Secondary Logo

Journal Logo

Original Research

Gestational Trophoblastic Neoplasia: Treatment Outcomes

Hoekstra, Anna V. MD, MPH; Lurain, John R. MD; Rademaker, Alfred W. PhD; Schink, Julian C. MD

Author Information
doi: 10.1097/AOG.0b013e31817f58ae
  • Free

The spectrum of cellular proliferations arising from placental villous trophoblast known as gestational trophoblastic disease encompasses hydatidiform mole (complete and partial) and gestational trophoblastic neoplasia (GTN), which includes invasive mole, choriocarcinoma, and placental site trophoblastic tumor. The outcomes for GTN were poor before the introduction of chemotherapy into their management approximately 50 years ago. The mortality rate for invasive mole approached 15%, most often due to sepsis, hemorrhage, embolic phenomena, or complications from surgery. Mortality from metastatic choriocarcinoma approached 100% and was still 60% when hysterectomy was done for apparent nonmetastatic disease. Gestational trophoblastic neoplasia is now one of the most curable of all human tumors, with a cure rate exceeding 90%.1 Improvements in survival are attributed to advances in chemotherapy, better assays for the tumor marker hCG, development of specialized treatment centers, identification of prognostic scoring systems to predict treatment response and enhance individualization of therapy, and use of combined modality treatment with chemotherapy, radiation, and surgery to treat the highest-risk patients.2

In 1982, we reported the results of treatment of GTN at the John I. Brewer Trophoblastic Disease Center of Northwestern University from 1962 through 1978.3 Overall survival rate for 396 patients was 89%: 100% for nonmetastatic disease and 78% for metastatic disease. Factors found to significantly influence survival were 1) clinicopathologic diagnosis of choriocarcinoma; 2) time greater than 4 months from pregnancy event to treatment; 3) pretreatment hCG more than 100,000 milliinternational units/mL; 4) metastases to sites other than the lung and vagina; 5) antecedent term pregnancy; and 6) prior failed chemotherapy. The report emphasized the high curability of GTN when appropriate treatment was applied and elucidated which patients were at highest risk of treatment failure and would benefit from aggressive multimodality therapy.

We present an updated series of 408 patients with GTN (invasive mole and choriocarcinoma) treated at the Brewer Trophoblastic Disease Center between January 1, 1979, and December 31, 2006. Outcomes and prognostic factors are analyzed with respect to response to chemotherapy and survival, and are compared with those for patients treated from 1962 to 1978.

MATERIALS AND METHODS

From 1979 through 2006, 408 patients with GTN (invasive mole and choriocarcinoma, excluding placental-site tumor) were treated at the John I. Brewer Trophoblastic Disease Center. The diagnosis of GTN was made using both clinical and histologic criteria. As outlined by the International Federation of Gynecology and Obstetrics (FIGO) in 2002, postmolar GTN was diagnosed after a plateau in hCG level for four consecutive tests over 3 weeks, a rise in the hCG level by 10% for three consecutive tests over 2 weeks, any elevation in hCG level for 6 months after evacuation, a histopathologic diagnosis of choriocarcinoma, or presence of metastatic disease.4 The trophoblastic disease was categorized as metastatic or nonmetastatic invasive mole or choriocarcinoma. Patients were diagnosed with invasive mole by pathologic diagnosis of tissue or, if tissue wasn’t available, after a molar pregnancy when there was a plateau or rise in hCG or detection of metastatic disease in the absence of a histologic diagnosis of choriocarcinoma. Choriocarcinoma was diagnosed if there was a pathologic diagnosis of choriocarcinoma made by examination of tissue specimens showing hyperplastic and dysplastic trophoblast in the absence of villi, or if the tumor developed after a nonmolar pregnancy, thus excluding the possibility of an invasive mole.

Staging of GTN took place after standard imaging and laboratory testing was performed. After physical examination, patients underwent serum testing for quantitative hCG, complete blood count, and metabolic profile including liver and renal function tests. Computed tomography scans of the chest, abdomen, and pelvis were completed, as well as either computed tomography or magnetic resonance imaging of the brain if the other scans revealed metastatic disease. The patient was classified according to the anatomic staging system adopted by FIGO in 1982 along with the addition of the modified World Health Organization (WHO) risk factor scoring system in 2002.3 World Health Organization score was determined by age, antecedent pregnancy, duration of disease, pretreatment hCG level, site of metastases, number of metastases, size of largest tumor, and previous failed chemotherapy. Patients with stage I and stages II/III, score less than 7 were considered low risk, whereas patients with stages II/III, score 7 or more and stage IV were considered high risk.

