:416-22. Ross GT. Clinical relevance of research on the structure of human chorionic gonadotropin. AM .I OBSTET GyJ\ECOL 1977;129:795-808. Morgan Fj, Birken S, Canfield RE. The amino acid sequence of human chorionic gonadotropin. The apha subunit and the beta subunit. J Bioi Chern 1975; 250:5247-58. Vaitukaitis JL, Braunstein GD, Ross GT. A radioimmunoassay which specifically measures human chorionic gonadotropin in the prese'nce of human luteinizing hormone. AMj OBSTET GYNECOL 1972;113:751-8. Jones WB, LewisjLjr, Lehr M. Monitor of chemotherapy in the gestational trophoblastic neoplasm by radioimmunoassay of the j3-subunit of human chorionic gonadotropin. AM j OBSTET GYNECOL 1975; 121:669-73. Yuen BH, Cannon W, Benedict jL, Boyes DA. Plasma j3-subunit human chorionic gonadotropin assay in molar pregnancy and choriocarcinoma. AMJ OBSTET GYNECOL 1977; 127:711-2. 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. A:M j OBSTET GYNECOL 1974; 120: 1025-8. Chen HC, Ayala AR, Hodgen GE, Birkell RE, Canfield RE, Ross GT. First specific assay for chorionic gonadotropin in human urinary extracts. Clin Res 1976; 24:375A. Schreiber jR, Rebar RW, Chen HC, Hodgen GD, Ross

294 Jones

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59. 60. 61. 62. 63. 64. 65.

GT. Limitation of the specific serum radioimmunoassay for human chorionic gonadotropin in the management of trophoblastic neoplasms. AM ] OBSTET GYNECOL 1976; 125:705-7. Ayala AR, Nisula BC, Chen HC, Hodgen GD, Ross GT. Highly sensitive radioimmunoassay for chorionic gonadotropin in human urine. ] Clin Endocrinol Metab 1978;47:767-73. Wehmann RE, Ayala AR, Birken S, Canfield RE, Nisula BC. Improved monitoring of gestational trophoblastic neoplasia using a highly sensitive assay for urinary human chorionic gonadotropin. AM ] OBSTET GYNECOL 1981; 140:753-7. Dawood MY, Saxena BB, Landesman R. Human chorionic gonadotropin and its subunits in hydatidiform mole and choriocarcinoma. Obstet Gynecol 1977;50: 172-81. Quigley MM, Tyrey L, Hammond CB. Utility of assay of alpha subunit of human chorionic gonadotropin in management of gestational trophoblastic malignancies. AM] OBSTET GYNECOL 1980; 138:545-9. Fan C, Goto S, Furuhashi Y, Tomoda Y. Radioimmunoassay of the serum free ~-subunit of human chorionic gonadotropin in trophoblastic disease.] Clin Endocrinol Metab 1987;64:313-8. Khazaeli MB, Buchina ES, Pattillo RA, Soong S], Hatch KD. Radioimmunoassay of free ~-subunit of human chorionic gonadotropin in diagnosis of high-risk, and lowrisk gestational trophoblastic disease. AM] OBSTET GyNECOL 1989; 160:444-9. Akar AH, Wehmann RE, Blithe DL, Blacker C, Nisula BC. A radioimmunoassay for the core fragment of the human chorionic gonadotropin ~-subunit. ] Clin Endocrinol Metab 1988;66:538-44. Amr S, Wehmann RE, Birken S, Canfield RE. Characterization of a carboxy terminal peptide fragment of the human choriogonadotropin ~-subunit excreted in the urine of a woman with choriocarcinoma. ] Clin Invest 1983;71 :329-39. O'Connor ]F, Schlatterer ]P, Birken S, et at. Development of highly sensitive immunoassays to measure human chorionic gonadotropin, its ~-subunit, and 13 core fragment in the urine: application to malignancies. Cancer Res 1988;48: 1361-6. Cole LA, Wang Y, Elliott M, et at. Urinary human chorionic gonadotropin free ~-subunit and ~-core fragment: a new marker of gynecological cancers. Cancer Res 1988;48:1356-60. Stenman UH, Myllynen L, Alfthan H, Seppala M. UItrarapid and highly sensitive time-resolved ftuoroimmunometric assay for chorionic gonadotropin. Lancet 1983;2:647-9. Alfthan H. Comparison of immunoradiometric and immunoftuorometric assays for serum hCG. ] Immunol Methods 1986;88:239-44. Iwasa S, Kitada C, Yoshida I, Kondo K, Hori M, Fujino M. Highly specific enzyme immunoassay for human chorionic gonadotropin.] Biochem 1981 ;89: 1091-9. Sekiya T, Furuhashi Y, Goto S, Kaseki S, Tomoda Y, Kato K. Specific enzyme immunoassay for human chorionic gonadotropin. Acta Endocrinol 1981;97:562-8. Bock ]L, Furgiuele ], Segen ]C. Choriogonadotropin measured with the Tandem-E immunoenzymetric assay system. Clin Chern 1985;31:441-4. Yazaki K, Armstrong FG, Koide SS. Unique variants of hCG in sera of choriocarcinoma patients. Cancer 1987; 59:795-7. Li MC, Hertz R, Spencer DB. Effects of methotrexate therapy upon choriocarcinoma and chorioadenoma. Proc Soc Exp Bioi Med 1956;93:361-6. Hertz R, Lewis]L]r, Lipsett MB. Five years' experience with the chemotherapy of metastatic choriocarcinoma and related trophoblastic tumors in women. AM] OBSTET GYNECOL 1961;82:631-40.

