EMA/CO for High-Risk Gestational Trophoblastic
Tumors: Results From a Cohort of 272 Patients
By M. Bower, E.S. Newlands, L. Holden, D. Short, C. Brock, G.J.S. Rustin, R.H.J. Begent, and K.D. Bagshawe
Purpose: To evaluate the results of etoposide, metho-
trexate, and doctinomycin alternating with cyclophos-
phamide and vincristine (EMA/CO) chemotherapy in
women with high-risk gestational trophoblastic tumors
(GTT) and to document the middle- and long-term toxicity
of the regimen.
Patients and Methods: A total of 272 consecutive
women with high-risk GTT, including 121 previously
treated patients, were treated with weekly EMA/CO. The
median follow-up duration is 4.5 years (range, I to 16).
Results: The cumulative 5-year survival rate is 86.2%
(95% confidence interval, 81.9% to 90.5%). No deaths
from GTT have occurred later than 2 years after the start
of EMA/CO. In a multivariate model, adverse prognostic
factors were the presence of liver metastases (P <
.0001), interval from antecedent pregnancy (P< .0001),
brain metastases (P = .0008), and term delivery of ante-
cedent pregnancy (P = .045). There were 11 (4%) early
THE MAJORITY OF WOMEN with gestational tro-
phoblastic disease will be cured by single-agent
methotrexate and folinic acid chemotherapy; however, a
small number will have high-risk disease or will relapse.
High-risk trophoblastic disease refers to gestational tropho-
blastic tumors (GTT) that are unlikely to be cured by sin-
gle-agent chemotherapy. Several scoring systems to iden-
tify adverse prognostic factors have been used in this
disease. We have used the system devised by Bagshawe'
in 1976, which is similar to the one adopted by the World
Health Organization (WHO). 2 The scoring is based on
patient age, parity and type of antecedent pregnancy, inter-
val since antecedent pregnancy, serum human chorionic
gonadotropin (hCG) level, paternal and maternal blood
groups, number and size of metastases, largest tumor mass,
and prior chemotherapy. Following the discovery of the
marked activity of etoposide in GTr,3'4 this drug was incor-
porated in 1979 into the etoposide, methotrexate, and dacti-
nomycin alternating with cyclophosphamide and vincris-
tine (EMA/CO) regimen for high-risk disease. This
From the Medical Oncology Unit, Charing Cross Hospital; De-
partment of Clinical Oncology, Royal Free Hospital, London; and
Mount Vernon Centrefor Cancer,Mount Vernon Hospital, Middle-
sex, United Kingdom.
Submitted September 16, 1996; accepted March 28, 1997.
Address reprintrequests to M. Bower, PhD, MRCP, Fulham Pal-
ace Road, London W6 8RF, United Kingdom; Email m.bower@
cxwms.ac.uk.
1997 by American Society of Clinical Oncology.
0732-183X/97/1507-0001$3.00/0
2636 Journal of Clinical Oncology, Vol 15, No 7 (July), 1997: pp 2636-2643
Downloaded from ascopubs.org by 125.161.44.184 on March 7, 2017 from 125.161.044.184
Copyright 2017 American Society of Clinical Oncology. All rights reserved.
deaths, while 213 patients (78%) achieved a complete
remission. Forty-seven (17%) developed drug resistance
to EMA/CO, of whom 33 (70%) were salvaged by further
cisplatin-based chemotherapy and surgery. Two women
developed acute myeloid leukemia, two cervical malig-
nancy, and one gastric adenocarcinoma after EMA/CO.
More than half (56%) of the women who had fertility-
conserving surgery and who have been in remission at
least 2 years have become pregnant since the completion
of EMA/CO, with 112 live births, including three infants
with congenital abnormalities.
Conclusion: EMA/CO is an effective and well-toler-
ated regimen for high-risk GTT. More than half of the
women will retain their fertility; however, there is a small
but significant risk of second malignancy.
J Clin Oncol 15:2636-2643. 1997 by American So-
ciety of Clinical Oncology.
alternating combination chemotherapy regimen requires
only one night of hospitalization every 2 weeks and re-
placed a 9-day, seven-drug combination (hydroxyurea, ac-
tinomycin D, methotrexate, folinic acid, vincristine, cyclo-
phosphamide, and adriamycin [CHAMOCA]) that induced
prolonged myelosuppression. Initial publications reported
high response rates with EMA/CO.5 -8 This schedule has
now been used to treat 272 consecutive patients at Charing
Cross Hospital, with a follow-up duration of 1 to 16 years.
Thus, the survival data, as well as late toxicities, are avail-
able for this large cohort of high-risk GTT patients treated
with EMA/CO.
PATIENTS AND METHODS
Between October 1979 and April 1995, 1,047 patients were treated
at Charing Cross Hospital for GTT, including 272 consecutive
women with high-risk disease treated with the EMA/CO regimen.
The median follow-up duration for this cohort of women is 4.5 years
and all have been monitored for a least 1 year. The clinicopathologic
features of the women at presentation are listed in Table 1.
One hundred fifty-one women had received no prior chemotherapy
for GTT, including 140 with high-risk disease at presentation (prog-
nostic score, -- 8) and 11 with intermediate-risk disease (prognostic
score, 5 to 7) according to the WHO prognostic scoring index. The
latter included two women with placental-site trophoblastic disease
(PSTT) and one who presented with a perforated uterus. In addition,
one woman with GTT presented following a normal-term pregnancy;
however, she had experienced a molar pregnancy 2 years previously
and if the trophoblastic disease was related to the molar pregnancy
she would have fallen into the high-risk group.' The remaining seven
women with genuine intermediate-risk disease were treated de novo
with EMA/CO chemotherapy because of difficult geographic access
for medium-risk regimens, which require chemotherapy 5 days a
week.
EMA/CO FOR HIGH-RISK GTr 2637
Table 1. Clinicopathologic Features of Patients survival distributions." The variables found to be significant were
put into a stepwise Cox regression model to establish which were
Characteristic No. %
independently prognostic.12
Total no. 272
Age, years RESULTS
Median 29.1
Range 17-60 Survival
Antecedent pregnancy
The cumulative overall survival rate for all patients at
Hydatidiform mole 120 44
Term delivery 99 36
5 years is 86.2% (95% confidence interval, 81.9% to
Spontaneous abortion 48 18 90.5%) (Fig 1). There have been no deaths from any
Stillbirth 2 1 causes in this cohort later than 5 years following EMA/
Unknown 3 1 CO chemotherapy. Indeed, the latest two deaths that oc-
Previous chemotherapy 121 44
curred 4 and 2.2 years after starting EMA/CO were due
CNS metastases 34 13
Liver metastases 17 6
to a traffic accident and colonic cancer, respectively. The
PSTT histology 9 3 last death due to GTT occurred 2.1 years after the start
of EMA/CO.
Adverse prognostic factors in univariate analysis in-
One hundred twenty-one women had received chemotherapy for cluded the presence of brain metastases (log-rank hetero-
G'IT before they began EMA/CO. Eighty-six had evidence of dis- geneity [LRH] test P = .0004), liver metastases (LRH P
ease resistant to prior chemotherapy at the start of EMA/CO, 50 < .0001), and a term pregnancy as the last known preg-
having received single-agent methotrexate and 36 having received nancy before diagnosis (LRH P = .0005), but not prior
combination chemotherapy. Thirty-five patients were treated with
EMA/CO at relapse following initial treatment to complete remis-
chemotherapy (LRH P = .27), pretreatment serum hCG
sion, 15 having previously been treated with methotrexate alone and level (LRH P = .19), or PSTT histology (LRH P = .07).
20 having been treated with combination chemotherapy. In addition, a prolonged interval separating the antecedent
The EMA/CO regimen is shown in Table 2. Patients with pulmo- pregnancy and the diagnosis was an adverse prognostic
nary metastases received cranial prophylaxis with intrathecal injec- factor (LRH P < .0001). Indeed, the 5-year survival rate
tions of methotrexate 12.5 mg (total dose) every 2 weeks with CO
cycles. Patients with documented cranial metastases received an in-
of women with an interval of more than 24 months is
2
creased dosage of systemic methotrexate of 1 g/m as an intravenous only 67% (95% confidence interval, 55% to 80%) com-
infusion over 24 hours followed by folinic acid rescue (15 mg orally pared with 92% (95% confidence interval, 88% to 96%)
every 6 hours for 12 doses starting 32 hours after methotrexate). for those women with a shorter interval (LRH P < .0001)
Response to treatment was assessed by twice-weekly serum hCG
(Fig 2 and Table 3).
measurement. Chemotherapy was administered until serum hCG lev-
els decreased to within the normal range (< 5 IU/L) and for a further In a Cox proportional hazards model (PHM), the pres-
6 to 8 weeks thereafter. Failure of the serum hCG level to decrease ence of liver metastases (PHM P < .0001), long interval
was used to diagnose resistance to EMA/CO chemotherapy and (PHM P < .0001), brain metastases (PHM P = .0008),
cisplatin-containing combination chemotherapy was introduced for and antecedent term pregnancy (PHM P = .045) were
these patients. In most cases, cisplatin was administered in the EP independent adverse prognostic factors.
schedule (etoposide 150 mg/m2 intravenously and cisplatin 75 mg/
m2) alternating at weekly intervals with the EMA schedule (in most
cases day 2 of EMA was omitted due to myelosuppression). Table 2. EMA/CO Regimen