Treatment protocols were standardized for all providers treating patients with GTN during the time of data collection.5 First-line chemotherapy for low-risk disease did not change at the Center from 1962 to 2006.6,7 Treatment for nonmetastatic and low-risk metastatic disease usually started with methotrexate 0.4mg/kg (maximum 25 mg) intravenously daily for 5 days every other week (352 patients). An alternative regimen of dactinomycin 0.5mg intravenously daily for 5 days every other week was given as initial treatment to seven patients and for methotrexate resistance or toxicity to 64 patients. Twenty-nine patients (8.1%) who failed sequential single-agent chemotherapy were managed with combination drug regimens: methotrexate, dactinomycin, and cyclophosphamide (MAC) from 1979 to 1982 (four patients); cyclophosphamide, hydroxyurea, dactinomycin, methotrexate with folinic acid, vincristine, and doxorubicin (CHAMOCA) from 1983 to 1985 (three patients); and etoposide, high-dose methotrexate with folinic acid, dactinomycin, cyclophosphamide, and vincristine (EMA-CO) from 1986 to 2006 (23 patients). Other drug combinations used in this group were EMA-EP (EMA-CO, but substituting etoposide and cisplatin for cyclophosphamide and vincristine) in four patients; bleomycin, etoposide, and cisplatin (BEP) in six patients; and ifosfamide, carboplatin, and etoposide (ICE) in two patients.

Forty-nine patients with metastatic high-risk disease initially underwent treatment with MAC (17 patients) from 1979 to 1982, CHAMOCA (1 patient) from 1983 to 1985, and EMA-CO (31 patients) from 1986 to 2006.8,9 Multiagent chemotherapeutic regimens used after failure of MAC, CHAMOCA or EMA-CO included EMA-EP in 10 patients, etoposide, ifosfamide, cisplatin (VIP) in two patients, ICE in one patient, BEP in nine patients, and paclitaxel, etoposide (TE)/paclitaxel, cisplatin (TP) in one patient. In patients with clear resistance to the methotrexate-based regimens, cisplatin/etoposide-based combinations, such as BEP, ICE, VIP, and TE/TP were used.10 No course of chemotherapy was given with a white blood cell count less than 3,000/mL or platelet count less than 100,000/mL. A chemotherapy failure was diagnosed after a plateau or rise of the hCG after two courses or if toxicity precluded an adequate dose or frequency of treatment.

Radiation and surgical treatment were also used in selected patients. Whole brain irradiation, 3000 cGy in 200-cGy fractions (nine patients) or surgical excision with stereotactic irradiation (two patients) was used in 11 patients with brain metastases, usually given simultaneously with the initiation of systemic chemotherapy.11 Adjuvant surgical procedures were employed in 70 patients (17%). Thirty-three patients (8%) underwent initial surgical procedures for the purposes of diagnosis, bleeding, or decreasing tumor burden, including 27 hysterectomies, two brain resections, one thoracotomy, one small bowel resection, and one liver resection. Thirty-seven patients (9%) underwent interval operations, including 27 hysterectomies, seven thoracotomies, three brain resections, and two uterine wedge resections, for removal of chemotherapy resistant/persistent disease or control of hemorrhage.12

Routine surveillance during treatment included complete blood counts, chemistry profiles, and hCG levels the first day of each course of treatment. Patients were considered to be responding if there was a decrease in hCG or the disappearance of metastases on imaging. The hCG assay used during this time period was either a serum radioimmunoassay or enzyme linked immunoassay for total hCG. Remission was diagnosed after three consecutive weekly hCG levels were within normal range in addition to a lack of symptoms or radiologic evidence of disease. Two courses of chemotherapy were usually given after the first normal hCG level. Follow-up after remission and completing treatment for GTN consisted of 1 year of monthly hCG levels and use of oral or barrier contraception to prevent pregnancy.

