Partial mole

The genetic origins of complete and partial molar pregnancies are demonstrated in Fig. 42.1. Partial moles are triploid with 69 chromosomes comprising two sets of paternal and one set of maternal chromosomes. Macroscopically and on ultrasound scanning during the first trimester, partial mole will often resemble normal products of conception. The embryo can appear viable on an early ultrasound scan but becomes non‐viable by week 10–12. The histology of partial mole shows less swelling of the chorionic villi than in a complete mole and there may be only focal changes. As a result the diagnosis of partial mole can often be missed after a miscarriage or evacuation, unless the products are sent for expert pathological review.

The clinical presentation of partial mole is most frequently via a failed pregnancy rather than irregular bleeding or by detection on routine ultrasound. The obstetric management is by suction evacuation and histological review; all patients with partial mole should be followed up by registration and serial hCG measurement.

Fortunately, partial mole rarely transforms into malignant disease, and there is an overall risk of 0.5–1% of patients requiring chemotherapy after a partial mole [10].

Complete mole

In complete moles the genetic material is totally male in origin (Fig. 42.1), resulting from the fertilization of an ‘empty’ anucleate oocyte lacking maternal DNA. The chromosome complement is most commonly 46XX, which results from one X chromosome‐carrying sperm that duplicates its DNA, or less frequently 46XY or 46XX from the presence of two separate sperm [1].

The clinical diagnosis of complete mole most often occurs as a result of first‐trimester bleeding or an abnormal ultrasound. There is no fetal material and the histology shows the characteristic oedematous villous stroma. The textbook ‘bunch of grapes’ appearance is only seen in the second trimester and as most cases are diagnosed earlier, this is now rarely observed in UK practice. The typical macroscopic appearances of a complete mole are shown in Fig. 42.2. Obstetric management is by suction evacuation followed by registration and serial hCG measurement. Complete molar pregnancies have an appreciable risk of proceeding to invasive disease, with approximately 15% of patients with complete mole requiring chemotherapy.

Familial recurrent hydatidiform mole syndrome

Occasionally we see patients who have repetitive complete molar pregnancies and who have failed to conceive a normal baby. Patients who have had three consecutive complete moles should be genetically tested to determine whether they have FRHM syndrome where the DNA is biparental in origin rather than androgenetic [11]. Thus far, mutations in two genes have been associated with this autosomal recessive condition, NLRP7 [12] and KHD3CL [13], that account for about 75% and 5–10% of cases respectively, so additional gene(s) remain to be discovered [14]. Affected women have rarely been reported to achieve a normal pregnancy, with most just suffering repeated complete moles or pregnancy losses where no pathology is available. Unfortunately, the risk of subsequent malignancy requiring chemotherapy is the same for these patients as for those with androgenetic complete moles. Therefore recommending repeated rounds of attempted natural conception does not seem appropriate. As these genes are normally expressed in both the egg and the genital tract/uterus, the question then arises as to whether the recurrent moles are due to a defective egg or a problem with uterine implantation. To help resolve this, one affected patient agreed to try egg donation from an unaffected relative and this resulted in a normal pregnancy [15]; since then, several other women with FRHM have similarly had normal babies.