May 1990

Am J Obstet Gynecol

66. Ross GT, Hammond CB, Lipsett MD. Chemotherapy of metastatic and non-metastatic gestational trophoblastic neoplasms. Tex Rep Bioi Med 1966;24:326-38. 67. Issell BF. The podophyllotoxin derivatives VP16-213 and VM26. Cancer Chemother Pharmacol 1982;7:7380. 68. Loike ]D. VP16-213 and podophyllotoxin. A study on the relationship between chemical structure and biological activity. Cancer Chemother Pharmacol 1982; 7: 103-

II.

69. O'Dwyer P], Leyland-]ones B, Alonso MT, Marsoni S, Wittes RE. Etoposide (VP-16-213). Current status of an active anticancer drug. N Engl] Med 1985;312:692-700. 70. Ross W, Rowet T, Glisson B, Yalowitch ], Liu L. Role of topoisomerase II in mediating epipodophyllotoxininduced DNA cleavage. Cancer Res 1984;44:5857-60. 71. Creaven PJ. The clinical pharmacology of VM26 and VPI6-213. A brief overview. Cancer Chemother PharmacoI1982;7:133-40. 72. Bagshawe KD. From methotrexate to EMA/CO in gestational trophoblastic disease. In: Szulman AE, Buchsbaum H], eds. Gestational trophoblastic disease. New York: Springer-Verlag, 1987:127-34. 73. Newlands ES, Bagshawe KD. The role ofVP16-213 (etoposide: NSC-141540) in gestational choriocarcinoma. Cancer Chemother Pharmacol 1982;7:211-4. 74. Rosenberg B. Cisplatin: its history and possible mechanisms of action. In: Prestayko AW, Crooke ST, Carter SK, eds. Cisplatin: current status and new developments. New York; Academic, 1980:9-21. 75. Johnson NP, Lapetoule P, Razaka H, Villani G. Biological and biochemical effects of DNA damage caused by platinum compounds. In: McBrien DCH, Slater TF, eds. Biochemical mechanisms of platinum antitumor drugs. Oxford: IRL Press, 1986: 1-28. 76. Madias NE, Harrington ]T. Platinum nephrotoxicity. Am] Med 1978;68:307-14. 77. Ozols RE, Corden B], Collins], Young RC. High dose cisplatin in hypertonic saline: renal effects and pharmacokineticsofa40 mg/m2 qd x 5 schedule. In: Hacker MP, Douple EB, Krakoff IH, eds. Platinum coordination complexes in cancer chemotherapy. Boston: Martinus Nijhoff, 1984:321-9. 78. Ross GT, Goldstein DP, Hertz R, Lipsett MB, Odell WD. Sequential use of methotrexate and actinomycin-D in the treatment of metastatic choriocarcinoma and related trophoblastic diseases in women. AM ] OBSTET GYNECOL 1965;93:223-9. 79. Hammond CB, Parker RT. Diagnosis and treatment of trophoblastic disease. A report from the Southeastern Regional Center. Obstet GynecoI1970;35:132-43. 80. Lewis]L]r. Treatment of metastatic trophoblastic neoplasms. AM] OBSTET GYNECOL 1979;136:163-72. 81. Bagshawe KD. Treatment of trophoblastic tumours. Recent results. Cancer Res 1977;62:192-9. 82. WHO Scientific Group. Gestational trophoblastic disease. Geneva: World Health Organization, 1982:692. 83. Hammond CB, Borchert LG, Tyrey L, Creasman WT, Parker RT. Treatment of metastatic trophoblastic disease: good and poor prognosis. AM] OBSTET GYNECOL 1973; 115:451-7. 84. Lurain]R, Brewer]l. Treatment of high-risk gestational trophoblastic disease with methotrexate, actinomycin D, and cyclophosphamide chemotherapy. Obstet Gynecol 1985;65:830-6. 85. Curry SL, Blessing ]A, DiSaia P], et at. A prospective randomized comparison of methotrexate, dactinomycin, and chlorambucil versus methotrexate, dactinomycin, cyclophosphamide, doxorubicin, melphalan, hydroxyurea, and vincristine in "poor prognosis" metastatic gestational trophoblastic disease: a Gynecologic Oncology Group Study. Obstet Gynecol 1989;73:357-62. 86. Berkowitz RS, Goldstein DP, Bernstein MR. Modified triple chemotherapy. In: Szulman AE, Buchsbaum H.J,

Gestational trophoblastic disease

Volume 162 1';umber !i

87.