EMA
Survey
Day 1
One hundred sixty patients had completed chemotherapy more Dactinomycin 0.5 mg IV bolus
2
than 5 years previously. Questionnaires were sent to 134 of these Etoposide 100 mg/m in 250 mL NS over 30 minutes
2
women (16 had died and 10 did not reside in the United Kingdom). Methotrexate 300 mg/m in 1 L NS over 12 hours
They were asked to complete a simple questionnaire that concerned Day 2
fertility, pregnancies, and major illnesses since treatment. Responses Dactinomycin 0.5 mg IV bolus
were obtained from 96 women (71%). Additional information con- Etoposide 100 mg/m2 in 250 mL NS over 30 minutes
cerning fertility and subsequent pregnancies was available for pa- Folinic acid 15 mg orally/IM twice a day for 2 days 24
tients on serologic follow-up evaluation at Charing Cross Hospital. hours after start of methotrexate
CO
Statistical Methods Vincristine 0.8 mg/m2 IV bolus (maximum, 2 mg)
Cyclophosphamide 600 mg/m2 in 250 mL NS over 30 minutes
Survival was calculated from the first day of treatment until death
Repeat EMA alternating weekly with CO to serologic remission plus a
or date of last follow-up visit. Overall survival duration curves were
further 6-8 weeks therapy
plotted according to the method of Kaplan and Meier.'o The log-
rank method was used to test for the significance of differences in Abbreviations: IV, intravenous; NS, normal saline; IM, intramuscularly.

