- Email: [email protected]
The pathology of gestational trophoblastic disease: recent advances
Pathology (February 2007) 39(1), pp. 88–96
MOLECULAR STUDIES
The pathology of gestational trophoblastic disease: recent advances MICHAEL WELLS
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
Academic Unit of Pathology, University of Sheffield Medical School, Sheffield, United Kingdom
Summary When inundated with numerous specimens of products of conception as the consequence of miscarriage, it is all too easy for histopathologists to forget that the biology of trophoblast and the events of early placental implantation continue to fascinate because of the inherently invasive properties of the non-villous (extravillous) trophoblast. However, unlike the invasion of a malignant tumour, the invasion of trophoblast is controlled. The failure of adequate conversion of maternal uteroplacental arteries is a major pathogenetic phenomenon of important disorders of pregnancy including pre-eclampsia. However, it is in the field of gestational trophoblastic disease that diagnostic acumen is most called for. There are several problematic areas that give rise to diagnostic error; e.g., the diagnosis of early complete mole as partial mole, the over-diagnosis of hydatidiform mole in tubal pregnancy and the diagnosis of placental site non-villous trophoblast as placental site trophoblastic tumour or choriocarcinoma, particularly if associated with atypia, as frequently observed in complete mole. The chorionic villi of early diploid complete mole show characteristic features of villous profile, stromal mucin and stromal nuclear debris. The distinction between complete mole and triploid partial mole can be facilitated by ploidy analysis and immunohistochemistry for the product of the paternally imprinted, maternally expressed gene, p57kip2. Persistent trophoblastic disease (PTD) is a clinical not a histopathological diagnosis and the role of the histopathologist once a diagnosis of PTD has been made is limited. Invasive mole and choriocarcinoma are encompassed by PTD. Tumours of the non-villous trophoblast are placental site trophoblastic tumour and the more recently recognised epithelioid trophoblastic tumour. The role of immunohistochemistry in the elucidation of trophoblastic lesions is discussed pragmatically. Key words: Trophoblast, hydatidiform mole, complete mole, partial mole, ploidy, p57kip2, extravillous trophoblast, placental site trophoblastic tumour, epithelioid trophoblastic tumour. Received 15 November, accepted 17 November 2006
INTRODUCTION Since January 1997, I have been privileged to be responsible for the histopathological reporting of gestational trophoblastic disease for the Trophoblast Disease Centre in Sheffield, England. For 17 years up to that time, I had a major interest in the biology of trophoblast, in particular, the extravillous or non-villous trophoblast of the placental
bed (see ‘Michael Wells: Personal Contribution’ following References) and, in the course of reporting gynaecological pathology in a major teaching hospital, had been exposed to a reasonable amount of gestational trophoblastic disease. However, now as one of only two histopathologists in England to work in a Trophoblastic Centre (my counterpart at Charing Cross Hospital being Dr Neil Sebire) I see fascinating examples of gestational trophoblastic disease on a regular basis, many of which are received as consultation cases because of the increased experience that I have accrued over the last 10 years. However, if one regards trophoblast and its diseases in purely diagnostic and morphological terms, without an understanding of its fundamental biological function, then one is intellectually the poorer. The physiological invasion of maternal tissues by trophoblast in the first trimester of pregnancy remains one of the most fascinating phenomena in the whole of biology, our understanding of which remains incomplete. An excellent review of the pathology of molar disease has recently been published by Paradinas and Elston.1 In fact, the most up to date reference that I inserted at the proof stage of the 2003 edition of Haines and Taylor was the paper by Genest et al. on p57kip2.2 Therefore, this article will focus on recent developments over the last 5 years or so. At the recent centenary anniversary congress of the International Academy of Pathology held in Montreal, September 2006, I participated in a symposium on Gestational Trophoblastic Diseases. The other participants were Dr Le-Ming Shih (Johns Hopkins Medical Institutions, USA), Dr P. K. Lala (University of Western Ontario, Canada), Dr Lars-Christian Horn (Institute of Pathology, University of Leipzig, Germany), Dr Tsui-Lien Mao (National Taiwan University Hospital) and Dr Annie N. Y. Cheung (Queen Mary Hospital, The University of Hong Kong). I benefited greatly from the new knowledge imparted by my colleagues during that symposium and intend to share some of those developments with you in this review. Trophoblastic lesions continue to cause diagnostic difficulty for general histopathologists and gynaecological pathologists alike. The most common diagnostic errors are: 1. The erroneous diagnosis of early complete hydatidiform mole as partial mole. 2. The over-diagnosis of hydatidiform mole in tubal pregnancy, because of the sheer exuberance of normal early first trimester trophoblastic proliferation. 3. The erroneous diagnosis of exuberant placental site non-villous trophoblast as placental site trophoblastic tumour.
ISSN 0031-3025 printed/ISSN 1465-3931 # 2007 Royal College of Pathologists of Australasia DOI: 10.1080/00313020601137367
PATHOLOGY OF GESTATIONAL TROPHOBLASTIC DISEASE
4.
89
The erroneous diagnosis of atypical non-villous trophoblast, often seen in the context of complete hydatidiform mole, as choriocarcinoma or placental site trophoblastic tumour. In fact, such atypia is of no clinicopathological significance.
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
THE SUBPOPULATIONS OF TROPHOBLAST There are two major populations of trophoblast; the villous and the non-villous. The former is composed of the cytotrophoblast and syncytiotrophoblast. The non-villous trophoblast is sometimes referred to synonymously as the extravillous or intermediate trophoblast. However, it should be remembered that, in biological terms, the developing cytotrophoblastic columns of first trimester chorionic vill are composed of non-villous or intermediate trophoblast. The immunological conundrum that leads to successful pregnancy in the presence of the fetally derived placenta carrying both paternal and maternal antigens has occupied reproductive biologists and immunologists for more than 30 years. The other vital biological question is why does the invasion of trophoblast, under physiological circumstances, stop at the inner one-third of the myometrium in the second trimester (at around 16 weeks of gestation)? Extravillous trophoblast arises by proliferation and differentiation of the cytotrophoblast of the anchoring chorionic villi. These cells migrate as cell columns in the first few weeks of pregnancy to invade the decidua and uteroplacental (spiral) arteries. There are several subpopulations of extravillous trophoblast: interstitial trophoblast, endovascular trophoblast, placental bed giant cells (which are non-invasive and arise by fusion of interstitial trophoblast cells), the persistent cells of the cytotrophoblastic shell and the extravillous trophoblast of the amniochorion (chorion leve). The antigen expression of the various trophoblast populations differs as demonstrated in Table 1 which is meant to be illustrative but by no means exhaustive.
THE BIOLOGY OF TROPHOBLAST INVASION Thus, invasion of maternal tissues by non-villous trophoblast is a physiological phenomenon of normal pregnancy. It was the pioneering work of Pijnenborg et al. in the early 1980s that established the role of the invasive non-villous trophoblast in establishing normal human placental implantation.1 A failure of migratory non-villous trophoblast in the early weeks of pregnancy is implicated in a wide
TABLE 1 Antigen expression by trophoblast populations
HLA-G14 p57kip2 2,20–23 p6312
Villous syncytiotrophoblast
Villous cytotrophoblast
Placental site non-villous (intermediate) trophoblast
Negative Negative Negative
Negative Positive Positive
Positive Positive Negative
Fig. 1 Endovascular trophoblast associated with fibrinoid necrosis of a maternal uteroplacental (spiral) artery. This is placental site trophoblastic tumour but the pattern of trophoblast invasion seen in this field is histologically indistinguishable from normal early placentation.
range of disorders of pregnancy, the most important of which is, of course, pre-eclampsia. It is beyond the scope of this review to discuss such conditions in further detail but they have recently been extensively reviewed.3 The histological appearances of the placental site trophoblastic tumour merely mirror those of normally invading non-villous trophoblast: myometrial infiltration, intravascular tumour and the induction of fibrinoid necrosis in the walls of the uteroplacental arteries (Fig. 1). The endovascular trophoblast acquires an endothelial phenotype (expressing VCAM-1 and ephrin B receptor EPHB4) and replaces the endothelium of these arteries.4 Xu et al. have shown that the regulation of trophoblast invasion into maternal tissues is controlled by a complex series of interactions between the extravillous trophoblast and molecules that are predominantly of decidual origin. Transforming growth factor (TGF)-b and the TGF-bbinding proteoglycan decorin that localises TGF-b in the decidual extracellular matrix inhibit extravillous trophoblast cell growth, migration and invasiveness.5 Neoplastic trophoblast is resistant to the negative regulation of TGF-b and this, in turn, may be due to disruption of the Smad post-receptor signalling genes and down-regulation of TIMP genes.6–8 There are numerous other factors which have the capacity to promote trophoblast invasion. Derangements in the production or function of some of these factors such as the urokinase type plasminogen activator (uPA)/uPA receptor system or insulin-like growth factor binding protein (IGFBP)-1 may be implicated in the impaired trophoblast invasion associated with pre-eclampsia.3,9–11 p63 p63 is a nuclear transcription factor belonging to the p53 family. The two major isoforms regulate different downstream targets resulting in distinct phenotypes. Cytotrophoblast expresses the N isoform of p63, whereas chorionic-type ‘intermediate trophoblast’ in the fetal membranes expresses the TA isoform. Implantation site intermediate trophoblast and syncytiotrophoblast do not
90
WELLS
Pathology (2007), 39(1), February
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
express either of the p63 isoforms. Expression of different p63 isoforms may be important in the control of trophoblastic differentiation and placental development. For example as cytotrophoblast differentiates into either syncytiotrophoblast or implantation site extravillous trophoblast in the trophoblast columns, there is a dramatic decrease in expression of the N isoform which may contribute to cell cycle arrest and loss of proliferative capacity.12 HLA-G In the early 1980s, one of the possible explanations for the survival of the ‘fetal allograft’ was the lack of histocompatibility antigen expression by the placental villous trophoblast. Then, in 1984, it was shown that non-villous trophoblast, the population of trophoblast in most intimate contact with maternal cells, does indeed express an unusual form of Class 1 histocompatibility antigen.13 This is now known as HLA-G. HLA-G immunoreactivity is present in all types of non-villous (intermediate trophoblast), but is not detected in villous cytotrophoblast or syncytiotrophoblast.14 All trophoblastic tumours and tumour-like lesions express HLA-G strongly and diffusely but the vast majority of non-trophoblastic tumours do not express HLA-G, although melanoma, renal cell, breast, ovarian and large cell carcinoma of lung may show focal expression.
HYDATIDIFORM MOLE Hydatidiform mole is a disorder of pregnancy characterised by hydropic change, cistern formation, excessive trophoblast, abnormal distribution of trophoblast and the presence of trophoblastic inclusions. It is classified into complete or partial types based on histopathological and genetic criteria: Genetically, a complete mole is diploid 46XX and the sex chromosome complement is paternal in origin. It may arise by a haploid sperm fertilising an empty ovum which then undergoes cytogenesis. Alternatively, a diploid sperm fertilises an empty ovum. All chorionic villi are affected and there is often associated pleomorphism of trophoblast including the non-villous trophoblast of the placental bed. There is an absence of fibrosis. In contrast the partial mole is triploid (69,XXY 69,XXX or 69,XYY) and arises by dispermic fertilisation of a haploid ovum. Fetal parts may be present and there are two populations of villi. Enlarged villi (>3–4 mm) with central cavitation are a prominent feature. The villi have an irregular (angulated) profile with trophoblastic inclusions. The apparent excess of trophoblast may be subtle. There may be abnormal (angioectatic) vasculature in the second trimester. Pathological mimics of partial mole include: Beckwith–Wiedemann syndrome, placental angiomatous malformation, twin gestation with complete mole and existing fetus (see below), early complete mole (see below) and hydropic spontaneous miscarriage.15 Discordance in histopathological diagnosis is frequently seen in partial mole versus hydropic miscarriage and results from difficulty in evaluating trophoblastic hyperplasia. Ploidy analysis can improve concordance16 (see below).