Twin pregnancies

The risk of having a twin pregnancy comprising a complete mole and healthy co‐twin is estimated to be 1–5 per 100 000 [16,17]. These pregnancies cause considerable anxiety for both the patient and managing clinicians as there is a perception that they are likely to be associated with an increased risk of serious and potentially life‐threatening complications for the mother, including haemorrhage, uterine perforation, pre‐eclampsia, toxaemia and malignant change in the mole requiring chemotherapy. In addition, it was thought that the possibility of carrying the healthy co‐twin to viability or term was remote. These concerns were underpinned by several small case series based on local hospital reports [17] and so suffered from case ascertainment bias. We examined this issue in a UK population‐based series of 77 women that was the largest of its kind and which was not subject to case ascertainment bias. To our surprise, we discovered that the risk of malignancy was not increased by allowing such pregnancies to continue, a fact that we subsequently substantiated in singleton molar pregnancies evacuated in the second and third rather than first trimester [18]. Moreover, nearly 40% of them resulted in a healthy take‐home baby with most of the rest undergoing spontaneous loss from presumed continued molar growth. Importantly, there was only a small increased risk in pre‐eclampsia that could be safely managed and no cases of uterine perforation or maternal death [17]. Since then a further 90 patients have been offered the choice of early termination versus continuing their pregnancies. This has confirmed our earlier findings, with no maternal losses and no increased risk of malignancy by allowing the pregnancy to continue. Moreover, the healthy take‐home baby rate has risen to just under 60% (unpublished observations). The mode of delivery has been a mixture of normal vaginal and caesarean section depending on local assessment of what was considered the safest option. Therefore, after appropriate counselling, it seems reasonable to allow women to continue such pregnancies but under close monitoring to ensure that potential complications are detected early and dealt with promptly.

Atypical placental site nodules

Following a normal pregnancy, placental remnants in the uterine wall usually spontaneously regress and disappear. Occasionally, they may persist and form placental site nodules. These may present with irregular bleeding that often results in dilation and curettage. The subsequent pathology reveals the placental site nodule material. This has always been regarded as a benign condition. However, recent data suggest that some placental site nodules may have atypical histopathological features and, if so, can in 10–15% of cases be associated with or subsequently develop into either a PSTT or ETT [5]. Consequently, patients found to have placental site nodules where there may be atypical features are now having their pathology reviewed centrally in the UK. If atypical placental site nodule is confirmed, then these patients are registered and seen centrally for monitoring. If the nodule is thought to be typical, then at present local MRI pelvis imaging is suggested and if this is normal no further monitoring is currently undertaken. However, as we learn more it is likely that the advice will change.

Registration and surveillance

Overall, 90% of patients with molar pregnancies will not need any additional treatment following their evacuation. In these patients the residual trophoblast tissue cells fail to proliferate and as the cells stop growing, hCG levels return to normal. At present there is no effective prognostic method that allows accurate distinction between the patients who will develop malignant disease after evacuation and the majority who will not [19]. As a result all patients with molar pregnancy should be registered for an hCG follow‐up system. The use of this approach allows the early identification of patients whose disease is continuing to proliferate, while also allowing careful observation of patients with more slowly falling hCG levels, so minimizing unnecessary chemotherapy.

Patients taking part in an hCG surveillance programme usually have the need for additional treatment determined by the pattern of their hCG results. Thus, an hCG level that has plateaued or is rising following evacuation of a molar pregnancy is the commonest indication of malignant change requiring chemotherapy. Table 42.2 shows the Charing Cross Hospital recommended indications for treatment. The first four of these are also used by the International Federation of Gynecology and Obstetrics (FIGO) [2]. However, based on our recent work, the second indication has been dropped because nearly all women with a falling but elevated hCG 6 months after evacuation will normalize hCG levels spontaneously without chemotherapy [20]. The last three indications are UK‐specific reasons for commencing treatment. Typically, about 1–2% of patients start therapy because of excessive bleeding requiring transfusion even if hCG is falling because the treatment helps to settle the bleeding. An hCG concentration above 20 000 IU/L 4 weeks after uterine evacuation is associated with a high risk of developing malignant change that requires chemotherapy and with a small risk of uterine perforation and accounts for probably another 1–2% of patients who require chemotherapy in the UK. The presence of more than 2 cm lung and/or vaginal metastases as a sole indication for treatment has not been seen for over 25 years, likely because of improved early diagnosis within the first trimester and better surveillance. Those patients recovering from molar pregnancy enrolled in the surveillance service and who go on to require treatment have a cure rate approaching 100% [21], and over 95% will initially fall into the low‐risk treatment group.