Downloaded from ascopubs.org by 125.161.44.184 on March 7, 2017 from 125.161.044.184

Copyright 2017 American Society of Clinical Oncology. All rights reserved.
2638
1111111111111111I1 ,.,I.,,I,,,I,.I,.I,
.8
u .4
.2
0
' ' I I I I I I I I I
0 2 4 6 8 10
Years
Response to EMA/CO
Survival is the most appropriate end point for this anal-
ysis due to the good follow-up data. There were 11 (4%)
early deaths within 4 weeks of the start of EMA/CO and,
overall, 47 (17%) patients required a change of treatment,
while 213 (78%) achieved a complete remission on EMA/
CO. One patient died of a ruptured aorta sustained at
thoracotomy for the management of an anaerobic lung
abscess after the third course of EMA/CO while re-
sponding serologically to chemotherapy (Table 4).
Drug Resistance to EMA/CO
Twenty-one women (14%) who had received no prior
chemotherapy needed to switch therapy because of
drug resistance to EMA/CO. All were treated with cis-
platin-containing combination chemotherapies (17 re-
ceived the EP/EMA regimen and four received earlier
cisplatin-based combination chemotherapy schedules),
and eight also had surgical resections of tumor sites
(four of these had hysterectomies for PSTT). Sixteen
of 21 patients (76%) underwent successful salvage
treatment and are alive in remission. The remaining
five women all died of GTT, despite surgical resections
and multiple chemotherapy regimens (including high-
Downloaded from ascopubs.org by 125.161.44.184 on March 7, 2017 from 125.161.044.184
Copyright 2017 American Society of Clinical Oncology. All rights reserved.
BOWER ET AL
Fig 1. Kaplan-Meier overall
survival plot of all 272 patients
treated with EMA/CO.
I I, '
. . I I
12 14 16 18
dose chemotherapy and progenitor stem-cell rescue in
two).
Twenty-six women (21%) who had received chemo-
therapy before EMA/CO developed serologic evidence
of drug resistance and switched therapy. This included
seven of 65 women (11%) who had received prior
methotrexate only and 19 of 56 (34%) who had received
combination chemotherapy. All were treated with cis-
platin-containing combination chemotherapy and 15
also underwent surgery. Seventeen women (65%) are
alive in serologic remission, while nine (35%) died of
drug-resistant GTT.
Relapse
Twenty patients who achieved complete remission
have relapsed, including 16 of 213 (8%) who attained
a complete remission on EMA/CO alone and four who
had switched therapy due to drug resistance to EMA/
CO. The 20 women who relapsed included eight who
had received no chemotherapy before EMA/CO. Six
(30%) have died following relapse, while 14 (70%)
have been salvaged with a combination of chemother-
apy (in most cases EP/EMA) and surgery (a total of
nine hysterectomies, 11 thoracotomies, and one crani-
EMA/CO FOR HIGH-RISK GTT 2639