Fig. 2 Early complete hydatidiform mole showing the very characteristic bulbous or polypoid profile of a molar chorionic villus accompanied by a myxoid stroma.
Early complete mole In the 1960s the mean age at evacuation of hydatidiform mole was 17 weeks. Nowadays, with the introduction of ultrasound examination as part of the routine clinical management of early pregnancy complications, it is 9.4 weeks, before the classical clinical or pathological features have developed. Thus, the histological features of complete hydatidiform mole are more subtle.17 A common error made by histopathologists in molar disease is the erroneous diagnosis of early complete hydatidiform mole as partial mole. Non-triploid partial moles probably do not exist.18 Certain typical features are seen only in early complete mole: abnormally shaped villi with a bulbous or polypoid profile, stromal mucin and stromal nuclear debris (Fig. 2, 3). We have recently shown (unpublished data) that the stromal debris which is a feature of early complete mole is the result of increased stromal proliferation and apoptosis (Table 2), a finding also recently demonstrated by Kim et al.19 Using an antibody to CD31, Kim et al. found that the number of mature blood vessels with distinct lumens in the
Fig. 3 High power view of early complete mole to show the villous stromal apoptosis.
PATHOLOGY OF GESTATIONAL TROPHOBLASTIC DISEASE
TABLE 2 Ki-67 indices of villous stroma* Diagnosis
Mean
n
SD
Early complete mole Partial mole Hydropic miscarriage
22.31 8.33 5.19
30 22 18
5.75 5.66 4.28
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
*Unpublished data. p,0.05. SD, standard deviation.
villous stroma was significantly reduced in both early complete mole and early partial mole compared with normal pregnancy. However, the number of CD31 positive primitive stromal cells or immature vascular networks without lumens did not differ significantly between the three groups. They attributed the retarded vasculogenic differentiation in early complete mole to increased apoptosis in the precursor components of blood vessels. In our study there was a statistically significant strong positive correlation between Ki-67 proliferation index and p53 expression (rho50.840; p,0.05). Ki-67 proliferation index and p53 expression of villous stromal cells were strongly correlated with the amount of stromal karyorrhetic debris (rho50.791; p,0.05). Ploidy analysis and p57kip2 Ploidy analysis is of great value in making the distinction between a diploid complete mole and a triploid partial mole but cannot be used to make a diagnosis of hydatidiform mole per se. It may also aid the distinction between a diploid hydropic miscarriage and a triploid partial mole. This may be achieved by flow cytometry or digital image analysis. In my own laboratory we have moved exclusively to the use of digital image analysis of microdissected tissue from formalin fixed, paraffin embedded blocks which, unlike flow cytometry, allows precise identification of the cells undergoing analysis (Fig. 4, 5). The greatest advance in the histopathology of molar pregnancy in recent years has been the introduction of immunohistochemistry for p57kip2.2,20–23 p57kip2 is a
Fig. 4 Digital image ploidy analysis of diploid complete hydatidiform mole. Note that the cells analysed are present in the gallery behind the histogram.
91
paternally imprinted gene located on chromosome 11p15.5, which is maternally expressed. The villous cytotrophoblast is therefore p57kip2 negative in androgenetic complete mole, whereas the villous cytotrophoblast is p57kip2 positive in normal placenta, hydropic miscarriage and partial mole. The syncytiotrophoblast is always p57kip2 negative. An interesting finding is that the non-villous (extravillous or intermediate) trophoblast is p57kip2 positive even in complete mole.23 As yet, there is no satisfactory biological explanation for this interesting phenomenon. Combined digital image ploidy analysis and p57kip2 can now be used in a complementary way to facilitate the diagnosis of complete and partial mole21 (Table 3). Twin molar pregnancy Twin molar pregnancies are dizygotic gestations, usually with a normal fetus and complete mole. Obstetric management may be complicated and fetal and maternal morbidity are increased. During the period 1988–2004, 7200 cases of gestational trophoblastic disease were registered with the Sheffield Trophoblast Centre. Among these patients, 30 cases of clinically or histopathologically suspected twin molar gestations were recorded (0.4% of all molar pregnancies). Nine cases were erroneously diagnosed on clinical grounds. Of those in which preliminary histological examination was suggestive of mole plus twin, tissue was available for expert review in 14 of 19 cases. Only seven (50%) of these were confirmed as complete mole/normal twin. During the same period, two cases provisionally diagnosed as partial mole were found on expert review to be a complete mole plus normal twin. Of the other seven, three were complete mole only, two were partial mole only and two were non-molar; one a normal fetus and the other a normal fetus with placental mosaicism.24 Thus, cases of twin molar gestations comprising a complete mole with coexisting fetus are often misinterpreted as partial molar singleton pregnancies because the presence of normal and molar villi is misinterpreted as indicating partial mole (Fig. 6). Immunohistochemistry for p57kip2 can facilitate the identification of two populations of villous cytotrophoblast in complete mole/twin pregnancy (Fig. 7).
Fig. 5
Digital image ploidy analysis of triploid partial hydatidiform mole.
92
WELLS
Pathology (2007), 39(1), February
TABLE 3 Complementary use of ploidy analysis and p57kip2 immunohistochemistry to facilitate the diagnosis of hydatidiform mole Suspected diagnosis Partial mole Complete mole Partial mole Partial mole
Flow cytometry
Image cytometry
p57kip2 status
Revised diagnosis
Diploid* Triploid Diploid Diploid
Triploid* Triploid Diploid Diploid
+ + 2 +
Partial mole Partial mole Complete mole Hydropic miscarriage
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
*Note greater reliability of digital image analysis compared with flow cytometry. +, positive; 2, negative.
Ectopic molar pregnancy Molar disease occurring in the context of ectopic pregnancy is rare and most reported cases are tubal pregnancies. Tubal molar pregnancy is over-diagnosed because histopathologists do not appreciate the exuberance of normal trophoblast in the early first trimester.25,26 Tertiary histopathological review is essential. However, normal early placentation, be it tubal or intrauterine, does not show trophoblast proliferation around the entire circumference of the villus, stromal apoptosis or scalloping of chorionic villi. Persistent trophoblastic disease (PTD) The histopathologist must appreciate that his or her role is limited once the diagnosis of hydatidiform mole has been made. The management of gestational trophoblastic disease is based on serial b-hCG estimations. If the b-hCG levels continue to rise or fail to fall, then the patient by definition has persistent trophoblastic disease (PTD). Thus, PTD is not a histopathological diagnosis and, indeed, the basic underlying cause may never be known if the patient responds to appropriate chemotherapy, for another biopsy specimen may never be taken. Usually PTD is due to invasive mole or choriocarcinoma. PTD occurs in 15–20% of patients with complete mole, and is rare following partial mole (remember that many early complete moles are misdiagnosed as partial moles). The majority of cases are due to invasive mole; however, this definitive diagnosis will only be made nowadays in those rare cases that are recalcitrant to chemotherapy for which hysterectomy is performed. It is entirely possible for a general histopathologist working in a district general hospital to go through an entire working lifetime without ever seeing such a case.
Fig. 6 Twin pregnancy with complete hydatidiform mole. In a single high power field normal chorionic villi and a molar villus are seen. Such an appearance may lead to an erroneous diagnosis of partial mole.
Prediction of persistent gestational trophoblastic disease Partial mole still carries a risk of developing gestational trophoblastic neoplasia, though the risk is much less than for complete moles, so follow-up is considered necessary for both complete and partial moles. Can the histopathologist predict those cases of hydatidifom mole that will go on to persistent trophoblastic disease? In asking this question I am not referring to the prognostic factors used by clinicians in their management of these patients to produce a prognostic score to separate patients into low risk and high risk groups for different treatments. Dr Cheung has addressed this question in a series of papers from Hong Kong. Hydatidiform moles which subsequently develop persistent trophoblastic disease, especially those which metastasise, are more likely to show telomerase activity.27 Apoptotic indices of those hydatidiform moles that spontaneously regress are statistically higher than those that develop persistent gestational trophoblastic neoplasia, using both the nick end labelling (TUNEL) technique28 and a M30 CytoDeath antibody.29 Using a human apoptosis array, increased expression of Mcl-1, an anti-apoptotic gene, has been detected in hydatidiform moles that subsequently develop into gestational trophoblastic neoplasia.30 This has been confirmed by quantitative PCR and protein expression analysis. cDNA microarrays and quantitative RNA analysis have shown down-regulation of ferritin light polypeptide (FTL) and IGFBP-1 in hydatidiform mole that subsequently developed gestational trophoblastic neoplasia when compared with those cases that regressed. Immunohistochemical analysis confirmed reduced IGFBP-1 protein expression in hydatidiform mole that developed gestational trophoblastic neoplasia.31
Fig. 7 p57kip2 immunohistochemistry in twin pregnancy with complete mole. The villous cytotrophoblast of the normal placenta is p57kip2 positive, whilst the complete mole is negative.
PATHOLOGY OF GESTATIONAL TROPHOBLASTIC DISEASE
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
Fig. 8
Hysterectomy specimen showing invasive complete mole.
The hysterectomy specimen performed for persistent gestational trophoblastic disease In rare cases, when the patient does not respond to chemotherapy and has persistently elevated b-hCG levels, there is little alternative to performing a hysterectomy. In such circumstances, it is likely that the hysterectomy specimen will show features of invasive mole, with venous invasion by molar chorionic villi (Fig. 8). There may also be extensive associated necrosis which is a reflection of the (albeit inadequate) cytopathic effect of the chemotherapy.