Invasive mole (chorioadenoma destruens)

Invasive mole nearly always arises from a complete mole and is characterized by invasion of malignant cells into the myometrium, which can lead to perforation of the uterus. Microscopically, invasive mole has a similar histological appearance to complete mole but is characterized by the ability to invade into the myometrium and local structures if untreated. Fortunately, the incidence of invasive mole in the UK has fallen substantially with the introduction of routine ultrasonography, the early evacuation of complete moles and effective hCG surveillance, and is now rare.

Choriocarcinoma

Choriocarcinoma is histologically and clinically overtly malignant and presents the most frequent emergency medical problems in the management of trophoblast disease. The diagnosis often follows a complete mole, when the patients are usually in a surveillance programme, but can also arise after any other type of pregnancy including a partial mole, a non‐molar abortion, miscarriage, ectopic or term pregnancy. The clinical presentation of choriocarcinoma can be due to local disease in the uterus leading to bleeding in two‐thirds of cases, or from distant metastases that can cause a wide variety of symptoms, with the lungs, central nervous system (CNS) and liver the most frequent sites of distant disease. About one‐third of patients will have no obvious residual abnormality within the uterus [1].

The cases of choriocarcinoma presenting with symptoms from distant metastases can be diagnostically challenging [22]. However, the combination of the gynaecological history and elevated serum hCG usually makes the diagnosis clear and so avoids biopsy, which can be hazardous due to the risk of severe haemorrhage, as demonstrated following a liver biopsy (Fig. 42.3). On the occasions when choriocarcinoma pathology is available, the characteristic findings show the structure of the villous trophoblast but sheets of syncytiotrophoblast or cytotrophoblast cells, haemorrhage, necrosis and intravascular growth are common [1]. The genetic profile of choriocarcinoma can reflect its origin from a complete or partial mole [10] or a normal diploid conception combined with a range of gross abnormalities without any specific characteristic patterns [23]. Some cases will prove to be non‐gestational choriocarcinomas that pathologically are identical to gestational choriocarcinomas but are in fact germ cell or epthelial tumours that have trophoblastic differentiation and genetically originate from the patient. These tumours may initially respond well to combination agent chemotherapy used to treat gestational trophoblastic tumours but then revert to their true phenotype, and if epithelial in origin do badly [3].

Placental site trophoblast tumour and epithelioid trophoblastic tumour

PSTT was originally described in 1976 [6] and is the rarest form of GTD, accounting for 0.2% of all registered cases of GTD in the UK [4]. ETT is an even rarer variant first described in 1998 [24] and although it is thought to behave in a similar fashion to PSTT, in reality we are still learning about the similarities and differences of these two entities [2]. PSTT/ETT most commonly follows a normal pregnancy but can also occur after a non‐molar abortion or a molar pregnancy.

In contrast to the more common types of trophoblast disease, which characteristically present fairly soon after the index pregnancy, in PSTT the average interval between the prior pregnancy and presentation is 3.4 years. The most frequent presentations are amenorrhoea or abnormal bleeding. In nearly all cases the serum hCG level is elevated, but is characteristically lower for the volume of disease than in the other GTNs. The tumour can arise after any type of pregnancy including complete and partial moles [25] and is believed to be derived from the non‐villous trophoblast. The pathology is characterized by intermediate trophoblastic cells with vacuolated cytoplasm, the expression of placental alkaline phosphatase and hCG, and the absence of cytotrophoblast and villi. ETT develops from the chorionic‐type intermediate trophoblast present in other parts of the placenta such as the chorionic plate and fetal membranes. Pathologically, ETT is distinguished from PSTT by its smaller cells, its smaller, more monomorphic cells and by its nested, nodular, well‐circumscribed growth pattern unlike the sheet‐like infiltrative pattern seen with PSTT [26]. In addition, clinically, ETT tends to secrete even less hCG than PSTT. The clinical presentation of PSTT/ETT can range from slow‐growing disease limited to the uterus to metastatic disease, with lung and liver the most common sites of distant spread [7]. Spread to lymphatics is a bit more common in PSTT/ETT compared with choriocarcinoma where it is exceptionally rare [4].