Data on all 408 patients were collected retrospectively from records maintained prospectively in the John I. Brewer Trophoblastic Disease Center from January 1, 1979, through December 31, 2006, to allow for at least 1 year of follow-up. Patient and disease characteristics were summarized, such as age and gravidity, clinicopathologic diagnosis, antecedent pregnancy, time to diagnosis, pretreatment hCG, extent of disease, and whether they had failed prior therapy. In patients treated before widespread application of WHO prognostic scores and FIGO stage, these were calculated and assigned retrospectively. Factors related to response to initial chemotherapy were analyzed using the Fisher exact test for categorical variables and the Wilcoxon rank sum test for continuous variables. In addition, univariable and multivariable analyses were done using logistic regression analysis. Only variables found to be significant at P<.05 in the univariable analysis were included in the multivariable analysis. Continuous data were reported as medians, and statistical significance was considered at P<.05. The Institutional Review Board of Northwestern University approved the collection of information for this report.

RESULTS

Table 1 presents the results of 804 patients with GTN (choriocarcinoma or invasive mole) treated at the Brewer Trophoblastic Disease Center of Northwestern University from 1962 through 2006. The overall survival rate was 93.3%, but improved from 88.6% (351 of 396) in 1962–1978 to 97.8% (399 of 408) in 1979–2006. All 491 patients with nonmetastatic disease treated during both time periods were cured; however, the cure rate for patients with metastatic disease increased from 77.8% (158 of 203) during 1962–1978 to 91.8% (101 of 110) during 1979–2006.

Table 1
Table 1:
Comparison of Survival in Patients With Gestational Trophoblastic Neoplasia Between 1962–1978 and 1979–2006

Of the nine patients who died from 1979 to 2006, four started treatment in 1979, and two presented after failed treatment at outside institutions, whereas only three patients who were primarily treated at the Brewer Center between 1980 and 2006 died of disease. All nine of the patients who did not survive had lung metastases, five patients had liver metastases, four patients had brain metastases, and four patients had additional lesions in the abdomen, including the spleen, pancreas, and/or small bowel. Of the three patients primarily treated at the Brewer Center since 1980 who did not survive, all had a histopathologic diagnosis of choriocarcinoma, two of the three had pretreatment hCG levels more than 100,000 milliinternational units/mL, and all three had WHO scores greater than 7. All three presented with metastatic disease to the lungs, and two patients had additional metastases to the vagina and brain (for one patient) and to the kidneys, liver, and spleen (for the other).

The overall complete response rate to the initial chemotherapy regimen given at the Brewer Center was 77% (Table 2). The complete response rate to sequential single-agent chemotherapy for nonmetastatic and low-risk metastatic disease was 92% (330 of 359). The remaining 29 low-risk patients were all placed into permanent remission with the use of multiagent chemotherapy and/or surgery. The complete response rate to initial multiagent chemotherapy for high-risk metastatic disease was 67% (33 of 49). An additional 44% (7 of 16) of patients had a complete response to secondary chemotherapy resulting in an overall complete response rate to sequential multiagent chemotherapy in high-risk patients of 82% (40 of 49).

Table 2
Table 2:
Chemotherapy for Patients With Gestational Trophoblastic Neoplasia Between 1979 and 2006

The characteristics of patients who responded completely to the first chemotherapy regimen differed from those who required additional treatments (Table 3). Those who were resistant to the first chemotherapy regimen were older (30 years compared with 28 years P =.008), with more advanced gravidity (four compared with one; P<.001); they had higher median pretreatment hCG (11,061 compared with 3,789 milliinternational units/mL; P<.001); and they had a higher mean WHO score (2.5 compared with 2; P<.001).

Table 3
Table 3:
Univariable and Multivariable Logistic Regression Analysis of Clinical Characteristics of Patients Treated for Gestational Trophoblastic Neoplasia from 1979 to 2006 With Respect to Resistance Rate to Initial Chemotherapy and Odds Radio for Response to Initial Chemotherapy

Patients with metastatic disease had an increased rate of resistance to initial chemotherapy (41%) when compared with patients with nonmetastatic disease (12%) (P<.001). The site of metastatic disease also influenced response to initial chemotherapy (Table 4). Among 110 patients with metastatic disease, 108 had lung metastases, 12 had brain metastases, 15 had vaginal metastases, 12 had liver metastases, and nine had other intraabdominal or extraabdominal sites of metastases. Resistance to initial chemotherapy was seen in 76% (19 of 25) of patients with a metastatic site other than the lung or vagina, compared with 31% (26 of 85) in patients with metastatic disease localized to the lung and/or vagina only (P<.001).