CHORIOCARCINOMA We are concerned here with choriocarcinoma as a gestational trophoblastic neoplasm; choriocarcinoma may also, of course, form part of the spectrum of nongestational malignant germ cell neoplasia. Under such circumstances, the tumour is often admixed with other malignant germ cell elements. In terms of the basic histopathology of choriocarcinoma, there have been few developments in recent years. The diagnosis rests on the recognition of the characteristic bimorphic appearance of malignant cyto- and syncytiotrophoblast, which is usually unmistakeable. However, insufficient attention has been paid in the literature to apparently ‘hybrid’ variants which show features that render them difficult to label either as choriocarcinoma on the one hand or placental site trophoblastic tumour on the other. Intraplacental choriocarcinoma I have recently encountered a 27-year-old female who presented with sudden onset of right-sided hemiparesis, headache and vomiting, 12 weeks after the spontaneous miscarriage of a 24 week (severely macerated but otherwise normal) fetus. She died 20 days after the onset of symptoms. Post-mortem examination showed disseminated trophoblastic disease but with no evidence of choriocarcinoma in the uterus. Review of the placental tissue showed a small focus of intraplacental choriocarcinoma comprising a biphasic proliferation of markedly atypical syncytio and cytotrophoblast with a high proliferation index (Fig. 9, 10). One clue to the diagnosis in this case was the presence of persistently elevated
93
Fig. 9 Intraplacental choriocarcinoma. A small focus of highly atypical trophoblast is arising from apparently normal chorionic villi.
serum a-fetoprotein levels, which are known to be a complication of choriocarcinoma. It is possible that the elevated serum a-fetoprotein is caused by increased amounts of fetal proteins crossing to the maternal circulation facilitated by mechanical breakdown of the materno-fetal barrier. Intraplacental choriocarcinoma is one of the most enigmatic conditions in the spectrum of gestational trophoblastic disease. It is rare and the placental primary is often small. It arises from the cytotrophoblast of apparently normal chorionic villi. Placental removal is sometimes curative but the lesion is more likely to be associated with maternal metastatic disease, as illustrated above. It is a difficult macroscopic diagnosis because the lesions may be small and resemble small infarcts. Unlike conventional choriocarcinoma, the presence of villi is not incompatible with the diagnosis. The tumour may present in pregnancy but is usually diagnosed several weeks or months after pregnancy (median 5 months). The lungs are the most common site of metastasis and, additionally, the possibility of choriocarcinoma should be borne in mind with any intracerebral bleed in a woman of child-bearing age.32
Fig. 10 High power view of intraplacental choriocarcinoma. Note the severe degree of nuclear atypia and the biphasic pattern of malignant cyto and syncytiotrophoblast.
94
WELLS
Pathology (2007), 39(1), February
LESIONS OF NON-VILLOUS TROPHOBLAST There are two benign lesions that are sometimes erroneously diagnosed as neoplastic: the (exaggerated) placental site reaction and the placental site nodule. In my opinion, too much emphasis has been placed in the descriptive morphological literature on the potential for diagnostic mishap associated with these lesions. The most important thing to remember when confronted with such cases is to gain access to the necessary clinical information which includes knowledge of the b-hCG status of the patient.
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
(Exaggerated) placental site reaction It is probably inappropriate to regard the so-called exaggerated placental site reaction as a lesion since, in my experience the normal implantation site often shows exuberant extravillous trophoblast that can seem alarming to the inexperienced. Placental site nodule This is usually a reasonably well circumscribed lesion of the endometrium composed of rather degenerate-looking extravillous trophoblast set in a hyaline-type matrix. It marks the residuum of involuted pregnancy and may be present many months after the pregnancy, of which there may be no clinical (or indeed patient) awareness. It is usually diagnosed as an incidental finding and is of no clinical significance except that it may perturb the histopathologist and lead him or her to an erroneous diagnosis of placental site trophoblastic tumour. However, the relatively discrete nature of the lesion, the degenerate appearance of the trophoblast and the lack of true cytological atypia in the form of nuclear pleomorphism should detract from this diagnosis. Kurman believes that the placental site nodule originates from the trophoblast of the chorion leve with a similar immunohistochemical profile to that of epithelioid trophoblastic tumour.33,34 Placental site trophoblastic tumour This is a neoplasm of the implantation site non-villous trophoblast and, in contrast to choriocarcinoma, occurs mostly after regular pregnancies or spontaneous miscarriages. It is generally not associated with the presence of chorionic villi. Invasion of the myometrium with a dissecting growth pattern is a prominent feature and the typical vascular changes have been referred to above. Despite the weak immunoreactivity for b-hCG of individual placental site trophoblastic tumour cells, serum levels are usually moderately elevated due to the bulk of tumour. Severe nuclear pleomorphism and high mitotic activity are associated with malignant behaviour.1 Epithelioid trophoblastic tumour This relatively recently described tumour of trophoblast resembles the trophoblastic cells of the chorion leve. It can be associated with any gestational event in younger women, but can also be diagnosed more than 10 years after the last known pregnancy or in post-menopausal women. It is composed of a relatively uniform population of mononucleate trophoblastic cells with eosinophilic cytoplasm surrounded by a well-defined cell membrane, intimately associated with a fibrillar hyaline-like matrix34 (Fig. 11).
Fig. 11 Epithelioid trophoblastic tumour. Note the uniform mononucleate cells set in a hyaline-like matrix. This case was initially erroneously diagnosed as epithelioid leiomyosarcoma.
PRAGMATIC IMMUNOHISTOCHEMICAL APPROACH TO THE DIAGNOSIS OF LESIONS OF THE NON-VILLOUS TROPHOBLAST IN BIOPSY SPECIMENS The first important task is to confirm the trophoblastic nature of the cells of interest. Though not specific for trophoblast, cytokeratin (e.g., CAM5.2; AE1/AE3; cytokeratin 18) immunohistochemistry will readily discriminate the intensely immunoreactive trophoblast of the placental site, whether neoplastic or non-neoplastic, from the cytokeratin negative maternal decidua (Fig. 12). Such a simple initial approach will also effectively exclude malignant melanoma. However, cytokeratin immunohistochemistry alone will not exclude other types of epithelial cells such as squamous carcinoma, which may be particularly important in the differential diagnosis of epithelioid trophoblastic tumour, though cytokeratin 18 is not expressed in squamous carcinoma of the cervix. Immunohistochemistry for human placental lactogen, inhibin-a, HLA-G and Mel-CAM (CD146) is helpful in this regard, all of which are positive in non-villous trophoblast. Squamous carcinoma cells are likely to be
Fig. 12 Intense immunoreactivity of non-villous trophoblast for low molecular weight cytokeratin (this is the epithelioid trophoblastic tumour seen in Fig. 11).
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
PATHOLOGY OF GESTATIONAL TROPHOBLASTIC DISEASE
p16 positive as a consequence of human papillomavirus infection, whereas trophoblast is negative for p16.35 The possibility of a smooth muscle lesion such as epithelioid leiomyosarcoma should always be borne in mind and excluded by demonstrating positivity with the trophoblast markers already cited and negativity with appropriate smooth muscle markers. However, the most difficult histological distinction in small biopsy specimens is between non-neoplastic and neoplastic trophoblast and, it must be conceded, that the immunohistochemical armamentarium is of little help in this regard, the exception to this being Ki-67; the Ki-67 labelling index is likely to be significantly elevated (.10%) in a neoplastic lesion of non-villous trophoblast.36,37 Recently, it has been shown that cyclin E is expressed in the trophoblastic columns of anchoring villi, implantation site non-villous trophoblastic cells and choriocarcinoma.35,37 Epithelioid trophoblastic tumours demonstrate a much higher cyclin E staining score than placental site nodules permitting distinction between the two.37,38 To be candid, I have never performed Ki-67 or cyclin immunohistochemistry in this situation and there really is no substitute for careful clinicopathological evaluation. A dogmatic diagnosis of placental site trophoblastic tumour will be followed by embarrassment if the clinicians are already aware that serum levels of b-hCG or human placental lactogen are not significantly elevated and that imaging shows no evidence of tumour. In practical clinical terms, the distinction between placental site trophoblastic tumour and epithelioid trophoblastic tumour is not really an important one.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
OUTSTANDING QUESTIONS Whilst there has been a dramatic increase in our knowledge of the biology of trophoblast and the various pathological lesions derived from it, certain outstanding questions remain. To my mind the most important of these are as follows: 1. The cell of origin of choriocarcinoma. 2. The key factor(s) in determining the cessation of trophoblastic invasion at the inner one-third of the myometrium. 3. The more accurate prediction of those cases of hydatidiform mole that will go on to persistent trophoblastic disease.
16.
17.
18.
19.
20. Address for correspondence: Professor M. Wells, Academic Unit of Pathology, University of Sheffield Medical School, Beech Hill Road, Sheffield S10 2RX, United Kingdom. Email: [email protected]
References 1. Paradinas FJ, Elston CW. Gestational trophoblastic diseases. In: Fox H, Wells M. Haines and Taylor Obstetrical and Gynaecological Pathology. 5th ed. London: Churchill Livingstone, 2003; 1359–430. 2. Genest DR, Dorfman DM, Castrillon DH. Ploidy and imprinting in hydatidiform moles. Complementary use of flow cytometry and immunohistochemistry of the imprinted gene product p57kip2 to assist molar classification. J Reprod Med 2002; 47: 342–6. 3. Lala PK, Chakraborty C. Factors regulating trophoblast migration and invasiveness: possible derangements contributing to preeclampsia and fetal injury. Placenta 2003; 24: 575–87. 4. Red-Horse K, Kapidzic M, Zhou Y, Feng KT, Singh H, Fisher SJ. EPHB4 regulates chemokine-evoked trophoblast responses: a
21.
22.
23.
24.
25.
95
mechanism for incorporating the human placenta into the maternal circulation. Development 2005; 132: 4097–106. Xu G, Guimond M-J, Chakraborty G, Lala PK. Control of proliferation, migration and invasiveness of human extravillous trophoblast by decorin, a decidual product. Biol Reprod 2002; 67: 681–9. Xu G, Chakraborty C, Lala PK. Expression of TGF-beta signalling genes in the normal, premalignant and malignant human trophoblast: loss of Smad3 in choriocarcinoma cells. Biochem Biophys Res Commun 2001; 287: 47–55. Xu G, Chakraborty C, Lala PK. Restoration of TGF-beta regulation of plasminogen activator inhibitor-1 in Smad3-restituted human choriocarcinoma cells. Biochem Biophys Res Commun 2002; 294: 1079–86. Xu G, Chakraborty C, Lala PK. Reconstitution of Smad3 restores TGF-b response of tissue inhibitor of metalloprotease-1 upregulation in human choriocarcinoma cells. Biochem Biophys Res Commun 2003; 300: 383–90. Mckinnon T, Chakraborty C, Gleeson L, Chidiac D, Lala PK. Stimulation of human extravillous trophoblast (EVT) migration by IGF-II is mediated by IGF type 2 receptor, involving inhibiting G proteins and phosphorylation of MAPK. J Clin Endocrinol Metab 2001; 86: 3665–74. Gleeson L, Chakraborty C, Mckinnon T, Lala PK. Insulin-like growth factor binding protein-1 stimulates human trophoblast migration by signalling through a5b1 integrin via mitogen-activated protein kinase pathway. J Clin Endocrinol Metab 2001; 86: 2484–93. Liu J, Chakraborty C, Barbin YP, Dixon SJ, Lala PK. Non catalytic domain of uPA stimulates human extravillous trophoblast migration by utilising phospholipase C, P-1-3 kinase and MAP kinase. Exp Cell Res 2003; 286: 138–51. Shih IM, Kurman RJ. p63 expression is useful in the distinction of epithelioid and placental site trophoblastic tumours by profiling trophoblastic subpopulations. Am J Surg Pathol 2004; 28: 1177–83. Wells M, Hsi B-L, Faulk WP. Class 1 antigens of the major histocompatibility complex on cytotrophoblast of the human placental basal plate. Am J Reprod Immunol 1984; 6: 167–74. Singer G, Kurman RJ, McMaster M, Shih I-M. HLA-G immunoreactivity is specific for intermediate trophoblast in gestational trophoblastic disease and can serve as a useful marker in differential diagnosis. Am J Surg Pathol 2002; 26: 914–20. Genest DR. Partial hydatidiform mole; clinicopathological features, differential diagnosis, ploidy and molecular studies, and gold standards for diagnosis. Int J Gynecol Pathol 2001; 20: 315–22. Fukunaga M, Katabuchi H, Nagasaka T, Mikami Y, Minamiguchi S, Lage JM. Interobserver and intraobserver variability in the diagnosis of hydatidiform mole. Am J Surg Pathol 2005; 29: 942–7. Sebire NJ, Makrydimas G, Agnantis NJ, Zagorianakou N, Rees H, Fisher RA. Updated diagnostic criteria for partial and complete hydatidiform moles in early pregnancy. Anticancer Res 2003; 23: 1723–8. Genest DR, Ruiz RE, Weremowicz S, Berkowitz RS, Goldstein DP, Dorfman DM. Do nontriploid partial hydatidiform moles exist? A histologic and flow cytometric reevaluation of nontriploid specimens. J Reprod Med 2002; 47: 363–8. Kim MJ, Kim KR, Ro JY, Lage JM, Lee HI. Diagnostic and pathogenetic significance of increased stromal apoptosis and incomplete vasculogenesis in complete hydatidiform moles in very early pregnancy periods. Am J Surg Pathol 2006; 30: 362–9. Jun SY, Ro JY, Kim KR. p57kip2 is useful in the classification and differential diagnosis of complete and partial hydatidiform moles. Histopathology 2003; 43: 17–25. Crisp H, Burton JL, Stewart R, Wells M. Refining the diagnosis of hydatidiform mole: image ploidy analysis and p57KIP2 immunohistochemistry. Histopathology 2003; 43: 363–73. Merchant SH, Amin MB, Viswanatha DS, Malhotra RK, Moehlenkamp C, Joste NE. p57KIP2 immunohistochemistry in early molar pregnancies: emphasis on its complementary role in the differential diagnosis of hydropic abortuses. Hum Pathol 2005; 36: 180–6. Sebire NJ, Lindsay I. p57KIP2 immunostaining in the diagnosis of complete versus partial hydatidiform moles. Histopathology 2006; 48: 873–4. Hancock BW, Martin K, Evans CA, Everard JE, Wells M. Twin mole and viable fetus: the case for misdiagnosis. J Reprod Med 2006; 51: 825–8. Burton JL, Lidbury EA, Gillespie AM, et al. Over-diagnosis of hydatidiform mole in early tubal ectopic pregnancy. Histopathology 2001; 38: 409–17.