Role of hCG in trophoblast disease diagnosis and management

Produced predominantly by syncytiotrophoblast cells, hCG is a glycosylated heterodimer protein consisting of α and β units held together by non‐covalent bonds. In malignant disease a number of hCG variants can occur, including hyperglycosylated hCG, nicked hCG, hCG missing the β subunit C‐terminal peptide and the free β subunit [1]. With the potential exception of a few atypical cases of PSTT/ETT, hCG is constitutively expressed by all forms of premalignant and malignant GTD.

The measurement of hCG allows estimation of tumour bulk, forms an important part of the assessment of the patient’s disease risk and provides a simple method to follow the response to treatment. The hCG level can be measured by a variety of immunoassays but at present there is no internationally standardized assay and the various commercially available kits used in different hospitals can vary in their ability to detect different forms of partially degraded hCG molecules and so can give divergent results and occasional false negatives [27]. It is also important to be aware that patients with exceptionally high hCG levels may produce a false‐negative result through the high‐dose ‘hook’ effect. Moreover, any assay might produce a false‐positive hCG value. Consequently, if the clinical situation does not correspond with an elevated hCG value, then another hCG assay should be employed and/or the hCG should be checked in the urine. This is because heterophilic antibodies, which are the commonest cause of false‐positive hCG values by occasionally allowing two independent assays to report positively, are too large (unlike hCG) to pass into the urine [1]. In the UK we are very fortunate to have an hCG assay that detects all forms of hCG produced in cancer and because of the way it is configured is highly unlikely to suffer false‐positive results.

In the absence of tumour hCG production, the serum half‐life of hCG is 24–36 hours; however, for patients receiving chemotherapy, total hCG levels characteristically show slower rates of fall as the tumour cells continue to produce some hCG as their number decreases with treatment.

Low‐risk disease management

In the UK, over 90% women with molar pregnancies who require additional treatment following their initial evacuation fall into the low‐risk treatment category as defined by the FIGO prognostic scoring system. The role of repeated uterine evacuation in the management of these patients has been a subject of uncertainty until recently. A number of studies looking at the impact of repeated evacuation in women with rising or static hCG levels following their first evacuation have suggested that a repeated procedure is rarely curative [30,31]. Based on these data, the current recommendation is that a repeated evacuation should only be considered if the hCG level is under 5000 IU/L and tissue is seen in the uterine cavity on ultrasound.

For patients meeting the FIGO standard for low‐risk treatment, the most widely used protocol is methotrexate given intramuscularly with oral folinic acid rescue following the schedule shown in Table 42.4. The first course of treatment should be administered in hospital with the subsequent courses administered at home. This is because they have a higher risk of bleeding as therapy starts, particularly if the tumour shrinks rapidly with the initial chemotherapy. Bleeding usually responds well to strict bed rest and less than 1% of low‐risk patients require emergency interventions such as embolization, vaginal packing or hysterectomy.

The low‐risk chemotherapy treatment is usually well tolerated without much major toxicity. Methotrexate does not cause alopecia or significant nausea and myelosuppression is extremely rare. Of the side effects that do occur, the most frequent problems, occurring in about 2–3% of cases, are mucositis, a gritty feeling in the eyes and, more rarely, pleuritic chest or abdominal pains from serositis. Mild elevation of liver function tests can occur but rarely prevents ongoing therapy. For low‐risk patients with lung metastases visible on their chest radiographs, the policy at Charing Cross Hospital is to perform MRI of the head and if normal to add CNS prophylaxis with intrathecal methotrexate administration to minimize the risk of development of CNS disease. A maximum of three doses of intrathecal methotrexate is given unless the cerebrospinal fluid to serum hCG ratio exceeds 1 : 60 [1].