Table 4
Table 4:
Influence of Location of Metastatic Disease on Resistance to Initial Chemotherapy and Survival Rate in Gestational Trophoblastic Neoplasia Patients Treated at the Brewer Center from 1979 to 2006

Patients with a pathologic diagnosis of choriocarcinoma or antecedent nonmolar gestation had a resistance rate of 40% compared with those patients with an antecedent molar pregnancy and no pathologic diagnosis of choriocarcinoma (presumed invasive mole) who had a resistance rate of 13% (P<.001). There was a significant difference in resistance to initial chemotherapy in patients with an antecedent term gestation (45%) as compared with the combined rates for disease after a hydatidiform mole, an elective or spontaneous abortion, or ectopic pregnancy (17%) (P<.001). Patients with a duration of disease greater than 4 months compared with less than 4 months showed an increased rate of resistance to initial chemotherapy (35% compared with 17%; P=.001). Pretreatment hCG did not significantly affect resistance to initial chemotherapy in this study.

The scoring system adopted by the WHO in 1983 and subsequently modified and incorporated into FIGO staging in 2002 has been used to identify a high-risk group more likely to fail single-agent chemotherapy. In our study, the WHO score also correlated with response to chemotherapy even though all high-risk patients were initially treated with multiagent chemotherapy upon presentation to our center. World Health Organization scores ranged from 0–20 (267 from 0–2; 93 from 3–6; 30 from 7–10; 9 from 11–15; and 9 from 16–20). Patients with a score 7 or more had a 57% rate of resistance to initial chemotherapy, whereas those with a score less than 7 had a 15% rate of resistance to initial chemotherapy (P<.001). Resistance increased with increasing WHO score: 9% for score 0–2; 28% for score 3–6; 40% for score 7–10; 89% for score 11–15; and 78% for score 16–20.

Patients who presented to the Brewer Center after failed treatment at an outside institution had a lower rate of complete response to initial chemotherapy than those treated primarily at our Center. Of the 408 patients treated from 1979 to 2006, 6.6% (27 patients) received cytotoxic drugs at another institution before referral to the Brewer Center. All of those patients were referred to the Brewer Center from gynecologic oncologists not at specialized treatment centers. Of those referred after prior unsuccessful chemotherapy, 63% were resistant to the initial chemotherapy regimen used at our center compared with 17% who were primarily treated at this center (P<.001).

Clinical characteristics found to have a significant effect on chemotherapy response in the univariable analysis were included in the multivariable analysis. Factors found to be independently associated with resistance to first-line chemotherapy included duration of disease more than 4 months (P=.036), presence of metastatic disease (P<.001), site of metastatic disease other than the lung and/or vagina (P=.004), and secondary treatment at the Brewer Center (P<.001).

DISCUSSION

The overall worldwide survival rate approaches 100% for low-risk GTN and 81–89% for high-risk GTN.13–20 Commonly cited risk factors affecting curability include stage, histologic type (choriocarcinoma compared with invasive mole), hCG level, duration of disease, site of metastases, antecedent pregnancy type, and extent of prior treatment.

Survival for patients with GTN has improved from 1962–1978 to 1979–2006 at the John I. Brewer Trophoblastic Disease Center. Improvements in survival through 1978 were attributed to development and use of sensitive hCG assays to monitor disease, identification of high-risk factors that allowed individualization of treatment, involvement of specialized treatment centers in treatment plans, and aggressive use of multiagent chemotherapy regimens to treat high-risk patients.3

The overall survival rate for patients with GTN treated at the Brewer Center since 1979 now approaches 98%. Continued success in treating this disease is a result of increased use of more effective initial chemotherapy regimens such as EMA-CO, improved salvage chemotherapy, and the aggressive use of individualized multimodality treatment, including surgery and radiation therapy, in the highest-risk patients.10–12,21 In addition to improved treatment, earlier diagnosis may play a role in improved survival, because patients with a longer duration of disease may also present with higher rates of metastatic disease and higher hCG levels. A slightly lower percentage of patients with metastatic disease were diagnosed within 4 months of the antecedent pregnancy in the earlier time period when compared with the later time period (67% compared with 76%). In the earlier time period, 48% (174 of 359) of patients presented with metastatic disease compared with 27% (110 of 408) patients who had metastatic disease in the later time period. Pretreatment hCG level more than 100,000 milliinternational units/mL was present in 37% (63 of 170) compared with 24% (26 of 110) of patients with metastatic disease in the earlier and later time periods, respectively. When comparing patients who presented with metastatic disease between time periods, however, better outcomes in the later time period indicate an improvement in multimodality treatment for advanced disease that contributed to improved overall survival. The survival rate increased from 37% to 68% between the two time periods for patients with metastatic disease to sites other than the lung and/or vagina, and from 78% to 92% for patients with any metastatic disease.

