INTRODUCTION

The true incidence of gestational trophoblastic disease has been challenging to determine accurately, mainly due to problems with the collection and interpretation of data in a diagnosis that can occur years after an antecedent pregnancy. Gestational trophoblastic disease has been cited as occurring in 2 to 7 per 100 000 pregnancies in North America and 5 to 200 per 100 000 pregnancies in Asia.1 2

Gestational trophoblastic disease is characterized by abnormal proliferation of the placental trophoblast, comprising the spectrum of benign, non-neoplastic tumors, including both the typical and atypical placental site nodule, exaggerated placental sites, and the hydatidiform mole.

Gestational trophoblastic neoplasia is characterized by abnormal proliferation of the trophoblast that manifests as the following malignant lesions:

Choriocarcinoma

Placental site trophoblastic tumor

Epithelioid trophoblastic tumor

Invasive mole.

Gestational trophoblastic neoplasia is sometimes diagnosed without histological confirmation, most commonly where there is a persistently elevated level of human chorionic gonadotropin (hCG) after evacuation of a hydatidiform mole.

Choriocarcinoma is an epithelial malignancy arising from trophoblastic proliferation and is histologically characterized by sheets of anaplastic cytotrophoblasts and syncytiotrophoblasts without chorionic villi. Choriocarcinoma can follow virtually any gestational event but rarely follows a partial hydatidiform mole. Most choriocarcinomas following a molar pregnancy are aneuploid and almost exclusively carry paternal or androgenetic DNA.3 4 Choriocarcinomas usually originate in the uterus, but metastatic disease is common, with the lungs being the most common metastatic site. Other common sites of metastasis in choriocarcinoma are the vagina, brain, and liver. With the introduction of systemic chemotherapy active against gestational trophoblastic neoplasia and sensitive assays for the β subunit of human chorionic gonadotropin (β-hCG), mortality due to gestational trophoblastic neoplasia is unlikely, even with metastatic disease.5 6

Placental site trophoblastic tumors are a rare expression of gestational trophoblastic neoplasia, with fewer than 1000 cases ever being reported. Both placental site trophoblastic tumors and epithelioid trophoblastic tumors arise from the intermediate trophoblast. Placental site trophoblastic tumor arises from the extra-villous intermediate trophoblast and is typically a proliferation of the extra-villous intermediate trophoblast in the myometrium. In contrast to choriocarcinoma, where two populations of cells are seen, placental site trophoblastic tumors are characterized by a single population of mononuclear trophoblastic cells. Placental site trophoblastic tumors and epithelioid trophoblastic tumors do not secrete high levels of hCG, unlike choriocarcinoma and invasive moles. The hemorrhage and necrosis typical of choriocarcinoma are also absent in placental site trophoblastic tumors. These tumors can occur months to years after any pregnancy event but most commonly follow term pregnancies. Most placental site trophoblastic tumors are benign; only 30% of cases will present with metastatic disease. International Federation of Gynecology and Obstetrics (FIGO) stage IV disease and time interval ≥48 months from the antecedent pregnancy are both associated with a significant decrease in overall survival.7

Epithelioid trophoblastic tumors are another rare expression of gestational trophoblastic neoplasia. These tumors arise from the chorionic intermediate trophoblast. They are behaviorally indolent tumors that tend to present with metastases because of a lack of symptoms in early disease.

This tumor, like placental site trophoblastic tumor, does not secrete hCG. A typical presentation is that of a discrete solid-cystic nodular mass in the myometrium. Like placental site trophoblastic tumor, time from antecedent pregnancy is strongly associated with overall survival.

Invasive moles represent the ‘invasive’ form or presentation of the hydatidiform mole, a benign diagnosis. Invasive moles are characterized by edematous chorionic villi and trophoblastic proliferation. Their occurrence most commonly follows a complete molar gestation (15–20% incidence after complete molar gestation) and less commonly occurs after a partial molar gestation (1–4% incidence after a partial molar gestation).8 One or more identified risk factor in a complete hydatidiform molar gestation can increase the risk of a subsequent invasive molar gestation. These factors include uterine size larger than dates, pre-evacuation hCG levels of >100 000 IU/L, the presence of theca lutein cysts, and previous cesarean delivery.9–11 Most invasive moles present as myometrial invasion, with about 5% presenting as metastatic disease. Metastatic spread is hematogenous, and so the most likely sites are the lung and the vagina.9

INDICATIONS FOR SURGERY IN GESTATIONAL TROPHOBLASTIC NEOPLASIA

The role of surgery in the management of gestational trophoblastic neoplasia has traditionally been in the management of disease sequelae, such as bleeding, or as an adjunct to systemic therapy to remove anatomical foci of chemoresistant disease. In a review of gestational trophoblastic neoplasia in a single institution over 20 years, 82% of patients with metastatic disease and high-risk WHO scores required surgical intervention, primarily for residual disease after completing systemic treatment. The disease-free survival in these patients was reported to be 87%.12 The findings in this report reflect the body of evidence surrounding the role of surgery in gestational trophoblastic neoplasia, which is primarily an adjuvant to systemic treatment in removing resistant foci.