Treatment is continued until normalization of the serum hCG level (0–4 IU/L) and then a further three cycles (6 weeks) to ensure eradication of any residual disease that is below the level of serological detection. In the Netherlands, the policy had been to only give two cycles of consolidation therapy after normalization of the hCG and their relapse rate was double that of the UK so the Dutch have recently modified their policy to emulate the UK [32]. A typical example of the treatment graph for a patient successfully completing methotrexate chemotherapy is shown in Fig. 42.4.

An overview of the data for the 1990s indicates that 67% of patients in the low‐risk group will only require treatment with the methotrexate protocol to successfully complete their therapy. Patients who have an inadequate response to methotrexate therapy, as shown by an hCG plateau or rise, have their treatment changed to second‐line therapy. This comprises single‐agent actinomycin D (1.25 mg/m² i.v. repeated every 2 weeks) which, like methotrexate, is relatively less toxic than the alternative etoposide, methotrexate, actinomycin D, cyclophosphamide and vincristine (EMA/CO) combination chemotherapy (Table 42.5). To minimize the number of patients proceeding to EMA/CO we have increased the hCG cut‐off at the point of methotrexate resistance from 100 to 300 IU/L, and since this resulted in equally good results [21,33] we are now testing an hCG cut‐off of 1000 IU/L.This means that only patients with an hCG above 1000 IU/L at the point of methotrexate resistance proceed to EMA/CO whilst the rest receive a trial of single‐agent actinomycin D.

An individual example of the pattern of hCG levels during the course of management is shown in Fig. 42.5. This demonstrates the rise in hCG that led to the introduction of methotrexate chemotherapy; following this, the hCG initially falls but after two cycles appears to plateau. The introduction of second‐line treatment with EMA/CO chemotherapy leads to a rapid fall in hCG to normal and the completion of chemotherapy after 6 weeks of further treatment. Overall the survival in this group is currently running at 100% and the sequential introduction of additional chemotherapy as necessary minimizes the potential long‐term carcinogenic risks of excess treatment.

High‐risk disease management

Historical data from treatment prior to the introduction of multiagent chemotherapy demonstrated that less than one‐third of high‐risk patients would be cured with single‐agent therapy [34]. The introduction of combination chemotherapy treatments in the 1970s transformed this situation, and modern data indicate a cure rate for high‐risk patients of 85–94% using EMA/CO chemotherapy []. This combination delivers a dose‐intense treatment with the five chemotherapy agents, delivered in two groups 1 week apart as shown in Table 42.5. Delivering therapy on a weekly as opposed to 3–4 weekly basis appears to be the most effective approach to this rapidly proliferating malignancy. EMA‐CO is myelosuppressive and most patients require granulocyte colony‐stimulating factor support each week to maintain treatment intensity and to minimize the risk of neutropenic sepsis. Nausea and vomiting are now rarely a problem with modern antiemetics that include the 5HT₃ receptor antagonists such as ondansetron combined with domperidone. The addition of dexamethesone is rarely required and we prefer to avoid this given the rare but unpleasant problem of avascular necrosis of the femoral and humeral heads requiring joint replacements in young patients. Increasing lethargy with progressive cycles of EMA/CO is common and some patients develop peripheral neuropathy that can be aided by stopping the vincristine. Overall, the regimen is well tolerated compared with many other multiagent therapies and serious or life‐threatening toxicity is very rare.

As in the low‐risk situation, treatment is continued for 6 weeks after the normalization of hCG. Moreover, the second day of the EMA comprising etoposide and actinomycin D can be dropped once the hCG is normal. An example of a high‐risk patient treated in this case with the high‐risk chemotherapy regimen is shown in Fig. 42.6, which demonstrates the resolution of extensive lung metastases and normalization of hCG levels in response to treatment.