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
96
WELLS
26. Sebire NJ, Lindsay I, Fisher RA, Savage P, Seckl MJ. Overdiagnosis of complete and partial hydatidiform mole in tubal ectopic pregnancies. Int J Gynecol Pathol 2005; 24: 260–4. 27. Cheung AN, Zhang DK, Liu Y, Ngan HY, Shen DH, Tsao SW. Telomerase activity in gestational trophoblastic disease. J Clin Pathol 1999; 52: 588–92. 28. Wong SY, Ngan HY, Chan CC, Cheung AN. Apoptosis in gestational trophoblastic disease is correlated with clinical outcome and Bcl-2 expression but not Bax expression. Mod Pathol 1999; 12: 1025–33. 29. Chiu PM, Ngan HY, Chan CC, Cheung AN. Apoptosis in gestational trophoblastic disease correlates with clinical outcome: assessment by the caspase-related M30 CytoDeath antibody. Histopathology 2001; 38: 243–9. 30. Fong PY, Xue WC, Ngan HY, et al. Mcl-1 expression in gestational trophoblastic disease correlates with clinical outcome: a differential expression study. Cancer 2005; 103: 268–76. 31. Feng HC, Tsao SW, Ngan HY, Xue WC, Chiu PM, Cheung AN. Differential expression of insulin-like growth factor binding protein 1 and ferritin light polypeptide in gestational trophoblastic neoplasia: combined cDNA suppression subtractive hybridization and microarray study. Cancer 2005; 104: 2409–16. 32. Fernando MS, Wells M, Carroll T, Hancock BW, Wharton SB, Intracranial metastasis from an intraplacental choriocarcinoma. (Submitted for publication.) 33. Shih IM, Seidman JD, Kurman RJ. Placental site nodule and characterization of distinctive types of intermediate trophoblast. Hum Pathol 1999; 30: 687–94. 34. Shih I-M, Kurman RJ. Epithelioid trophoblastic tumour – a neoplasm distinct from choriocarcinoma and placental site trophoblastic tumor simulating carcinoma. Am J Surg Pathol 1998; 22: 1393–403. 35. Mao TL, Seidman JD, Kurman RJ, Shih I-M. Cyclin E and p16 immunoreactivity in epithelioid trophoblastic tumour – an aid in differential diagnosis. Am J Surg Pathol 2006; 30: 1105–10. 36. Shih IM, Kurman RJ. Ki-67 labeling index in the differential diagnosis of exaggerated placental site, placental site trophoblastic tumour and choriocarcinoma: a double immunohistochemical staining technique using Ki-67 and Mel-CAM antibodies. Hum Pathol 1998; 29: 27–33. 37. Bamberger AM, Aupers S, Milde-Langosch K, Loning T. Expression pattern of the cell cycle promoter cyclin E in benign extravillous trophoblast and gestational trophoblastic lesions: correlation with expression of Ki-67. Int J Gynecol Pathol 2003; 22: 156–61.
Michael Wells: personal contribution Original articles Sinha D, Wells M, Faulk WP. Immunological studies of human placentae: complement components in pre–eclamptic chorionic villi. Clin Exp Immunol 1984; 156: 175–84. Wells M, Hsi B-L, Yeh C-JG, Faulk WP. Spiral (utero–placental) arteries of the human placental bed show the presence of amniotic basement membrane antigens. Am J Obstet Gynecol 1984; 150: 937–77. Wells M, Hsi B-L, Faulk WP. Class 1 antigens of the major histocompatibility complex on cytotrophoblast of the human placental basal plate. Am J Reprod Immunol 1984; 6: 167–74. Earl UM, Bulmer JN, Wells M. Immunohistological identification of antigen expression by trophoblast populations in a case of ectopic cornual implantation. Br J Obstet Gynaecol 1985; 92: 843–6. Earl UM, Wells M, Bulmer JN. The expression of major histocompatibility complex antigens by trophoblast in ectopic tubal pregnancy. J Reprod Immunol 1985; 8: 13–24. Bulmer JN, Wells M, Bhabra K, Johnson PM. Immunohistological characterisation of endometrial gland epithelium and extra–villous fetal trophoblast in third trimester human placental bed tissues. Br J Obstet Gynaecol 1986; 93: 823–32.
Pathology (2007), 39(1), February
Earl UM, Wells M, Bulmer JN. Immunohistochemical characterisation of trophoblast membrane antigens and secretory products in ectopic tubal pregnancy. Int J Gynecol Pathol 1986; 5: 132–42. Lalani E-NMA, Bulmer JN, Wells M. Peroxidase labelled lectin binding of human extra–villous trophoblast. Placenta 1987; 8: 15–26. Hemming JD, Quirke P, Womack C, Wells M, Elston CW, Bird CC. The diagnosis of molar pregnancy and persistent trophoblastic disease by flow cytometry. J Clin Pathol 1987; 40: 615–20. Bulmer JN, Wells M, Lunny DP, Yeh C-J, Hsi B-L. An investigation of the expression of amnion antigens by spiral arteries in human uteroplacental tissues. Am J Reprod Immunol 1987; 14: 79–83. Wells M, Bennett J, Bulmer JN, Jackson P, Holgate CS. Complement component deposition in uteroplacental spiral arteries in normal human pregnancy. J Reprod Immunol 1987; 12: 125–35. Bulmer JN, Smith JC, Morrison L, Wells M. Maternal and fetal cellular relationships in the human placental basal plate. Placenta 1988; 9: 237–46. Hemming JD, Quirke P, Womack C, Wells M, Elston CW, Pennington GW. Flow cytometry in persistent trophoblast disease. Placenta 1988; 9: 615–20. Thornton JG, Lewis FA, Linton G, Wells M, Tyrrell S, Lilford RJ. Fetal sexing by chorionic villus biopsy and in–situ DNA hybridization with a Y probe and biotin–streptavidin polyalkaline phosphatase labelling. J Obstet Gynaecol 1989; 10: 1–4. Andrew AC, Bulmer JN, Wells M, Morrison L, Buckley CH. Subinvolution of the uteroplacental arteries in the placental bed. Histopathology 1989; 15: 431–3. Thrower S, Bulmer JN, Wells M. Expression of epidermal growth factor receptor and transferrin receptor by human trophoblast populations. Am J Reprod Immunol 1989; 21: 87–93. Thrower S, Bulmer JN, Griffin NR, Wells M. Further studies of lectin biding by villous and extravillous trophoblast in normal and pathological pregnancy. Int J Gynecol Pathol 1991; 10: 238–51. Andrew A, Bulmer JN, Morrison L, Wells M, Buckley CH. Subinvolution of the uteroplacental arteries: an immunohistochemical study. Int J Gynecol Pathol 1993; 12: 28–33. Burton Jl, Lidbury EA, Gillespie AM, et al. Over-diagnosis of hydatidiform mole in early tubal ectopic pregnancy. Histopathology 2001; 38: 409–417. Crisp H, Burton JL, Smith O, Stewart R, Wells M. Refining the diagnosis of hydatidiform mole: image ploidy analysis and p57kip2 immunohistochemistry. Histopathology 2003; 43: 363–73. Hassadia A, Gillespie A, Tidy J, Everard Rgn J, Wells M, Coleman R, Hancock B. Placental site trophoblastic tumour: Clinical features and management. Gynecol Oncol 2005; 99: 603–7. Hancock BW, Martin K, Evans CA, Everard JE, Wells M. Twin mole and viable fetus: the case for misdiagnosis. J Reprod Med 2006; 51: 825–8. Chapters in books Wells M, Fox H. Immunology and immunopathology of the materno-fetal interface. In: Coulam CB, McIntyre JA, Faulk WP, editors. Immunological Obstetrics. New York: Norton Medical Books, 1992; 166–76. Coulam CB, Wells M. Ectopic pregnancy loss. In: Coulam CB, McIntyre JA, Faulk WP, editors. Immunological Obstetrics. New York: Norton Medical Books, 1992; 464–78. Wells M, Mohamdee O. Pathology of the pregnant uterus. In: Fox H, Wells M, editors. Haines and Taylor: Obstetrical and Gynaecological Pathology. London: Churchill Livingstone, 1995; 1509–37. Fox H, Wells M. Pathology of the first trimester. In: Kurjak A, editor-inchief. A Textbook of Perinatal Medicine. London: Parthenon, 1998; 991–8. Hustin J, Wells M. Pathology of the pregnant uterus. In: Fox H, Wells M, editors. Haines and Taylor: Obstetrical and Gynaecological Pathology. London: Churchill Livingstone, 2003; 1327–57. Review articles Wells M, Bulmer JN. The human placental bed: anatomy immunohistochemistry and pathology. Histopathology 1988; 13: 483–98. Berry CW, Brambati B, Eskes TKAB, et al. The Euro-Team Early Pregnancy (ETEP) protocol for recurrent miscarriage. Hum Reprod 1995; 10: 1516–20.