(1) Prior Chemotherapy (a) (2) Brain metastases (a)

or none (b) or none (b)

I I I I, i . =I I i. i , r . . . . . = i i , I
a. n=121 1 b. n=237

b. n=151 .6- a. n-34

.6
t .6
.4
.4
.4-
.2 .2 -
E -
0- 0-

0 2 4 6 8 10 12 14 16 18 0 2 4 6 8 10 12 14 16 18
Years Years

(3) Liver metastases (a) (4) PSTT histology (a)

or none (b). or not (b).

b. n=254 1 b. n-263
Fig 2. Univariate analysis of
survival according to (1) prior
a. n-S9
chemotherapy, (2) presence of "L.6 -
brain metastases, (3) presence of
liver metastases, (4) PSTr histol- .4
E - a. n-17
.4
ogy, (5) interval since anteced-
ent pregnancy, and (6) anteced- .2 .2
ent term pregnancy.
0- O-

0 2 4 6 8 10 12 14 16 18 0 2 4 6 8 10 12 14 16 18
Years Years

(5) Interval from antecedent pregnancy

(a) <4months, (b)4-7 months, (6) Grouping variable: Term antecedent
(c) 7-12 months, (d) 12-24 months pregnancy (a) or not (b)
& (e) >24 months.
b. n-1 71
a. n-10 1-
1-
d. n-4Z
n-38
.8 c.n-29 .8
a.n-101
.6
e. n=54 " .6-
.4
.4
C)
.2
-
0
I I . i = i E I " I = j , =

0 2 4 6 8 10 12 14 16 18 0 2 4 6 8 10 12 14 16 18
Years Years

otomy). The median time to relapse following comple- of polyposis coli and this is thought to be the major
tion of chemotherapy was 4 months (range, 1 to 72). etiologic factor, rather than chemotherapy. A second
woman died of acute myeloid leukemia French-Ameri-
Deaths can-British (FAB) subtype M1 16 months after she com-
There have been 34 deaths in this cohort of 272 patients pleted EMA/CO, having received a total etoposide dosage
(12.5%). Thirty-one patients died of GTT (11 early of 3.0 g/m 2 . Karyotypic analysis of the blast population
deaths, one treatment-related, six relapsed, and 13 refrac- demonstrated a clonal t(9:11)(p:q23) translocation-an
tory disease) and three patients have died of other causes. abnormality found frequently in secondary leukemia fol-
One woman died of colonic cancer 1.7 years after she lowing topoisomerase II inhibitor therapy. A third woman
completed treatment with EMA/CO. There was a history died in a traffic accident 4 years after EMA/CO. All three

Downloaded from ascopubs.org by 125.161.44.184 on March 7, 2017 from 125.161.044.184

Copyright 2017 American Society of Clinical Oncology. All rights reserved.
2640 BOWER ET AL