Ultra‐high‐risk disease

Ultra‐high‐risk disease is a new term to try to define those high‐risk patients at greatest risk of death. Analysis of the causes of death in patients with high‐risk disease revealed that this occurred either early or late. The early deaths took place within 4 weeks of admission due to organ failure associated with one or more of the following: fatal haemorrhage, metabolic upset from tumour lysis and very adavanced disease. The late deaths were from multiresistant disease. In addition, some deaths were due to cancers that were in fact non‐gestational epithelial malignancies histopathologically mimicking gestational choriocarcinomas. Factors associated with early or late deaths included a FIGO score above 12, very advanced lung disease, liver with or without brain metastases, and an interval from the last‐known and presumed causative pregnancy greater than 2.8 years [3]. To reduce early deaths, we developed a gentle induction chemotherapy comprising low‐dose etoposide 100 mg/m² and cisplatin 20 mg/m² (EP) given intravenously on days 1 and 2 and repeated weekly for up to 3 weeks until the patient is well enough to tolerate standard‐dose therapy. This has almost completely eliminated early deaths from patients with advanced GTN [3]. Those patients with liver metastases or liver and brain metastases have a particularly poor long‐term survival and so they should be considered for more prolonged platinum‐containing therapy. In the UK these patients now all receive etoposide and cisplatin alternating with 1 day of EMA [38]. Moreover, for patients with brain metastases, the methotrexate dose is increased to 1 g/m² in the EMA and intrathecal methotrexate is given with the EP to ensure adequate CNS penetration. We also extend the duration of consolidation therapy in ultra‐high‐risk patients to 8 weeks. Further details of how to manage such complex patients is beyond the scope of this chapter but can be found elsewhere [38–40].

Salvage of EMA/CO failures

Of the high‐risk patients treated with EMA/CO, approximately 18–20% develop resistance to this combination and require a change to second‐line drug treatment. In this situation, the EP/EMA [41] or TE/TP [42] regimens may be used, which incorporate cisplatin and additional etoposide into the combination along with paclitaxel in the TE/TP regimen. These treatments, combined with surgery, both appear to salvage about 80% of patients failing EMA/CO. Since TE/TP seems much better tolerated and is given every 2 weeks rather than the weekly EP/EMA, TE/TP is now frequently the preferred option for EMA/CO failures [2].

Patients relapsing after EP/EMA and/or TE/TP can still be salvaged by surgery and/or further chemotherapy. Several other agents have activity in these tumours including gemcitabine, pemetrexed and capecitabine either alone or in combinations with other drugs. In addition, high‐dose chemotherapy with peripheral stem cell support may help to salvage some patients [2]. Recent work presented at international meetings has suggested that some antibodies that target surface molecules on either the trophoblastic tumour cells or surrounding immune cells may also salvage some patients who have failed all other therapies. Consequently, there is considerable optimism that the remaining 5% of deaths from highrisk disease will be eliminated in the near future.

Management of PSTT/ETT

PSTT/ETT, unlike choriocarcinoma, tend to be much more slow‐growing, remain confined to the uterus for longer, have a greater liklihood of lymphatic spread, produce less hCG, and tend to be less chemosensitive. However, like choriocarcinoma, they can metastasize widely.

