1074

Gestational trophoblastic tumours following initial

diagnosis of partial hydatidiform mole

patients registered with an initial diagnosis of

partial hydatidiform mole (PHM) subsequently
required chemotherapy for a gestational
trophoblastic tumour. In a retrospective review by
histopathological examination and measurement
of DNA ploidy, the diagnosis was confirmed as
PHM in 5 cases and revised to complete
hydatidiform mole in 4; in 2 cases there was no
evidence of a molar pregnancy. 4 of the patients
11

with PHM had no other known pregnancy before

the gestational trophoblastic tumour and in 2 of
these patients the tumour was diagnosed
histologically as choriocarcinoma. Not all patients
in whom PHM was diagnosed at referring
hospitals proved to have the condition. Although
the risk of a patient with PHM requiring
chemotherapy for gestational trophoblastic
tumour is of the order of 1 in 200, compared with 1
in 12 after a complete mole, there is no justification
for excluding a patient from follow-up after the
evacuation of a PHM.

Introduction
In 1973, the Royal College of Obstetricians and
Gynaecologists and the Department of Health and Social
Security established a registration and follow-up scheme for
patients in the United Kingdom who had a clinically or
histologically diagnosed hydatidiform mole. The
pathological classifications, as amended by a World Health
Organisation Scientific group in 1980, are (a) classic or
complete hydatidiform mole (CHM), (b) partial
hydatidiform mole (PHM), and (c) hydropic degeneration.
After evacuation of a molar pregnancy patients are
instructed to send samples of urine or blood at regular
intervals for estimation of human chorionic gonadotropin
(HCG). We have earlier published details of the scheme and
the indications for chemotherapy.2
Both histological and genetic criteria are now used to
distinguish CHM and PHM. There is no evidence of fetal
development in CHM, which is characterised by visible
swelling of most chorionic villi, cistern formation, and
trophoblastic hyperplasia. On chromosomal analysis,
complete moles are diploid; most have 46XX karyotype, and
few 46XY. There are no maternal chromosomes in the
nuclei of CHM, which arise by androgenesis; duplication of
a haploid male gamete,3 or dispermy.4 There is evidence of
fetal development in PHM; normal and hydropic villi are
present with some trophoblastic hyperplasia.5 Genetically,
a

PHM are usually triploid5 and they arise by dispermy-one

maternal and two paternal contributions to the genome.6,7
Malignant complications after a molar pregnancy are
usually associated with complete mole but have been
reported with PHM.8,9
However, most PHM undergo spontaneous resolution.to
DNA ploidy of molar tissue, including formalin-fixed
material, can be determined by flow cytometry and so the
diagnosis of PHM can be confirmed. 11,12 Through the
registration scheme, we have reviewed women with an initial
diagnosis of PHM who subsequently required
chemotherapy for gestational trophoblastic tumour.

Patients and methods

Jan 1, 1979, and Dec 31, 1988, 5989 patients with HM
registered for human chorionic gonadotropin (HCG) follow
at
up
Charing Cross Hospital; 424 (7-1%) subsequently required
chemotherapy. 857 (14 3%) had a diagnosis, made at a referring
hospital, of PHM, 10 of whom (1 2%) later needed chemotherapy.
We studied these 10 patients and 1 additional patient who was
initially registered at Dundee with a referring hospital diagnosis of
PHM and was subsequently referred to Charing Cross Hospital for
treatment. The histological features of the products of molar
pregnancy, and in 3 cases tissue from subsequent curettage, were
reviewed and the DNA ploidy was determined.
The diagnosis of molar pregnancy was based on the pathological
criteria of Szulman and Surti.S DNA ploidy of the cells was
determined by flow cytometry (method of Hedley et al)." Briefly,
30-40 urn formalin-fixed paraffm sections were dewaxed,
rehydrated, and digested with pepsin to produce a suspension of
nuclei which were stained with propidium iodide (50 Ilg/ml) and
analysed on an EPICS C (Coulter Electronics) flow cytometer.
Between
were

Results
Details of the 11 patients are given in the accompanying
table. The age given is that at time of evacuation of "molar
pregnancy". Length of gestation is taken from the first day
of the last menstrual period. Standard criteria were used to
determine risk categories.14 All patients were treated
successfully, none have relapsed.
The diagnosis of PHM was confirmed on review of
pathological features and by presence of triploidy in 5 cases.
ADDRESSES: Cancer Research Campaign Laboratories and
Department of Medical Oncology (Prof K. D. Bagshawe, FRS,
Visiting Prof Sylvia D. Lawler, MD, Joan Dent, SRN, P. Brown, MSc.G.M
Boxer, FIMLT), and Department of Histopathology (F.J Paradinas,
FRCPath), Charing Cross and Westminster Medical School,
London, UK. Correspondence to Prof K. D Bagshawe, CRC
Laboratories, Department of Medical Oncology, Charing Cross Hospital,
Fulham Palace Road, London W6 8RF, UK.