Data from this series of patients treated between 1979 and 2006 is reported in terms of resistance to initial chemotherapy regimen due to the near complete survival of patients during this time period. This study provides an updated set of characteristics of high-risk patients who require aggressive initial multiagent chemotherapy with or without surgery and radiotherapy, close monitoring, and a low threshold for referral to a center specializing in the treatment of GTN. When compared with prognostic factors determined from the earlier time period, pretreatment hCG was no longer associated with chemoresistance, whereas presence of any metastatic disease and WHO score of 7 or more were added to the significant prognostic factors. The present study confirms earlier reports from our center, which noted a poorer prognosis in patients with a diagnosis of choriocarcinoma, with metastatic disease to sites other than the lung and/or vagina, and with failed treatment at other institutions.2,3

The WHO score assigned to this series of patients accurately predicted resistance to initial chemotherapy, with a decrease in complete response rate when the score was 7 or more. The FIGO stage also correlated with complete response rate to initial chemotherapy: 88% for stage I, 75% for stage II, 69% for stage III, and 24% for stage IV (P<.001). Several components of the WHO score, including term pregnancy as the antecedent pregnancy event and duration of disease more than 4 months significantly affected resistance to initial chemotherapy. Pretreatment hCG was not associated with chemotherapy resistance.

The survival rate has improved over time for patients with GTN treated at our center, from 88.6% in 1962–1978 to 97.8% in 1979–2006. Improved survival rate in patients with advanced disease is demonstrated by the increase in survival rate in patients with brain metastases (50% to 75%) and with liver or other intraabdominal metastases (0 to 73%) between the 1962–1978 and 1979–2006 time periods, respectively. Even those patients with metastatic disease to the liver and gastrointestinal tract now have a 67% survival rate. The challenge currently is to provide all patients with appropriate therapy for the greatest chance of cure. The Brewer Center more frequently provided appropriate initial chemotherapy than institutions without specialized trophoblastic disease centers. Patients referred to our center after initial treatment failure at other institutions most commonly received the wrong drug schedule or frequency for low-risk disease or single-agent instead of multiagent chemotherapy for high-risk disease. Centers specialized in treatment of GTN may also be more adept at applying salvage chemotherapy, often in conjunction with surgical resection of sites of persistent tumor in a timely fashion which results in improved survival.1,22 Factors that are known to increase resistance to chemotherapy include the presence of metastatic disease, a clinicopathologic diagnosis of choriocarcinoma, antecedent term gestation, prior unsuccessful chemotherapy, metastatic site other than the lung or vagina, duration of disease more than 4 months, and WHO score of 7 or more. The presence of these characteristics suggests provision of aggressive multiagent chemotherapy, often combined with surgery and brain radiation, to maximize the chance for cure in these patients with high-risk GTN.

Specific management recommendations from our center for nonmetastatic and low-risk metastatic (WHO score less than 7) GTN are the provision of a 5-day methotrexate chemotherapy regimen, with transition to 5-day dactinomycin if resistance develops or adverse effects are not tolerated. When single-agent therapy fails, referral to a specialized trophoblastic disease center is appropriate. For high-risk metastatic disease (WHO score of 7 or more), the recommended initial chemotherapy is EMA-CO, with transition to platinum-etoposide combinations with bleomycin, ifosfamide, or paclitaxel if resistance to initial therapy occurs. Referral is appropriate for any high-risk patient. However, referral is highly recommended after failure of initial therapy in this group of patients.