MOLECULAR STUDIES
The pathology of gestational trophoblastic disease: recent advances MICHAEL WELLS
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
Academic Unit of Pathology, University of Sheffield Medical School, Sheffield, United Kingdom
Summary When inundated with numerous specimens of products of conception as the consequence of miscarriage, it is all too easy for histopathologists to forget that the biology of trophoblast and the events of early placental implantation continue to fascinate because of the inherently invasive properties of the non-villous (extravillous) trophoblast. However, unlike the invasion of a malignant tumour, the invasion of trophoblast is controlled. The failure of adequate conversion of maternal uteroplacental arteries is a major pathogenetic phenomenon of important disorders of pregnancy including pre-eclampsia. However, it is in the field of gestational trophoblastic disease that diagnostic acumen is most called for. There are several problematic areas that give rise to diagnostic error; e.g., the diagnosis of early complete mole as partial mole, the over-diagnosis of hydatidiform mole in tubal pregnancy and the diagnosis of placental site non-villous trophoblast as placental site trophoblastic tumour or choriocarcinoma, particularly if associated with atypia, as frequently observed in complete mole. The chorionic villi of early diploid complete mole show characteristic features of villous profile, stromal mucin and stromal nuclear debris. The distinction between complete mole and triploid partial mole can be facilitated by ploidy analysis and immunohistochemistry for the product of the paternally imprinted, maternally expressed gene, p57kip2. Persistent trophoblastic disease (PTD) is a clinical not a histopathological diagnosis and the role of the histopathologist once a diagnosis of PTD has been made is limited. Invasive mole and choriocarcinoma are encompassed by PTD. Tumours of the non-villous trophoblast are placental site trophoblastic tumour and the more recently recognised epithelioid trophoblastic tumour. The role of immunohistochemistry in the elucidation of trophoblastic lesions is discussed pragmatically. Key words: Trophoblast, hydatidiform mole, complete mole, partial mole, ploidy, p57kip2, extravillous trophoblast, placental site trophoblastic tumour, epithelioid trophoblastic tumour. Received 15 November, accepted 17 November 2006
INTRODUCTION Since January 1997, I have been privileged to be responsible for the histopathological reporting of gestational trophoblastic disease for the Trophoblast Disease Centre in Sheffield, England. For 17 years up to that time, I had a major interest in the biology of trophoblast, in particular, the extravillous or non-villous trophoblast of the placental
bed (see ‘Michael Wells: Personal Contribution’ following References) and, in the course of reporting gynaecological pathology in a major teaching hospital, had been exposed to a reasonable amount of gestational trophoblastic disease. However, now as one of only two histopathologists in England to work in a Trophoblastic Centre (my counterpart at Charing Cross Hospital being Dr Neil Sebire) I see fascinating examples of gestational trophoblastic disease on a regular basis, many of which are received as consultation cases because of the increased experience that I have accrued over the last 10 years. However, if one regards trophoblast and its diseases in purely diagnostic and morphological terms, without an understanding of its fundamental biological function, then one is intellectually the poorer. The physiological invasion of maternal tissues by trophoblast in the first trimester of pregnancy remains one of the most fascinating phenomena in the whole of biology, our understanding of which remains incomplete. An excellent review of the pathology of molar disease has recently been published by Paradinas and Elston.1 In fact, the most up to date reference that I inserted at the proof stage of the 2003 edition of Haines and Taylor was the paper by Genest et al. on p57kip2.2 Therefore, this article will focus on recent developments over the last 5 years or so. At the recent centenary anniversary congress of the International Academy of Pathology held in Montreal, September 2006, I participated in a symposium on Gestational Trophoblastic Diseases. The other participants were Dr Le-Ming Shih (Johns Hopkins Medical Institutions, USA), Dr P. K. Lala (University of Western Ontario, Canada), Dr Lars-Christian Horn (Institute of Pathology, University of Leipzig, Germany), Dr Tsui-Lien Mao (National Taiwan University Hospital) and Dr Annie N. Y. Cheung (Queen Mary Hospital, The University of Hong Kong). I benefited greatly from the new knowledge imparted by my colleagues during that symposium and intend to share some of those developments with you in this review. Trophoblastic lesions continue to cause diagnostic difficulty for general histopathologists and gynaecological pathologists alike. The most common diagnostic errors are: 1. The erroneous diagnosis of early complete hydatidiform mole as partial mole. 2. The over-diagnosis of hydatidiform mole in tubal pregnancy, because of the sheer exuberance of normal early first trimester trophoblastic proliferation. 3. The erroneous diagnosis of exuberant placental site non-villous trophoblast as placental site trophoblastic tumour.
ISSN 0031-3025 printed/ISSN 1465-3931 # 2007 Royal College of Pathologists of Australasia DOI: 10.1080/00313020601137367
PATHOLOGY OF GESTATIONAL TROPHOBLASTIC DISEASE
4.
89
The erroneous diagnosis of atypical non-villous trophoblast, often seen in the context of complete hydatidiform mole, as choriocarcinoma or placental site trophoblastic tumour. In fact, such atypia is of no clinicopathological significance.
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
THE SUBPOPULATIONS OF TROPHOBLAST There are two major populations of trophoblast; the villous and the non-villous. The former is composed of the cytotrophoblast and syncytiotrophoblast. The non-villous trophoblast is sometimes referred to synonymously as the extravillous or intermediate trophoblast. However, it should be remembered that, in biological terms, the developing cytotrophoblastic columns of first trimester chorionic vill are composed of non-villous or intermediate trophoblast. The immunological conundrum that leads to successful pregnancy in the presence of the fetally derived placenta carrying both paternal and maternal antigens has occupied reproductive biologists and immunologists for more than 30 years. The other vital biological question is why does the invasion of trophoblast, under physiological circumstances, stop at the inner one-third of the myometrium in the second trimester (at around 16 weeks of gestation)? Extravillous trophoblast arises by proliferation and differentiation of the cytotrophoblast of the anchoring chorionic villi. These cells migrate as cell columns in the first few weeks of pregnancy to invade the decidua and uteroplacental (spiral) arteries. There are several subpopulations of extravillous trophoblast: interstitial trophoblast, endovascular trophoblast, placental bed giant cells (which are non-invasive and arise by fusion of interstitial trophoblast cells), the persistent cells of the cytotrophoblastic shell and the extravillous trophoblast of the amniochorion (chorion leve). The antigen expression of the various trophoblast populations differs as demonstrated in Table 1 which is meant to be illustrative but by no means exhaustive.
THE BIOLOGY OF TROPHOBLAST INVASION Thus, invasion of maternal tissues by non-villous trophoblast is a physiological phenomenon of normal pregnancy. It was the pioneering work of Pijnenborg et al. in the early 1980s that established the role of the invasive non-villous trophoblast in establishing normal human placental implantation.1 A failure of migratory non-villous trophoblast in the early weeks of pregnancy is implicated in a wide
TABLE 1 Antigen expression by trophoblast populations
HLA-G14 p57kip2 2,20–23 p6312
Villous syncytiotrophoblast
Villous cytotrophoblast
Placental site non-villous (intermediate) trophoblast
Negative Negative Negative
Negative Positive Positive
Positive Positive Negative
Fig. 1 Endovascular trophoblast associated with fibrinoid necrosis of a maternal uteroplacental (spiral) artery. This is placental site trophoblastic tumour but the pattern of trophoblast invasion seen in this field is histologically indistinguishable from normal early placentation.
range of disorders of pregnancy, the most important of which is, of course, pre-eclampsia. It is beyond the scope of this review to discuss such conditions in further detail but they have recently been extensively reviewed.3 The histological appearances of the placental site trophoblastic tumour merely mirror those of normally invading non-villous trophoblast: myometrial infiltration, intravascular tumour and the induction of fibrinoid necrosis in the walls of the uteroplacental arteries (Fig. 1). The endovascular trophoblast acquires an endothelial phenotype (expressing VCAM-1 and ephrin B receptor EPHB4) and replaces the endothelium of these arteries.4 Xu et al. have shown that the regulation of trophoblast invasion into maternal tissues is controlled by a complex series of interactions between the extravillous trophoblast and molecules that are predominantly of decidual origin. Transforming growth factor (TGF)-b and the TGF-bbinding proteoglycan decorin that localises TGF-b in the decidual extracellular matrix inhibit extravillous trophoblast cell growth, migration and invasiveness.5 Neoplastic trophoblast is resistant to the negative regulation of TGF-b and this, in turn, may be due to disruption of the Smad post-receptor signalling genes and down-regulation of TIMP genes.6–8 There are numerous other factors which have the capacity to promote trophoblast invasion. Derangements in the production or function of some of these factors such as the urokinase type plasminogen activator (uPA)/uPA receptor system or insulin-like growth factor binding protein (IGFBP)-1 may be implicated in the impaired trophoblast invasion associated with pre-eclampsia.3,9–11 p63 p63 is a nuclear transcription factor belonging to the p53 family. The two major isoforms regulate different downstream targets resulting in distinct phenotypes. Cytotrophoblast expresses the N isoform of p63, whereas chorionic-type ‘intermediate trophoblast’ in the fetal membranes expresses the TA isoform. Implantation site intermediate trophoblast and syncytiotrophoblast do not
90
WELLS
Pathology (2007), 39(1), February
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
express either of the p63 isoforms. Expression of different p63 isoforms may be important in the control of trophoblastic differentiation and placental development. For example as cytotrophoblast differentiates into either syncytiotrophoblast or implantation site extravillous trophoblast in the trophoblast columns, there is a dramatic decrease in expression of the N isoform which may contribute to cell cycle arrest and loss of proliferative capacity.12 HLA-G In the early 1980s, one of the possible explanations for the survival of the ‘fetal allograft’ was the lack of histocompatibility antigen expression by the placental villous trophoblast. Then, in 1984, it was shown that non-villous trophoblast, the population of trophoblast in most intimate contact with maternal cells, does indeed express an unusual form of Class 1 histocompatibility antigen.13 This is now known as HLA-G. HLA-G immunoreactivity is present in all types of non-villous (intermediate trophoblast), but is not detected in villous cytotrophoblast or syncytiotrophoblast.14 All trophoblastic tumours and tumour-like lesions express HLA-G strongly and diffusely but the vast majority of non-trophoblastic tumours do not express HLA-G, although melanoma, renal cell, breast, ovarian and large cell carcinoma of lung may show focal expression.
HYDATIDIFORM MOLE Hydatidiform mole is a disorder of pregnancy characterised by hydropic change, cistern formation, excessive trophoblast, abnormal distribution of trophoblast and the presence of trophoblastic inclusions. It is classified into complete or partial types based on histopathological and genetic criteria: Genetically, a complete mole is diploid 46XX and the sex chromosome complement is paternal in origin. It may arise by a haploid sperm fertilising an empty ovum which then undergoes cytogenesis. Alternatively, a diploid sperm fertilises an empty ovum. All chorionic villi are affected and there is often associated pleomorphism of trophoblast including the non-villous trophoblast of the placental bed. There is an absence of fibrosis. In contrast the partial mole is triploid (69,XXY 69,XXX or 69,XYY) and arises by dispermic fertilisation of a haploid ovum. Fetal parts may be present and there are two populations of villi. Enlarged villi (>3–4 mm) with central cavitation are a prominent feature. The villi have an irregular (angulated) profile with trophoblastic inclusions. The apparent excess of trophoblast may be subtle. There may be abnormal (angioectatic) vasculature in the second trimester. Pathological mimics of partial mole include: Beckwith–Wiedemann syndrome, placental angiomatous malformation, twin gestation with complete mole and existing fetus (see below), early complete mole (see below) and hydropic spontaneous miscarriage.15 Discordance in histopathological diagnosis is frequently seen in partial mole versus hydropic miscarriage and results from difficulty in evaluating trophoblastic hyperplasia. Ploidy analysis can improve concordance16 (see below).