1075

DETAILS OF PATIENTS WITH PHM WHO REQUIRED CHEMOTHERAPY

I

PT= Persistent trophoblast UN = Unclassified, EP= Earty pregnancy, Chono=Chonocarcinoma

Spina blflda, tSubsequent curettage

a subsequent pregnancy aborted so we

be certain that gestational trophoblastic tumour
was associated with PHM pregnancy. Choriocarcinoma
later occurred in 2 patients (D and E) with PHM (figs 1 and
2). On flow cytometry, the specimen of choriocarcinoma
from case D produced a single G1 peak with a small
contiguous hump in the hyperdiploid region (fig I b). This
pattern is typical of autolysed tissues.15 There was no
evidence of triploid nuclei. In case E a very small peak was
found in the triploid region (DNA index == 1-51).
On histological review, the initial diagnosis of PHM was
revised to CHM in 4 cases and confirmed in 3 by the
presence of diploid cells with no triploid population.
In 2 patients (J and K) the molar pregnancy was not
confirmed. When reviewed, the pathological specimen from

patient (A) had

could

not

Fluorescence mtensity

(DNA content)

Fig 1-Case D: DNA histograms of (a) partial hydatidiform mole,

and (b) choriocarcinoma.

thought to show evidence of an early non-molar

pregnancy. Two samples of tissues were received from case
J. The first sample was from material evacuated from the
uterus after diagnosis of incomplete abortion in April, 1982.
The second sample was obtained at a subsequent dilatation
and curettage carried out in October 1982. Both samples
showed persistent trophoblastic tissue without villi. The
case

K was

HCG level 8 months after the initial uterine evacuation was

27050 IU/1. There was no evidence of a triploid cell
population in cases J and K. No maternal tissues were
present, so there was no diploid control peak on flow
cytometry. There are no satisfactory external diploid
standards for formalin-fixed paraffin-embedded material. 1-6
Samples without maternal cells which produced a single G1
peak on flow cytometry were therefore left unclassified.

fluorescence intensity

(DNA content)

Fig 2-Case E: DNA histograms of (a) partial hydatidiform mole,

and (b) choriocarcinoma

1076

Discussion

registering and monitoring molar

pregnancies is to identify those in whom gestational
trophoblastic tumour develops. The tumour types are
invasive mole, which can cause death by local invasion,
choriocarcinoma, and the rare placental-site tumour; the last
two are frankly malignant. Malignant tumour can follow any
type of conceptus, though the risk of choriocarcinoma is
The

purpose

of

1000-times greater after a molar pregnancy than a

normal pregnancy. 17 It is essential to examine the material of
an induced abortion and diagnose either molar or normal
conceptus. This can be difficult because there may be little
material for study. Fetal parts should be sought by the
obstetrician; they are present in PHM and can be crucial to
the diagnosis. Distinction between PHM and CHM is
important since the risk of gestational trophoblastic tumour
is greater after a complete mole. Nevertheless, all patients
with a hydatidiform mole need to be followed up. According
to revised criteria many patients with CHM and most with
PHM have HCG levels monitored for 6 months, rather than
the 2 years previously advised.
The clinical features of patients with CHM and PHM
differ (table). In CHM the gestation of the pregnancy is
shorter (range 9-11 weeks) than in PHM (range 16-33
weeks) fetal development. Fetal parts were observed in all
the confirmed cases of PHM but were not seen in CHM. In
this series the two main types of HM were identifiable, even
without knowledge of tissue ploidy. In the two patients
registered for follow-up of molar pregnancy in whom this
diagnosis was not confirmed were fortunate that the
referring hospital pathologists thought that there were some
suspicious histological features. The patients subsequent
course bore out this concern. If GTT had not been
diagnosed early, and chemotherapy started promptly, the
outcome might have been different. If there is any doubt
about the pathological features of the products of a
pregnancy, the patient should be registered and followedup.
The diagnosis of PHM was confirmed in 5 of 11patients
who had gestational trophoblastic tumour. In 4 the tumour
occurred after the molar pregnancy though PHM might not
have been the causal pregnancy.18 In 1 patient (D) the
malignant tissue was diploid and therefore choriocarcinoma
may not have developed from PHM. Another pregnancy
after the PHM was unlikely because the patient was on oral
contraceptives. The chemotherapy risk category was high
because of the long interval, nearly 2 years, between the last
known pregnancy and the start of treatrnent.14 Flow
cytometry of tissue from the other histologically confirmed
case of choriocarcinoma after PHM (E) showed a small
triploid cell population. The low ratio of triploid to diploid
cells makes interpretation difficult, but the triploid
population might represent malignant cells.
The diagnosis of PHM is further complicated by the
occurrence of twin pregnancies in which one twin is CHM
or PHM and the other is a normal or abnormal fetus. 1920
Genetic techniques or flow cytometry may be needed to
establish the diagnosis.
The genetic origin of the tumours can be determined by
molecular genetic techniques on fresh material and this has
been applied to choriocarcinoma .21 In the future genetic
diagnosis on fixed material might be possible.
Although the evidence of GTT arising from PHM is
circumstantial, 4 patients in this series required
chemotherapy for GTT when the last known pregnancy
over