Table 3. Univariate Analysis of Survival by Variables

n Deaths 5-Year Survival (%) 95% CI X2. P*
Prior chemotherapy
Yes 121 12 88.9 t 5.9 1.20 .27
No 151 22 84.2 - 6.0
Brain metastases
No 237 23 89.1 _ 4.3 12.5 .0004
Yes 34 10 69.2 15.9
Liver metastases
No 254 22 90.3 _ 3.9 71.9 < .0001
Yes 17 11 31.5 _ 23.3
PSTT histology
No 263 31 87.0+ 4.3 3.3 .07
Yes 9 3 64.8 32.3
Antecedent term pregnancy
No 171 12 92.2 - 4.3 12.1 .0005
Yes 101 21 77.0 8.6
Interval between antecedent pregnancy and diagnosis, months
0-4 107 2 97,5 t 3.3 32.0 < .0001t
4-7 38 5 84 1 _ 13.1
7-12 29 5 81.1 - 14.9
12-24 42 3 92.7- 8.0
> 24 54 17 67.3 -! 12.7
Abbreviation: CI, confidence interval.
*Log-rank test unless otherwise indicated.
tTrend test.

women were in serologic remission of GTT at the time
of death.
Late Toxicity
Second malignancy. In addition to the two deaths
from second malignancies (vide supra), another woman
developed acute myeloid leukemia of FAB subtype M5
15 months after she completed EMA/CO (total etoposide
dosage, 2.0 g/m 2) while in serologic remission of GTT.
No clonal karyotypic abnormality was identified. She was
successfully treated with chemotherapy and a matched
unrelated bone marrow allograft and remains in remission
of both leukemia and GTT 4 years later.
One woman has developed gastric adenocarcinoma at
the age of 52 years, 13 years after she completed initial
EMA/CO therapy and 6 years after further EMA/CO for
relapsed disease. She remains in serologic remission of
GTT. Two other women have reported that they have
received treatment for cervical malignancy (one invasive
and one intraepithelial 2 and 1 years following completion
of chemotherapy). A further two women have reported
benign breast lumps and one woman received treatment
for a benign ovarian cyst.
Other toxicities. Other potential late treatment-related
toxicities identified from the survey responses from 96 pa-
tients included thyroid dysfunction (n = 2), hypoparathy-
roidism (n = 1), epilepsy following craniotomy (n = 1),
and avascular necrosis of the femoral heads (n = 1).

Table 4. Response to EMA/CO

Total No Prior Chemotherapy Prior Chemotherapy

Response No. % No. % No. %

Early death 11 4 11 7 0 0
Treatment-related death 1 0.4 1 0.7 0 0
Progressive disease 47 17 21 14 26 21
Complete remission 213 78 118 78 95 79

Total 272 151 121

Dead (any cause) 34 13 22 14 12 10

Relapse 20 7 8 5 12 10

Downloaded from ascopubs.org by 125.161.44.184 on March 7, 2017 from 125.161.044.184