Fig. 2 Early complete hydatidiform mole showing the very characteristic bulbous or polypoid profile of a molar chorionic villus accompanied by a myxoid stroma.
Early complete mole In the 1960s the mean age at evacuation of hydatidiform mole was 17 weeks. Nowadays, with the introduction of ultrasound examination as part of the routine clinical management of early pregnancy complications, it is 9.4 weeks, before the classical clinical or pathological features have developed. Thus, the histological features of complete hydatidiform mole are more subtle.17 A common error made by histopathologists in molar disease is the erroneous diagnosis of early complete hydatidiform mole as partial mole. Non-triploid partial moles probably do not exist.18 Certain typical features are seen only in early complete mole: abnormally shaped villi with a bulbous or polypoid profile, stromal mucin and stromal nuclear debris (Fig. 2, 3). We have recently shown (unpublished data) that the stromal debris which is a feature of early complete mole is the result of increased stromal proliferation and apoptosis (Table 2), a finding also recently demonstrated by Kim et al.19 Using an antibody to CD31, Kim et al. found that the number of mature blood vessels with distinct lumens in the
Fig. 3 High power view of early complete mole to show the villous stromal apoptosis.
PATHOLOGY OF GESTATIONAL TROPHOBLASTIC DISEASE
TABLE 2 Ki-67 indices of villous stroma* Diagnosis
Mean
n
SD
Early complete mole Partial mole Hydropic miscarriage
22.31 8.33 5.19
30 22 18
5.75 5.66 4.28
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
*Unpublished data. p,0.05. SD, standard deviation.
villous stroma was significantly reduced in both early complete mole and early partial mole compared with normal pregnancy. However, the number of CD31 positive primitive stromal cells or immature vascular networks without lumens did not differ significantly between the three groups. They attributed the retarded vasculogenic differentiation in early complete mole to increased apoptosis in the precursor components of blood vessels. In our study there was a statistically significant strong positive correlation between Ki-67 proliferation index and p53 expression (rho50.840; p,0.05). Ki-67 proliferation index and p53 expression of villous stromal cells were strongly correlated with the amount of stromal karyorrhetic debris (rho50.791; p,0.05). Ploidy analysis and p57kip2 Ploidy analysis is of great value in making the distinction between a diploid complete mole and a triploid partial mole but cannot be used to make a diagnosis of hydatidiform mole per se. It may also aid the distinction between a diploid hydropic miscarriage and a triploid partial mole. This may be achieved by flow cytometry or digital image analysis. In my own laboratory we have moved exclusively to the use of digital image analysis of microdissected tissue from formalin fixed, paraffin embedded blocks which, unlike flow cytometry, allows precise identification of the cells undergoing analysis (Fig. 4, 5). The greatest advance in the histopathology of molar pregnancy in recent years has been the introduction of immunohistochemistry for p57kip2.2,20–23 p57kip2 is a
Fig. 4 Digital image ploidy analysis of diploid complete hydatidiform mole. Note that the cells analysed are present in the gallery behind the histogram.
91
paternally imprinted gene located on chromosome 11p15.5, which is maternally expressed. The villous cytotrophoblast is therefore p57kip2 negative in androgenetic complete mole, whereas the villous cytotrophoblast is p57kip2 positive in normal placenta, hydropic miscarriage and partial mole. The syncytiotrophoblast is always p57kip2 negative. An interesting finding is that the non-villous (extravillous or intermediate) trophoblast is p57kip2 positive even in complete mole.23 As yet, there is no satisfactory biological explanation for this interesting phenomenon. Combined digital image ploidy analysis and p57kip2 can now be used in a complementary way to facilitate the diagnosis of complete and partial mole21 (Table 3). Twin molar pregnancy Twin molar pregnancies are dizygotic gestations, usually with a normal fetus and complete mole. Obstetric management may be complicated and fetal and maternal morbidity are increased. During the period 1988–2004, 7200 cases of gestational trophoblastic disease were registered with the Sheffield Trophoblast Centre. Among these patients, 30 cases of clinically or histopathologically suspected twin molar gestations were recorded (0.4% of all molar pregnancies). Nine cases were erroneously diagnosed on clinical grounds. Of those in which preliminary histological examination was suggestive of mole plus twin, tissue was available for expert review in 14 of 19 cases. Only seven (50%) of these were confirmed as complete mole/normal twin. During the same period, two cases provisionally diagnosed as partial mole were found on expert review to be a complete mole plus normal twin. Of the other seven, three were complete mole only, two were partial mole only and two were non-molar; one a normal fetus and the other a normal fetus with placental mosaicism.24 Thus, cases of twin molar gestations comprising a complete mole with coexisting fetus are often misinterpreted as partial molar singleton pregnancies because the presence of normal and molar villi is misinterpreted as indicating partial mole (Fig. 6). Immunohistochemistry for p57kip2 can facilitate the identification of two populations of villous cytotrophoblast in complete mole/twin pregnancy (Fig. 7).
Fig. 5
Digital image ploidy analysis of triploid partial hydatidiform mole.
92
WELLS
Pathology (2007), 39(1), February
TABLE 3 Complementary use of ploidy analysis and p57kip2 immunohistochemistry to facilitate the diagnosis of hydatidiform mole Suspected diagnosis Partial mole Complete mole Partial mole Partial mole
Flow cytometry
Image cytometry
p57kip2 status
Revised diagnosis
Diploid* Triploid Diploid Diploid
Triploid* Triploid Diploid Diploid
+ + 2 +
Partial mole Partial mole Complete mole Hydropic miscarriage
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
*Note greater reliability of digital image analysis compared with flow cytometry. +, positive; 2, negative.
Ectopic molar pregnancy Molar disease occurring in the context of ectopic pregnancy is rare and most reported cases are tubal pregnancies. Tubal molar pregnancy is over-diagnosed because histopathologists do not appreciate the exuberance of normal trophoblast in the early first trimester.25,26 Tertiary histopathological review is essential. However, normal early placentation, be it tubal or intrauterine, does not show trophoblast proliferation around the entire circumference of the villus, stromal apoptosis or scalloping of chorionic villi. Persistent trophoblastic disease (PTD) The histopathologist must appreciate that his or her role is limited once the diagnosis of hydatidiform mole has been made. The management of gestational trophoblastic disease is based on serial b-hCG estimations. If the b-hCG levels continue to rise or fail to fall, then the patient by definition has persistent trophoblastic disease (PTD). Thus, PTD is not a histopathological diagnosis and, indeed, the basic underlying cause may never be known if the patient responds to appropriate chemotherapy, for another biopsy specimen may never be taken. Usually PTD is due to invasive mole or choriocarcinoma. PTD occurs in 15–20% of patients with complete mole, and is rare following partial mole (remember that many early complete moles are misdiagnosed as partial moles). The majority of cases are due to invasive mole; however, this definitive diagnosis will only be made nowadays in those rare cases that are recalcitrant to chemotherapy for which hysterectomy is performed. It is entirely possible for a general histopathologist working in a district general hospital to go through an entire working lifetime without ever seeing such a case.
Fig. 6 Twin pregnancy with complete hydatidiform mole. In a single high power field normal chorionic villi and a molar villus are seen. Such an appearance may lead to an erroneous diagnosis of partial mole.
Prediction of persistent gestational trophoblastic disease Partial mole still carries a risk of developing gestational trophoblastic neoplasia, though the risk is much less than for complete moles, so follow-up is considered necessary for both complete and partial moles. Can the histopathologist predict those cases of hydatidifom mole that will go on to persistent trophoblastic disease? In asking this question I am not referring to the prognostic factors used by clinicians in their management of these patients to produce a prognostic score to separate patients into low risk and high risk groups for different treatments. Dr Cheung has addressed this question in a series of papers from Hong Kong. Hydatidiform moles which subsequently develop persistent trophoblastic disease, especially those which metastasise, are more likely to show telomerase activity.27 Apoptotic indices of those hydatidiform moles that spontaneously regress are statistically higher than those that develop persistent gestational trophoblastic neoplasia, using both the nick end labelling (TUNEL) technique28 and a M30 CytoDeath antibody.29 Using a human apoptosis array, increased expression of Mcl-1, an anti-apoptotic gene, has been detected in hydatidiform moles that subsequently develop into gestational trophoblastic neoplasia.30 This has been confirmed by quantitative PCR and protein expression analysis. cDNA microarrays and quantitative RNA analysis have shown down-regulation of ferritin light polypeptide (FTL) and IGFBP-1 in hydatidiform mole that subsequently developed gestational trophoblastic neoplasia when compared with those cases that regressed. Immunohistochemical analysis confirmed reduced IGFBP-1 protein expression in hydatidiform mole that developed gestational trophoblastic neoplasia.31
Fig. 7 p57kip2 immunohistochemistry in twin pregnancy with complete mole. The villous cytotrophoblast of the normal placenta is p57kip2 positive, whilst the complete mole is negative.
PATHOLOGY OF GESTATIONAL TROPHOBLASTIC DISEASE
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
Fig. 8
Hysterectomy specimen showing invasive complete mole.
The hysterectomy specimen performed for persistent gestational trophoblastic disease In rare cases, when the patient does not respond to chemotherapy and has persistently elevated b-hCG levels, there is little alternative to performing a hysterectomy. In such circumstances, it is likely that the hysterectomy specimen will show features of invasive mole, with venous invasion by molar chorionic villi (Fig. 8). There may also be extensive associated necrosis which is a reflection of the (albeit inadequate) cytopathic effect of the chemotherapy.