was

registered and

confirmed

as a

PHM. Between 1973,

when the HM registration scheme began, and 1988, there

increase in the proportion of PHM among
registered pregnancies. 14-3% of all registrations were with
a referring hospital diagnosis of PHM but in 1988 this had
risen to 28 %. A subset of the 317 patients registered in 1983,
for whom histological sections were available, were reviewed
(by K. D. B.) and 16% were classified as PHM. Whereas
PHM was probably underdiagnosed in the early years of the
scheme, in the later years it may have been overdiagnosed. It
is difficult to estimate how many women will get
complications of PHM for which they need chemotherapy
but, on the basis of 4 cases among 857 registrations of PHM,
the risk is about 1 in 200 compared with about 1 in 30 000 for
term births and 1 in 12 for CHM. However approximate
these figures are, they suggest that a patient with a diagnosis
of PHM, even if confirmed by flow cytometry, should be
followed up through the molar registration scheme.
wassubstantial

REFERENCES
1. World Health Organisation: Gestational Trophoblastic Diseases. Report
of a WHO Scientific Group. Technical Report Series 692. Geneva:

WHO 1983.
2.

Bagshawe KD, Dent J and Webb J. Hydatidiform mole in England and

Wales 1973-1983. Lancet 1986; ii: 673-77.

Kajii T, Ohama K. Androgenetic origin of hydatidiform mole. Nature
1977; 268: 633-34.
4. Ohama K, Kajii T, Okamoto E, et al. Dispermic origin of XY
hydatidiform moles. Nature 1981; 29: 551-52.
5. Szulman AE, Surti U. The syndromes of hydatidiform mole. I.
Cytogenetic and morphologic correlations. Am J Obstet Gynecol 1978;
3.

131: 665-71.

6. Lawler SD, Fisher RA, Pickthall VJ, et al. Genetic studies on

hydatidiform moles. I. The ongin of partial moles. Cancer Genet
7.

Cytogenet 1982; 5: 309-20.

Jacobs PA, Szulman AE, Funkhouser J, et al. Human triploidy:
relationship between parental origin of the additional haploid
complement and development of partial hydatidiform mole. Ann Hum

Genet 1982; 46: 223-31.

8. Stone M, Bagshawe KD. Hydatidiform mole: two entities. Lancet 1976; i:
535-36.
9. Szulman AE, Surti U. Strict clinicopathological criteria in the diagnosis
of partial hydatidiform mole. A plea renewed. Am J Obstet Gynecol
1985; 152: 1107-08.
10. Lawler SD, Fisher RA. Genetic studies in hydatidiform mole with
clinical correlations. Placenta 1987; 8: 77-88.
11. Fisher RA, Lawler SD, Ormerod MG, et al. Flow cytometry used to
distinguish between complete and partial hydatidiform moles. Placenta
1987; 8: 249-56.
12. Hemming JD, Quirke P, Womack C, et al. Diagnosis of molar pregnancy
and persistent trophoblastic disease by flow cytometry. J Clin Pathol
1987; 40: 615-20.
13. Hedley DW, Friedlander ML, Taylor LW, et al. Method for analysis of
cellular DNA content of paraffin-embedded pathological material
using flow cytometry. J Histochem Cytochem 1983; 31: 1333-35.
14. Bagshawe KD. Risk and prognostic factors in trophoblastic neoplasia.
Cancer 1976; 38: 1373-85.
15. Alanen KA, Joensuu H, Klemi PJ. Autolysis is a potential source of false
aneuploid peaks in flow cytometric DNA histograms. Cytometry 1989;
10: 417-25.
16. Hedley DW. Flow cytometry using paraffin-embedded tissue: five years
on. Cytometry 1989; 10: 229-41.
17. Bagshawe KD, Lawler SD. Choriocarcinoma. In: Schottenfeld D,
Fraumeni JF, eds. Cancer Epidemiology and Prevention, pp 909-924,
Philadelphia: Saunders, 1982.
18. Lawler SD, Klouda PT, Bagshawe, KD. The relationship between HLA
antibodies and the causal pregnancy in choriocarcinoma Br J Obstet
Gynaecol 1976; 83: 651-55.
19. Beischer NN, Fortuno DW. Significance of chromatin patterns in cases
of hydatidiform mole with an associated fetus. Am J Obstet Gynec 1968;
100: 276-82.
20. Fisher RA, Sheppard DM, Lawler SD. Twin pregnancy with complete
hydatidiform mole (46xx) and fetus (46, xy); genetic origin proved by
analysis of chromosome polymorphisms. Br Med J 1982; 284: 1218-20.
21. Fisher RA, Lawler SD, Povey S, et al Genetically homozygous
choriocarcinoma following pregnancy with hydatidiform mole. BrJ
Cancer 1988; 58: 788-92.