Copyright 2017 American Society of Clinical Oncology. All rights reserved.
EMA/CO FOR HIGH-RISK G1T
Table 5. Fertility in Patients More Than 2 Years After Completion of
EMA/CO Chemotherapy
Dead
Lost to long-term follow-up
Total dead & lost to follow-up
Hysterectomy
Laparoscopic sterilization
High-dose chemotherapy*
latrogenic infertility
Not attempted
No pregnancies
Unsuccessful
Uncertain status
At least 1 live birth
TOPs only
Miscarriages only
Currently pregnant
Fertile
Total
* For acute myelogenous leukemia.
Abbreviation: TOP, termination of pregnancy.
Fertility and Pregnancy
One hundred ninety-two patients are alive and free of
disease at least 2 years after completing chemotherapy
and continue on serologic follow-up at Charing Cross
Hospital. Forty women (21%) have either had a hysterec-
tomy (n = 36) or laparoscopic sterilization. Eighty-five
(56%) of the remaining 152 women are or have been
pregnant since the completion of treatment (70 have had
- one live birth, four have had terminations only, three
have had miscarriages only, and three are currently preg-
nant). The other 67 women include 13 who reported that
they have not attempted to become pregnant and one who
has received high-dose chemotherapy and a bone marrow
allograft (Table 5). There have been a total of 112 live
births in this cohort, including three infants with congeni-
tal abnormalities (one Down syndrome, one clubfoot, and
one umbilical hernia). This is a similar incidence to that
reported in other series.13
DISCUSSION
As previously reported, EMA/CO is effective therapy
for high-risk GTT. The overall cumulative survival rate
at 5 years is 86.2% (95% confidence interval, 81.9% to
90.5%) and the disease-specific survival rate at 5 years
is 87.7%. This compares with previously published re-
ports of EMA/CO from smaller cohorts that documented
overall survival rates of 85%,6 88%,7 and 100%.8
Prior treatment with chemotherapy does not signifi-
cantly alter survival. However, all of the early deaths
and the single treatment-related death occurred in the
Downloaded from ascopubs.org by 125.161.44.184 on March 7, 2017 from 125.161.044.184
Copyright 2017 American Society of Clinical Oncology. All rights reserved.
2641
previously untreated group of women. In this group of
patients, early deaths are the major cause of mortality.
25 The early deaths were all related to the extent of disease
17 at initiation of therapy. Earlier diagnosis may reduce these
42
deaths; however, all 11 followed term deliveries or spon-
36
4
taneous abortions, rather than hydatidiform moles, and so
1 these patients were not on the surveillance program. To
41 reduce the deaths in untreated patients, improved respira-
13 tory management is required, since over half of the deaths
49
were due to respiratory complications.14 Mechanical ven-
4
66
tilation in these patients presents problems, as barotrauma
75 can cause intrapulmonary hemorrhage because the pul-
4 monary metastases are highly vascular. The use of an
3 intravascular oxygenator may be an alternative approach,
3
as anticoagulation is not necessary for this procedure.
85
The other fatalities among previously untreated patients
234 include five patients with drug-resistant primary disease
(including one patient with PSTT) and two patients with
drug-resistant disease at relapse (including one patient
with PSTT). All three deaths due to causes other than
GTT occurred in patients with untreated disease.
In contrast, all deaths in previously treated patients
were a consequence of drug resistance. Nine women had
resistant disease and never achieved remission after start-
ing EMA/CO, while three achieved remission, but died
of drug-resistant disease at relapse. Drug resistance to
EMA/CO occurs more frequently after failure of combi-
nation chemotherapy (34%) than failure of single-agent
chemotherapy (11%). Three quarters of untreated women
and two thirds of previously treated women who devel-
oped resistance to EMA/CO were successfully salvaged
by a combination of cisplatin-based combination chemo-
therapy and surgery, so that the initial treatment of low-
risk disease with methotrexate does not make salvage
difficult if drug resistance emerges. The ability to salvage
the majority of patients after initial treatment failure is one
of the unusual features of GTT and, given the relatively
successful outcome in this disease, it is clinically justifi-
able to try to achieve a complete remission even after
several treatment failures. However, improvements in the
therapy of drug-resistant GTT are needed to improve fur-
ther the survival in this group of women. High-dose che-
motherapy followed by autologous progenitor cell rescue