CHORIOCARCINOMA We are concerned here with choriocarcinoma as a gestational trophoblastic neoplasm; choriocarcinoma may also, of course, form part of the spectrum of nongestational malignant germ cell neoplasia. Under such circumstances, the tumour is often admixed with other malignant germ cell elements. In terms of the basic histopathology of choriocarcinoma, there have been few developments in recent years. The diagnosis rests on the recognition of the characteristic bimorphic appearance of malignant cyto- and syncytiotrophoblast, which is usually unmistakeable. However, insufficient attention has been paid in the literature to apparently ‘hybrid’ variants which show features that render them difficult to label either as choriocarcinoma on the one hand or placental site trophoblastic tumour on the other. Intraplacental choriocarcinoma I have recently encountered a 27-year-old female who presented with sudden onset of right-sided hemiparesis, headache and vomiting, 12 weeks after the spontaneous miscarriage of a 24 week (severely macerated but otherwise normal) fetus. She died 20 days after the onset of symptoms. Post-mortem examination showed disseminated trophoblastic disease but with no evidence of choriocarcinoma in the uterus. Review of the placental tissue showed a small focus of intraplacental choriocarcinoma comprising a biphasic proliferation of markedly atypical syncytio and cytotrophoblast with a high proliferation index (Fig. 9, 10). One clue to the diagnosis in this case was the presence of persistently elevated
93
Fig. 9 Intraplacental choriocarcinoma. A small focus of highly atypical trophoblast is arising from apparently normal chorionic villi.
serum a-fetoprotein levels, which are known to be a complication of choriocarcinoma. It is possible that the elevated serum a-fetoprotein is caused by increased amounts of fetal proteins crossing to the maternal circulation facilitated by mechanical breakdown of the materno-fetal barrier. Intraplacental choriocarcinoma is one of the most enigmatic conditions in the spectrum of gestational trophoblastic disease. It is rare and the placental primary is often small. It arises from the cytotrophoblast of apparently normal chorionic villi. Placental removal is sometimes curative but the lesion is more likely to be associated with maternal metastatic disease, as illustrated above. It is a difficult macroscopic diagnosis because the lesions may be small and resemble small infarcts. Unlike conventional choriocarcinoma, the presence of villi is not incompatible with the diagnosis. The tumour may present in pregnancy but is usually diagnosed several weeks or months after pregnancy (median 5 months). The lungs are the most common site of metastasis and, additionally, the possibility of choriocarcinoma should be borne in mind with any intracerebral bleed in a woman of child-bearing age.32
Fig. 10 High power view of intraplacental choriocarcinoma. Note the severe degree of nuclear atypia and the biphasic pattern of malignant cyto and syncytiotrophoblast.
94
WELLS
Pathology (2007), 39(1), February
LESIONS OF NON-VILLOUS TROPHOBLAST There are two benign lesions that are sometimes erroneously diagnosed as neoplastic: the (exaggerated) placental site reaction and the placental site nodule. In my opinion, too much emphasis has been placed in the descriptive morphological literature on the potential for diagnostic mishap associated with these lesions. The most important thing to remember when confronted with such cases is to gain access to the necessary clinical information which includes knowledge of the b-hCG status of the patient.
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
(Exaggerated) placental site reaction It is probably inappropriate to regard the so-called exaggerated placental site reaction as a lesion since, in my experience the normal implantation site often shows exuberant extravillous trophoblast that can seem alarming to the inexperienced. Placental site nodule This is usually a reasonably well circumscribed lesion of the endometrium composed of rather degenerate-looking extravillous trophoblast set in a hyaline-type matrix. It marks the residuum of involuted pregnancy and may be present many months after the pregnancy, of which there may be no clinical (or indeed patient) awareness. It is usually diagnosed as an incidental finding and is of no clinical significance except that it may perturb the histopathologist and lead him or her to an erroneous diagnosis of placental site trophoblastic tumour. However, the relatively discrete nature of the lesion, the degenerate appearance of the trophoblast and the lack of true cytological atypia in the form of nuclear pleomorphism should detract from this diagnosis. Kurman believes that the placental site nodule originates from the trophoblast of the chorion leve with a similar immunohistochemical profile to that of epithelioid trophoblastic tumour.33,34 Placental site trophoblastic tumour This is a neoplasm of the implantation site non-villous trophoblast and, in contrast to choriocarcinoma, occurs mostly after regular pregnancies or spontaneous miscarriages. It is generally not associated with the presence of chorionic villi. Invasion of the myometrium with a dissecting growth pattern is a prominent feature and the typical vascular changes have been referred to above. Despite the weak immunoreactivity for b-hCG of individual placental site trophoblastic tumour cells, serum levels are usually moderately elevated due to the bulk of tumour. Severe nuclear pleomorphism and high mitotic activity are associated with malignant behaviour.1 Epithelioid trophoblastic tumour This relatively recently described tumour of trophoblast resembles the trophoblastic cells of the chorion leve. It can be associated with any gestational event in younger women, but can also be diagnosed more than 10 years after the last known pregnancy or in post-menopausal women. It is composed of a relatively uniform population of mononucleate trophoblastic cells with eosinophilic cytoplasm surrounded by a well-defined cell membrane, intimately associated with a fibrillar hyaline-like matrix34 (Fig. 11).
Fig. 11 Epithelioid trophoblastic tumour. Note the uniform mononucleate cells set in a hyaline-like matrix. This case was initially erroneously diagnosed as epithelioid leiomyosarcoma.
PRAGMATIC IMMUNOHISTOCHEMICAL APPROACH TO THE DIAGNOSIS OF LESIONS OF THE NON-VILLOUS TROPHOBLAST IN BIOPSY SPECIMENS The first important task is to confirm the trophoblastic nature of the cells of interest. Though not specific for trophoblast, cytokeratin (e.g., CAM5.2; AE1/AE3; cytokeratin 18) immunohistochemistry will readily discriminate the intensely immunoreactive trophoblast of the placental site, whether neoplastic or non-neoplastic, from the cytokeratin negative maternal decidua (Fig. 12). Such a simple initial approach will also effectively exclude malignant melanoma. However, cytokeratin immunohistochemistry alone will not exclude other types of epithelial cells such as squamous carcinoma, which may be particularly important in the differential diagnosis of epithelioid trophoblastic tumour, though cytokeratin 18 is not expressed in squamous carcinoma of the cervix. Immunohistochemistry for human placental lactogen, inhibin-a, HLA-G and Mel-CAM (CD146) is helpful in this regard, all of which are positive in non-villous trophoblast. Squamous carcinoma cells are likely to be
Fig. 12 Intense immunoreactivity of non-villous trophoblast for low molecular weight cytokeratin (this is the epithelioid trophoblastic tumour seen in Fig. 11).
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
PATHOLOGY OF GESTATIONAL TROPHOBLASTIC DISEASE
p16 positive as a consequence of human papillomavirus infection, whereas trophoblast is negative for p16.35 The possibility of a smooth muscle lesion such as epithelioid leiomyosarcoma should always be borne in mind and excluded by demonstrating positivity with the trophoblast markers already cited and negativity with appropriate smooth muscle markers. However, the most difficult histological distinction in small biopsy specimens is between non-neoplastic and neoplastic trophoblast and, it must be conceded, that the immunohistochemical armamentarium is of little help in this regard, the exception to this being Ki-67; the Ki-67 labelling index is likely to be significantly elevated (.10%) in a neoplastic lesion of non-villous trophoblast.36,37 Recently, it has been shown that cyclin E is expressed in the trophoblastic columns of anchoring villi, implantation site non-villous trophoblastic cells and choriocarcinoma.35,37 Epithelioid trophoblastic tumours demonstrate a much higher cyclin E staining score than placental site nodules permitting distinction between the two.37,38 To be candid, I have never performed Ki-67 or cyclin immunohistochemistry in this situation and there really is no substitute for careful clinicopathological evaluation. A dogmatic diagnosis of placental site trophoblastic tumour will be followed by embarrassment if the clinicians are already aware that serum levels of b-hCG or human placental lactogen are not significantly elevated and that imaging shows no evidence of tumour. In practical clinical terms, the distinction between placental site trophoblastic tumour and epithelioid trophoblastic tumour is not really an important one.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
OUTSTANDING QUESTIONS Whilst there has been a dramatic increase in our knowledge of the biology of trophoblast and the various pathological lesions derived from it, certain outstanding questions remain. To my mind the most important of these are as follows: 1. The cell of origin of choriocarcinoma. 2. The key factor(s) in determining the cessation of trophoblastic invasion at the inner one-third of the myometrium. 3. The more accurate prediction of those cases of hydatidiform mole that will go on to persistent trophoblastic disease.
16.
17.
18.
19.
20. Address for correspondence: Professor M. Wells, Academic Unit of Pathology, University of Sheffield Medical School, Beech Hill Road, Sheffield S10 2RX, United Kingdom. Email: [email protected]
References 1. Paradinas FJ, Elston CW. Gestational trophoblastic diseases. In: Fox H, Wells M. Haines and Taylor Obstetrical and Gynaecological Pathology. 5th ed. London: Churchill Livingstone, 2003; 1359–430. 2. Genest DR, Dorfman DM, Castrillon DH. Ploidy and imprinting in hydatidiform moles. Complementary use of flow cytometry and immunohistochemistry of the imprinted gene product p57kip2 to assist molar classification. J Reprod Med 2002; 47: 342–6. 3. Lala PK, Chakraborty C. Factors regulating trophoblast migration and invasiveness: possible derangements contributing to preeclampsia and fetal injury. Placenta 2003; 24: 575–87. 4. Red-Horse K, Kapidzic M, Zhou Y, Feng KT, Singh H, Fisher SJ. EPHB4 regulates chemokine-evoked trophoblast responses: a
21.
22.
23.
24.
25.
95
mechanism for incorporating the human placenta into the maternal circulation. Development 2005; 132: 4097–106. Xu G, Guimond M-J, Chakraborty G, Lala PK. Control of proliferation, migration and invasiveness of human extravillous trophoblast by decorin, a decidual product. Biol Reprod 2002; 67: 681–9. Xu G, Chakraborty C, Lala PK. Expression of TGF-beta signalling genes in the normal, premalignant and malignant human trophoblast: loss of Smad3 in choriocarcinoma cells. Biochem Biophys Res Commun 2001; 287: 47–55. Xu G, Chakraborty C, Lala PK. Restoration of TGF-beta regulation of plasminogen activator inhibitor-1 in Smad3-restituted human choriocarcinoma cells. Biochem Biophys Res Commun 2002; 294: 1079–86. Xu G, Chakraborty C, Lala PK. Reconstitution of Smad3 restores TGF-b response of tissue inhibitor of metalloprotease-1 upregulation in human choriocarcinoma cells. Biochem Biophys Res Commun 2003; 300: 383–90. Mckinnon T, Chakraborty C, Gleeson L, Chidiac D, Lala PK. Stimulation of human extravillous trophoblast (EVT) migration by IGF-II is mediated by IGF type 2 receptor, involving inhibiting G proteins and phosphorylation of MAPK. J Clin Endocrinol Metab 2001; 86: 3665–74. Gleeson L, Chakraborty C, Mckinnon T, Lala PK. Insulin-like growth factor binding protein-1 stimulates human trophoblast migration by signalling through a5b1 integrin via mitogen-activated protein kinase pathway. J Clin Endocrinol Metab 2001; 86: 2484–93. Liu J, Chakraborty C, Barbin YP, Dixon SJ, Lala PK. Non catalytic domain of uPA stimulates human extravillous trophoblast migration by utilising phospholipase C, P-1-3 kinase and MAP kinase. Exp Cell Res 2003; 286: 138–51. Shih IM, Kurman RJ. p63 expression is useful in the distinction of epithelioid and placental site trophoblastic tumours by profiling trophoblastic subpopulations. Am J Surg Pathol 2004; 28: 1177–83. Wells M, Hsi B-L, Faulk WP. Class 1 antigens of the major histocompatibility complex on cytotrophoblast of the human placental basal plate. Am J Reprod Immunol 1984; 6: 167–74. Singer G, Kurman RJ, McMaster M, Shih I-M. HLA-G immunoreactivity is specific for intermediate trophoblast in gestational trophoblastic disease and can serve as a useful marker in differential diagnosis. Am J Surg Pathol 2002; 26: 914–20. Genest DR. Partial hydatidiform mole; clinicopathological features, differential diagnosis, ploidy and molecular studies, and gold standards for diagnosis. Int J Gynecol Pathol 2001; 20: 315–22. Fukunaga M, Katabuchi H, Nagasaka T, Mikami Y, Minamiguchi S, Lage JM. Interobserver and intraobserver variability in the diagnosis of hydatidiform mole. Am J Surg Pathol 2005; 29: 942–7. Sebire NJ, Makrydimas G, Agnantis NJ, Zagorianakou N, Rees H, Fisher RA. Updated diagnostic criteria for partial and complete hydatidiform moles in early pregnancy. Anticancer Res 2003; 23: 1723–8. Genest DR, Ruiz RE, Weremowicz S, Berkowitz RS, Goldstein DP, Dorfman DM. Do nontriploid partial hydatidiform moles exist? A histologic and flow cytometric reevaluation of nontriploid specimens. J Reprod Med 2002; 47: 363–8. Kim MJ, Kim KR, Ro JY, Lage JM, Lee HI. Diagnostic and pathogenetic significance of increased stromal apoptosis and incomplete vasculogenesis in complete hydatidiform moles in very early pregnancy periods. Am J Surg Pathol 2006; 30: 362–9. Jun SY, Ro JY, Kim KR. p57kip2 is useful in the classification and differential diagnosis of complete and partial hydatidiform moles. Histopathology 2003; 43: 17–25. Crisp H, Burton JL, Stewart R, Wells M. Refining the diagnosis of hydatidiform mole: image ploidy analysis and p57KIP2 immunohistochemistry. Histopathology 2003; 43: 363–73. Merchant SH, Amin MB, Viswanatha DS, Malhotra RK, Moehlenkamp C, Joste NE. p57KIP2 immunohistochemistry in early molar pregnancies: emphasis on its complementary role in the differential diagnosis of hydropic abortuses. Hum Pathol 2005; 36: 180–6. Sebire NJ, Lindsay I. p57KIP2 immunostaining in the diagnosis of complete versus partial hydatidiform moles. Histopathology 2006; 48: 873–4. Hancock BW, Martin K, Evans CA, Everard JE, Wells M. Twin mole and viable fetus: the case for misdiagnosis. J Reprod Med 2006; 51: 825–8. Burton JL, Lidbury EA, Gillespie AM, et al. Over-diagnosis of hydatidiform mole in early tubal ectopic pregnancy. Histopathology 2001; 38: 409–17.