REFERENCES

1. Soper JT. Gestational trophoblastic disease. Obstet Gynecol 2006;108:176–87.
2. Lurain JR, Casanova LA, Miller DS, Rademaker AW. Prognostic factors in gestational trophoblastic tumors: a proposed new scoring system based on multivariate analysis. Am J Obstet Gynecol 1991;164:611–6.
3. Lurain JR, Brewer JI, Torok EE, Halpern B. Gestational trophoblastic disease: treatment results at the Brewer Trophoblastic Disease Center. Obstet Gynecol 1982;60:354–60.
4. Kohorn EI. The new FIGO 2000 staging and risk factor scoring system for gestational trophoblastic disease: description and critical assessment. Int J Gynecol Cancer 2001;11:73–7.
5. Lurain JR. Pharmacotherapy of gestational trophoblastic disease. Expert Opin Pharmacother 2003;11:2005–17.
6. Lurain JR, Elfstrand EP. Single-agent methotrexate chemotherapy for treatment of nonmetastatic gestational trophoblastic tumors. Am J Obstet Gynecol 1995;172:574–9.
7. Roberts JP, Lurain JR. Treatment of low-risk metastatic gestational trophoblastic tumors with single-agent chemotherapy. Am J Obstet Gynecol 1996;174:1917–23.
8. Lurain JR, Brewer JI. Treatment of high-risk gestational trophoblastic disease with methotrexate, actinomycin D, and cyclophosphamide chemotherapy. Obstet Gynecol 1985;65:830–6.
9. Escobar PF, Lurain JR, Singh DK, Bozorgi K, Fishman DA. Treatment of high-risk gestational trophoblastic neoplasia with etoposide, methotrexate, actinomycin D, cyclophosphamide, and vincristine chemotherapy. Gynecol Oncol 2003;91:552–7.
10. Lurain JR, Nejad B. Secondary chemotherapy for high-risk gestational trophoblastic neoplasia. Gynecol Oncol 2005;97:618–23.
11. Small W Jr, Lurain JR, Shetty RM, Huang CF, Applegate GL, Brand WN. Gestational trophoblastic disease metastatic to the brain. Radiology 1996;200:277–80.
12. Lurain JR, Singh DK, Schink JC. Role of surgery in the management of high-risk gestational trophoblastic neoplasia. J Reprod Med 2006;51:773–6.
13. Berkowitz RS, Goldstein DP, Bernstein MR. Ten year’s experience with methotrexate and folinic acid as primary therapy for gestational trophoblastic disease. Gynecol Oncol 1986;23:111–8.
14. Ngan S, Seckl MJ. Gestational trophoblastic neoplasia management: an update. Curr Opin Oncol 2007;19:486–91.
15. Moodley M, Tunkyi K, Moodley J. Gestational trophoblastic syndrome: an audit of 112 patients. A South African experience. Int J Gynecol Cancer 2003;13:234–9.
16. Lok CA, Ansink AC, Grootfaam D, van der Velden J, Verheijen R, ten Kate-Booij M. Treatment and prognosis of post term choriocarcinoma in the Netherlands. Gynecol Oncol 2006;103:698–702.
17. El-Lamie IK, El Sayed HM, Badawie AG, Bayomi WA, El-Ghazaly HA, Khalaf-Allah AE, et al. Evolution of treatment of high-risk metastatic gestational trophoblastic tumors: Ain Shams University experience. Int J Gynecol Cancer 2006;16:866–74.
18. Szigetvari I, Szepesi J, Vegh G, Batorfi J, Arato G, Gati I, et al. 25 years’ experience in the treatment of gestational trophoblastic neoplasia in Hungary. J Reprod Med 2006;51:841–8.
19. Garrett AP, Garner EO, Goldstein DP, Berkowitz RS. Methotrexate infusion and folinic acid as primary therapy for nonmetastatic and low-risk metastatic gestational trophoblastic tumors. 15 years of experience. J Reprod Med 2002;47:355–62.
20. Chakrabarti BK, Mondal NR, Chatterjee T. Gestational trophoblastic tumor at a tertiary level cancer center: a retrospective study. J Reprod Med 2006;51:875–8.
21. Lurain JR, Singh DK, Schink JC. Primary treatment of metastatic high-risk gestational trophoblastic neoplasia with EMA-CO chemotherapy. J Reprod Med 2006;51:767–72.
22. Lurain JR. Advances in management of high-risk gestational trophoblastic tumors. J Reprod Med 2002;47:451–9.
Figure
Figure
© 2008 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.