88. 89.

90.

91.

92. 93. 94.

eds. Gestational trophoblastic disease. New York: Springer-Verlag, 1987: 146-54. Weed .lC, Barnard DE, Currie .lL, et al: Chemotherapy with the modified Bagshawe protocol for poor prognosis metastatic trophoblastic disease. Obstet Gynecol 1982; 59:377-80. Newlands ES, Bagshawe KD. Epipodophyllin derivative (VPI6-213) in malignant teratomas and choriocarcinomas. Lancet 1977;2:87. Newlands ES, Bagshawe KD. Anti-tumour activity of the epipodophyllin derivative VP16-213 (etoposide: NSC141540) in gestational choriocarcinoma. Eur .l Cancer 1980;16:401-5. Newlands ES, Bagshawe KD, Begent RHI, Rustin G.lS, Holden L. Developments in chemotherapy for mediumand high-risk patients with gestational trophoblastic tumours (1979-1984). Br J Obstet Gynaecol 1986;93:63-9. Newlands ES, Rustin G.lS, Bagshawe KD, Begent RH.l, Holden L. Weekly EMA/CO chemotherapy for high risk gestational trophoblastic tumours (GTTs). Proc Am Soc Clin Oncol 1986;5: 112. Bolis G, Bonazzi C, Landoni F, et al. EMA/CO regimen in high-risk gestational trophoblastic tumor (GTT). Gynecol Oncol 1988;31 :439-44. Wong LC, Choo YC, Ma HK. Etoposide, methotrexate, and bleomycin in drug resistant gestational trophoblastic disease. Gynecol Oncol 1986;24:51-7. Wong LC, Choo YC, Ma HK. Use of oral VP16-213 as

95.

96. 97.

98.

99. 100.

lOl.

1295

primary chemotherapeutic agent in treatment of gestational trophoblastic disease. A:-"I .J OSHEr (~\,:\f.(:01 1984;150:924-7. Newlands ES, Bagshawe KD. Activity of high dose cisplatinum (NCI 119875) in combination with vincristine and methotrexate in drug-resistant gestational clIO riocarcinoma. A report of 17 cases. Br .l Cancer 1979; 40:943-5. Newlands ES. New chemotherapeutic agents in the management of gestational trophoblastic disease. Semin Oncol 1982;9:239-43. DuBeschter B, Berkowitz RS, Goldstein DP, Bernstein M. Vinblastine. cisplatin and bleomycin as salvage therapy for refractory high-risk metastatic gestational trophoblastic disease . .l Reprod Med 1989;34: 189-92. Gordon AN, Kavanagh J.J, Gershenson DM, Saul PB, Copeland LJ, Stringer CA. Cisplatin, vinblastine, and bleomycin combination therapy in resistant gestational trophoblastic disease. Cancer 1986;58: 1407-10. Surwit EA. Management of high-risk gestational trophoblastic disease. J Reprod Med 1987;32:657-62. DuBeschter B, Berkowitz RS, Goldstein DP. Cramer OW. Bernstein M. Metastatic gestational trophoblastic disease: experience at the New England Trophoblastic Disease Center, 1965 to 1985. Obstet Gynecol 1987;69: 390-5. Lurain JR. Causes of treatment failure in gestational trophoblastic disease . .J Reprod Med 1987;32:675-9.

Bound volumes available to subscribers Bound volumes of the AMERICA1\: jOUR1';AL OF OBSTETRICS A1';n GY1\:ECOLO(;Y are available to subscribers (only) for the 1990 issues from the Publisher, at a cost of $56.00 ($79.00 international) for Vol. 162 (january-June) and Vol. 163 (j uly-December). Shipping charges are included. Each bound volume contains a subject and author index and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram with the jOUR1\:AL name, volume number, and year stamped in gold on the spine. PUJII/ent must accompany all ordl'l's. Contact Mosby- Vear Book, Inc., Circulation Department, 11830 Westline Industrial Drive, Sl. Louis, Missouri 63146-3318, USA; phone (800) 325-4177, ext. 7351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular JOURNAL subscription.