has been advocated."5 Two women with drug-resistant
GTT in this series underwent this procedure, but neither
achieved a meaningful response and both died of rapidly
progressive disease 1 and 4 months after transplantation.'16
A number of new cytotoxic agents are available (particu-
larly taxanes and topoisomerase I inhibitors), but they
have not yet been evaluated in drug-resistant GTT on
account of the small number of patients with this problem.
2642
Many centers have used one of several prognostic scor-
ing systems to identify patients with high-risk disease.
Both the Charing Cross Hospital system and that recom-
mended by the WHO are similar.12 The results reported
in this series confirm that the major risk factors in both
of these scoring systems (interval from antecedent preg-
nancy > 12 months, initial serum hCG level > 105 IU/
L, CNS metastases, > eight metastases, and failure of
prior multiagent chemotherapy) are important. However,
both of these prognostic scoring schemes need to take
into account developments in the last decade. PSTT is
a rare and unpredictable variant of trophoblastic tumor,
although it did not have a significantly worse prognosis
in this series. The molecular genetics, clinical features,
treatment response, and outcome of the PSTT patients
included in this cohort are described fully elsewhere."7
The presence of liver metastases has been underestimated
as an adverse prognostic variable and should be included
along with CNS metastases as a major risk factor. Since
the presence of both CNS and liver metastases carries an
even worse prognosis," each site of disease should be
scored independently (ie, a WHO score of 4 x 2 if both
CNS and liver metastases are present). Furthermore, a
prolonged interval of 2 years or more between antecedent
pregnancy and diagnosis is a major adverse prognostic
variable that is underscored in the current system. Our
own data base is probably large enough to define the
relative importance of the major adverse risk factors to
refine the current prognostic scoring systems. However,
the results will need to be confirmed independently on
data sets from other centers and should be evaluated pro-
spectively at several centers.
The published early toxicities of EMA/CO include alo-
pecia, nausea, reversible neurotoxicity, and myelosup-
pression.' 6 The late toxicities reported here include sec-
ond malignancies, most prominently acute myeloid
leukemia, which has been linked particularly to etoposide
administration. The mechanism of topoisomerase II inhib-
itor-induced leukemogenesis is obscure, but the myeloid
blasts frequently have rearrangements of the MLL gene
on chromosome 11q23.' 9 The two cervical tumors and
the gastric adenocarcinoma are not well-recognized sec-
ondary tumors and are possibly incidental associations in
REFERENCES
1. Bagshawe KD: Risk and prognostic factors in trophoblastic
neoplasia. Cancer 38:1373-1385, 1976
2. World Health Organization Scientific Group on Gestational
Trophoblastic Diseases: Gestational Trophoblastic Diseases. Ge-
neva, Switzerland, WHO, 1983
3. Newlands E, Bagshawe K: Antitumour activity of epipodophy-
Ilin derivative VP16-213 (etoposide: NSC-141540) in gestational
choriocarcinoma. Eur J Cancer 16:401-405, 1980
Downloaded from ascopubs.org by 125.161.44.184 on March 7, 2017 from 125.161.044.184
Copyright 2017 American Society of Clinical Oncology. All rights reserved.
BOWER ET AL
this cohort. Among 1,394 women treated with chemother-
apy for GTT at Charing Cross Hospital, the incidence of
second malignancies was significantly increased, with 37
cases occurring, compared with an expected rate of 24
(P = .011).20 The greatest increase in second tumors in
this large cohort was myeloid leukemia, with lesser but
significant increases in melanoma, colon, and breast can-
cer. Other late events reported here include thyroid and
parathyroid dysfunction, although the relationship to the
chemotherapy is uncertain.
Over half of the women monitored for at least 2 years
from the start of EMA/CO who had fertility-preserving
surgery have become pregnant. Fertility is a major issue
for many of the women with high-risk GTT and the con-
servation of fertility with this regimen is valuable. By
comparison, in an analysis of 445 women treated with
chemotherapy for GTT at Charing Cross Hospital, 97%
of those who wished to become pregnant succeeded and
86% of these had at least one live birth.21
However, there have been three congenital abnormali-
ties among the 112 babies born to these women and the
role of the chemotherapy in the teratogenicity causes con-
cern. This compares with reported rates of 0.8% to 3.4%
congenital abnormalities in infants born to mothers