The FIGO scoring system is not an effective way to determine management. Instead analysis of the UK experience has shown that the key independent prognostic factor is the interval from the last‐known or causative pregnancy. Patients presenting beyond 4 years do exceptionally badly whereas those presenting within 4 years have a 98% long‐term survival regardless of stage [4]. A more recent and as yet unpublished update suggests that stage is also an independent prognostic factor. Therapy is therefore stage and interval adapted. For patients with stage I disease within 4 years of the causative pregnancy, a simple hysterectomy with lymph node sampling and sparing of the ovaries but without adjuvant chemotherapy is all that is needed. However, those with stage I diease presenting more than 4 years from the causative pregnancy should be considered for adjuvant chemotherapy with or without high‐dose chemotherapy given the comparative reduced sensitivity to standard‐dose chemotherapy. Women who present with metastatic disease within 4 years are currently treated with EP/EMA chemotherapy until their hCG level is normal for 8 weeks and then offered surgery to remove all residual disease sites. If there is still active disease then high‐dose therapy is considered. The latter is now included for all metastatic patients presenting beyond 4 years, along with other novel therapies. Some young patients presenting with PSTT/ETT within 4 years of the presumed causative pregnncy are understandably keen to preserve their fertility and wish to avoid hysterectomy. Several experimental approaches are being examined including focal resection of uterine disease using a modified Strassman procedure, followed by either nothing or adjuvant chemotherapy. Alternatively, neoadjuvant chemotherapy has been tried followed by observation or focal resection if a residual uterine lesion is seen. The problem with these approaches resides in persisting diffuse microscopic disease causing subsequent relapse with the potential to place patients in a poor prognostic group [43]. These approaches therefore remain experimental.

To gain more insight into PSTT/ETT the International Society for the Study of Trophoblastic Diseases has established a database for these cases which is run by Sheffield and Charing Cross in the UK (http://stdc.group.shef.ac.uk/psttuhr/index.html).

Risk of relapse and late treatment complications

For the majority of patients with trophoblast disease who achieve a serological remission the outlook is very bright in terms of the very low risks of relapse, the high possibility of further successful pregnancies and only modest long‐term health risks from the chemotherapy exposure. Once the hCG level has fallen to normal, the risk of relapse is less than 4% for patients starting in the low‐risk category and 8% for patients in the high‐risk EMA/CO category [44]. Generally, recurrences occur within the first 12 months after treatment.

Subsequent fertility

Importantly, fertility is not clearly affected by single‐agent therapy with methotrexate and apart from hastening the menapause by 3 years [45], EMA/CO appears to also have no impact on the chances of a successful pregnancy, which for both methotrexate alone or EMA/CO is 83% [46]. Despite this good news, some patients attending for fertility testing have been very worried by reports of low to undetectable anti‐Müllerian hormone levels. Fortunately, all these women were menstruating normally, subsequently conceived and have had babies [47].

Further pregnancy should be deferred for 12 months after treatment to avoid any teratogenic effects on developing oocytes and to minimize possible confusion from the rising hCG levels between a new pregnancy and disease relapse. Despite this advice 500 of our treated patients have conceived in the first year of follow‐up. Interestingly, the risk of relapse appeared to be lower, there were no abnormal babies and only the expected rate of miscarriages/pregnancy losses and second molar pregnancies. Only one patient suffered severe difficulty with lung metastases but fortunately both the patient and her baby were saved [48,49]. Therefore, for women who are worried about their declining fertility, it seems reasonable to let them make their own informed choice about earlier attempts to conceive.

Many patients after experiencing one molar pregnancy, and particularly those who require chemotherapy, are anxious about the problem occurring again in any subsequent pregnancy. While the data suggest that the risk of a further molar pregnancy is about 10‐fold higher than in the normal population, this only equates to an approximate 1 in 100 risk [8,9]. This risk appears to be independent of chemotherapy exposure, being similar for those patients who required chemotherapy and those where the molar pregnancy was cured by evacuation alone.

Long‐term toxicities

Although our prior analysis with 15 000 patient‐years of follow‐up suggested a slightly increased risk of second cancers in patients treated with combination chemotherapies [50], more recent data with over 30 000 patient‐years of follow‐up has revealed a different story [51]. These data show that EMA/CO if kept under 6 months in total duration carries no overall increased risk of triggering a second cancer in later life [51]. This is very important information to share with our patients. Of course, those that rarely need high‐dose chemotherapy may not be in the same position and it is also true that high‐dose chemotherapy will almost certainly destroy fertility. Fortunately, though, very few of our patients end up needing this.