Pathology Downloaded from informahealthcare.com by University of Chicago on 09/09/13 For personal use only.
96
WELLS
26. Sebire NJ, Lindsay I, Fisher RA, Savage P, Seckl MJ. Overdiagnosis of complete and partial hydatidiform mole in tubal ectopic pregnancies. Int J Gynecol Pathol 2005; 24: 260–4. 27. Cheung AN, Zhang DK, Liu Y, Ngan HY, Shen DH, Tsao SW. Telomerase activity in gestational trophoblastic disease. J Clin Pathol 1999; 52: 588–92. 28. Wong SY, Ngan HY, Chan CC, Cheung AN. Apoptosis in gestational trophoblastic disease is correlated with clinical outcome and Bcl-2 expression but not Bax expression. Mod Pathol 1999; 12: 1025–33. 29. Chiu PM, Ngan HY, Chan CC, Cheung AN. Apoptosis in gestational trophoblastic disease correlates with clinical outcome: assessment by the caspase-related M30 CytoDeath antibody. Histopathology 2001; 38: 243–9. 30. Fong PY, Xue WC, Ngan HY, et al. Mcl-1 expression in gestational trophoblastic disease correlates with clinical outcome: a differential expression study. Cancer 2005; 103: 268–76. 31. Feng HC, Tsao SW, Ngan HY, Xue WC, Chiu PM, Cheung AN. Differential expression of insulin-like growth factor binding protein 1 and ferritin light polypeptide in gestational trophoblastic neoplasia: combined cDNA suppression subtractive hybridization and microarray study. Cancer 2005; 104: 2409–16. 32. Fernando MS, Wells M, Carroll T, Hancock BW, Wharton SB, Intracranial metastasis from an intraplacental choriocarcinoma. (Submitted for publication.) 33. Shih IM, Seidman JD, Kurman RJ. Placental site nodule and characterization of distinctive types of intermediate trophoblast. Hum Pathol 1999; 30: 687–94. 34. Shih I-M, Kurman RJ. Epithelioid trophoblastic tumour – a neoplasm distinct from choriocarcinoma and placental site trophoblastic tumor simulating carcinoma. Am J Surg Pathol 1998; 22: 1393–403. 35. Mao TL, Seidman JD, Kurman RJ, Shih I-M. Cyclin E and p16 immunoreactivity in epithelioid trophoblastic tumour – an aid in differential diagnosis. Am J Surg Pathol 2006; 30: 1105–10. 36. Shih IM, Kurman RJ. Ki-67 labeling index in the differential diagnosis of exaggerated placental site, placental site trophoblastic tumour and choriocarcinoma: a double immunohistochemical staining technique using Ki-67 and Mel-CAM antibodies. Hum Pathol 1998; 29: 27–33. 37. Bamberger AM, Aupers S, Milde-Langosch K, Loning T. Expression pattern of the cell cycle promoter cyclin E in benign extravillous trophoblast and gestational trophoblastic lesions: correlation with expression of Ki-67. Int J Gynecol Pathol 2003; 22: 156–61.
Michael Wells: personal contribution Original articles Sinha D, Wells M, Faulk WP. Immunological studies of human placentae: complement components in pre–eclamptic chorionic villi. Clin Exp Immunol 1984; 156: 175–84. Wells M, Hsi B-L, Yeh C-JG, Faulk WP. Spiral (utero–placental) arteries of the human placental bed show the presence of amniotic basement membrane antigens. Am J Obstet Gynecol 1984; 150: 937–77. Wells M, Hsi B-L, Faulk WP. Class 1 antigens of the major histocompatibility complex on cytotrophoblast of the human placental basal plate. Am J Reprod Immunol 1984; 6: 167–74. Earl UM, Bulmer JN, Wells M. Immunohistological identification of antigen expression by trophoblast populations in a case of ectopic cornual implantation. Br J Obstet Gynaecol 1985; 92: 843–6. Earl UM, Wells M, Bulmer JN. The expression of major histocompatibility complex antigens by trophoblast in ectopic tubal pregnancy. J Reprod Immunol 1985; 8: 13–24. Bulmer JN, Wells M, Bhabra K, Johnson PM. Immunohistological characterisation of endometrial gland epithelium and extra–villous fetal trophoblast in third trimester human placental bed tissues. Br J Obstet Gynaecol 1986; 93: 823–32.
Pathology (2007), 39(1), February
Earl UM, Wells M, Bulmer JN. Immunohistochemical characterisation of trophoblast membrane antigens and secretory products in ectopic tubal pregnancy. Int J Gynecol Pathol 1986; 5: 132–42. Lalani E-NMA, Bulmer JN, Wells M. Peroxidase labelled lectin binding of human extra–villous trophoblast. Placenta 1987; 8: 15–26. Hemming JD, Quirke P, Womack C, Wells M, Elston CW, Bird CC. The diagnosis of molar pregnancy and persistent trophoblastic disease by flow cytometry. J Clin Pathol 1987; 40: 615–20. Bulmer JN, Wells M, Lunny DP, Yeh C-J, Hsi B-L. An investigation of the expression of amnion antigens by spiral arteries in human uteroplacental tissues. Am J Reprod Immunol 1987; 14: 79–83. Wells M, Bennett J, Bulmer JN, Jackson P, Holgate CS. Complement component deposition in uteroplacental spiral arteries in normal human pregnancy. J Reprod Immunol 1987; 12: 125–35. Bulmer JN, Smith JC, Morrison L, Wells M. Maternal and fetal cellular relationships in the human placental basal plate. Placenta 1988; 9: 237–46. Hemming JD, Quirke P, Womack C, Wells M, Elston CW, Pennington GW. Flow cytometry in persistent trophoblast disease. Placenta 1988; 9: 615–20. Thornton JG, Lewis FA, Linton G, Wells M, Tyrrell S, Lilford RJ. Fetal sexing by chorionic villus biopsy and in–situ DNA hybridization with a Y probe and biotin–streptavidin polyalkaline phosphatase labelling. J Obstet Gynaecol 1989; 10: 1–4. Andrew AC, Bulmer JN, Wells M, Morrison L, Buckley CH. Subinvolution of the uteroplacental arteries in the placental bed. Histopathology 1989; 15: 431–3. Thrower S, Bulmer JN, Wells M. Expression of epidermal growth factor receptor and transferrin receptor by human trophoblast populations. Am J Reprod Immunol 1989; 21: 87–93. Thrower S, Bulmer JN, Griffin NR, Wells M. Further studies of lectin biding by villous and extravillous trophoblast in normal and pathological pregnancy. Int J Gynecol Pathol 1991; 10: 238–51. Andrew A, Bulmer JN, Morrison L, Wells M, Buckley CH. Subinvolution of the uteroplacental arteries: an immunohistochemical study. Int J Gynecol Pathol 1993; 12: 28–33. Burton Jl, Lidbury EA, Gillespie AM, et al. Over-diagnosis of hydatidiform mole in early tubal ectopic pregnancy. Histopathology 2001; 38: 409–417. Crisp H, Burton JL, Smith O, Stewart R, Wells M. Refining the diagnosis of hydatidiform mole: image ploidy analysis and p57kip2 immunohistochemistry. Histopathology 2003; 43: 363–73. Hassadia A, Gillespie A, Tidy J, Everard Rgn J, Wells M, Coleman R, Hancock B. Placental site trophoblastic tumour: Clinical features and management. Gynecol Oncol 2005; 99: 603–7. Hancock BW, Martin K, Evans CA, Everard JE, Wells M. Twin mole and viable fetus: the case for misdiagnosis. J Reprod Med 2006; 51: 825–8. Chapters in books Wells M, Fox H. Immunology and immunopathology of the materno-fetal interface. In: Coulam CB, McIntyre JA, Faulk WP, editors. Immunological Obstetrics. New York: Norton Medical Books, 1992; 166–76. Coulam CB, Wells M. Ectopic pregnancy loss. In: Coulam CB, McIntyre JA, Faulk WP, editors. Immunological Obstetrics. New York: Norton Medical Books, 1992; 464–78. Wells M, Mohamdee O. Pathology of the pregnant uterus. In: Fox H, Wells M, editors. Haines and Taylor: Obstetrical and Gynaecological Pathology. London: Churchill Livingstone, 1995; 1509–37. Fox H, Wells M. Pathology of the first trimester. In: Kurjak A, editor-inchief. A Textbook of Perinatal Medicine. London: Parthenon, 1998; 991–8. Hustin J, Wells M. Pathology of the pregnant uterus. In: Fox H, Wells M, editors. Haines and Taylor: Obstetrical and Gynaecological Pathology. London: Churchill Livingstone, 2003; 1327–57. Review articles Wells M, Bulmer JN. The human placental bed: anatomy immunohistochemistry and pathology. Histopathology 1988; 13: 483–98. Berry CW, Brambati B, Eskes TKAB, et al. The Euro-Team Early Pregnancy (ETEP) protocol for recurrent miscarriage. Hum Reprod 1995; 10: 1516–20.