treated with all forms of chemotherapy for GTT.21-24 All
of our patients are advised that pregnancy is contraindi-
cated for the first year after the completion of chemother-
apy to allow serologic monitoring of hCG levels to con-
firm remission and to minimize the potential teratogenic
effects of chemotherapy.
The EMA/CO regimen is a well-tolerated and effective
treatment for high-risk GTT that conserves fertility in a
high proportion of women. The medium- to long-term
toxicity recorded so far with the EMA/CO schedule is
acceptable in view of the high-risk nature of the disease
being treated. However, the small increase in second tu-
mors that is associated with combination chemotherapy
for this disease is worrying and emphasizes the need to
develop effective chemotherapeutic agents with a better
long-term toxicity profile than the agents in current clini-
cal use. In our analysis of second tumors in women treated
for GTT, only methotrexate used as a single agent is free
of the risk of inducing second malignancies. 20
4. Newlands E, Bagshawe K: Role of VP16-213 (etoposide: NSC-
141540) in gestational choriocarcinoma. Cancer Chemother Pharma-
col 7:211-214, 1982
5. Newlands ES, Bagshawe KD, Begent RHJ, et al: Developments
in chemotherapy for medium and high risk patients with gestational
trophoblastic tumours (1974-1984). Br J Obstet Gynaecol 93:63-69,
1986
6. Newlands ES, Bagshawe KD, Begent RHJ, et al: Results with
EMA/CO FOR HIGH-RISK GTr
the EMA/CO (etoposide, methotrexate, actinomycin D, cyclophos-
phamide, vincristine) regime in high risk gestational trophoblastic
tumours, 1979 to 1989. Br J Obstet Gynaecol 98:550-557, 1991
7. Bolis G, Belloni C, Bonazzi C, et al: EMA/CO regimen in
high-risk gestational trophoblastic tumor (GTT). Obstet Gynaecol
Surv 44:394-395, 1989
8. Schink J, Singh D, Rademaker A, et al: Etoposide, methotrex-
ate, actinomycin D, cyclophosphamide and vincristine for the treat-
ment of metastatic high-risk gestational trophoblastic disease. Obstet
Gynecol 80:817-820, 1992
9. Fisher RA, Soteriou B, Meredith L, et al: Previous hydatidi-
form mole identified as the causative pregnancy of choriocarcinoma
following birth of normal twins. Int J Gynecol Cancer 5:64-70, 1995
10. Kaplan E, Meier P: Nonparametric estimation from incom-
plete observations. J Am Stat Assoc 53:457-481, 1958
11. Peto R, Pike M, Armitage P, et al: Design and analysis of
randomised clinical trials requiring prolonged observation of each
patient: H analysis and examples. Br J Cancer 35:1-39, 1977
12. Cox D: Regression models and life tables. J R Stat Soc B
34:187-220, 1972
13. Berkowitz RS, Bernstein MR, Laborde O, et al: Subsequent
pregnancy experiences in patients with gestational trophoblastic dis-
ease. J Reprod Med 39:228-232, 1994
14. Kelly M, Rustin G, Ivory C, et al: Respiratory failure due to
choriocarcinoma: A study of 103 dyspneic patients. Gynecol Oncol
38:149-154, 1990
15. Lotz J-P, Andre T, Donsimoni R, et al: High dose chemother-
apy with ifosfamide, carboplatin, and etoposide combined with autol-
Downloaded from ascopubs.org by 125.161.44.184 on March 7, 2017 from 125.161.044.184
Copyright 2017 American Society of Clinical Oncology. All rights reserved.
2643
ogous bone marrow transplantation for the treatment of poor-progno-
sis germ cell tumors and metastatic trophoblastic disease in adults.
Cancer 75:874-885, 1995
16. Bower M, Brock C, Fisher R, et al: Gestational choriocarci-
noma. Ann Oncol 6:503-508, 1995
17. Bower M, Paradinas F, Fisher R, et al: Placental site tropho-
blastic tumor: Molecular analysis and clinical experience. Clin Can-
cer Res 2:897-902, 1996
18. Crawford R, Newlands E, Rustin G, et al: Gestational tropho-
blastic disease with liver metastases: The Charing Cross experience.
Br J Obstet Gynaecol 104:105-109, 1997
19. Bower M, Parry P, Gibbons B, et al: Human trithorax gene
rearrangements in therapy-related acute leukaemia after etoposide
treatment. Leukemia 8:226-229, 1994
20. Rustin G, Newlands E, Lutz J-M, et al: Combination but not
single agent methotrexate chemotherapy for gestational trophoblastic
tumors increases the incidence of second tumors. J Clin Oncol
14:2769-2773, 1996
21. Rustiri GJS, Booth M, Dent J, et al: Pregnancy after cytotoxic
chemotherapy gestational trophoblastic tumours. Br Med J 288:103-
106, 1984
22. Van Thiel D, Ross G, Lipsett M: Pregnancies after chemother-
apy of trophoblastic neoplasms. Science 169:1326-1327, 1970
23. Ngan H, Wong L, Ma H: Reproductive performance of pa-
tients with gestational trophoblastic disease in Hong Kong. Acta
Obstet Gynecol Scand 67:11-14, 1988
24. Feldman S, Goldstein D, Berkowitz R: Low-risk metastatic
gestational trophoblastic tumors. Semin Oncol 22:166-171, 1995