Another critical objective when considering surgery for gestational trophoblastic neoplasia is complete surgical resection of all gross disease. Data from a retrospective review of surgical management at a single institution over 30 years suggested that patients who had complete resection of all gross disease had better epithelioid trophoblastic tumor progression-free and overall survival. In the cohort of patients that had surgery with no gross residual disease, adjuvant chemotherapy did not confer survival benefit. The nature of the surgery, whether emergent or elective, did not affect outcomes.13 This raises interesting questions about the relationship between surgery and chemotherapy in the management of gestational trophoblastic neoplasia—for example, could there be a linear relationship between the amount of residual disease, the number of cycles of chemotherapy required to achieve no measurable disease, and progression-free and overall survival?

SURGICAL TECHNIQUES IN GESTATIONAL TROPHOBLASTIC NEOPLASIAHysterotomy

Hysterotomy is an archaic surgical approach to removing molar pregnancy from the uterus. It accesses the molar pregnancy through a laparotomy incision and a uterine incision. This approach usually necessitates surgical delivery in subsequent pregnancies due to concerns about uterine rupture. Hysterotomy also results in a lower rate of spontaneous regression than dilatation and curettage,14 but is very useful in removing isolated chemoresistant disease in women desirous of future fertility15 and may be considered part of the contemporary armamentarium in the management of gestational trophoblastic neoplasia.

We present our single-institution experience where one of the authors (AI) has managed six cases of gestational trophoblastic neoplasia with chemoresistant foci in the uterus with hysterotomy and excision of the foci to preserve fertility (unpublished). Five of the patients had invasive mole, and one had choriocarcinoma. The surgical technique is described as follows and represented in Figures 1–4. (1) A minilaparotomy is performed, and the uterine lesion is identified. (2) A dilute solution of vasopressin or epinephrine is infiltrated into the myometrium in the dilution and fashion of a myomectomy. (3) A hysterotomy is performed over the location of the resistant focus. It is extremely important that the location is correctly identified by pre-operative ultrasound and, if possible, by intra-operative ultrasound, so that a precisely located incision is made. The resistant focus is always in the myometrium, surrounded by a false capsule (probably a reaction to the prior chemotherapy). (4) The focus is then quite easily shelled out without any significant bleeding. The defect in the myometrium is then repaired as one would repair a myometrial defect in a myomectomy, and the laparotomy incision closed in the usual fashion. In our opinion, the judicious use of this procedure can avoid escalation to multiagent chemotherapy in cases of invasive moles.

Figure 1

The site of the intramural lesion is identified.

Figure 2

Vasopressin is infused into the myometrium at the site of incision as in myomectomy.

Figure 3

The lesion is usually located within a capsule and readily shelled out.

Figure 4

The excision can be accomplished quite bloodlessly and the defect is closed in the usual fashion as in myomectomy.

Dilatation and curettage

Dilatation of the cervix and suction evacuation of the uterus is the mainstay of surgical treatment of molar pregnancy. Induction of labor, with oxytocin or prostaglandin, has also been used in the treatment of molar pregnancy but is associated with higher rates of incomplete evacuation, with most cases ultimately requiring dilatation and curettage to accomplish complete evacuation.16 17 A key prognostic factor of successful evacuation of the uterus is uterine size.18 Uterine size >16 weeks has also been cited as a significant risk factor for complications, such as pulmonary tumor embolization, uterine perforation, and hemorrhage.19 20 It is advisable to start an infusion of oxytocin at the beginning of the evacuation, not before, to reduce the risk of hemorrhage, and to continue the infusion for a few more hours after the procedure. There is no need for pre-operative cervical dilatation of the cervix with prostaglandins, as the cervix is already very soft due to the high levels of hCG. While the idea of repeat dilatation and curettage to reduce the trophoblastic burden, and therefore reduce the need for chemotherapy, is attractive in theory, data from retrospective series suggest that there is little to no effect on the number of cycles of chemotherapy required, and repeat curettage is associated with a higher rate of complications like uterine perforation and hemorrhage.21 22

Hysterectomy

In women where fertility is no longer desired, hysterectomy to treat gestational trophoblastic neoplasia localized to the uterus is not an unreasonable option. Primary hysterectomy may be the best option in older patients presenting with a very enlarged uterus due to molar pregnancy. More commonly, adjuvant hysterectomy is performed for patients with chemoresistant disease in the uterus. The rate of recurrence is, as expected, lower than for dilatation and curettage: 3% in hysterectomy versus 20% for dilatation and curettage.14 While the traditional role of hysterectomy in gestational trophoblastic neoplasia is in an adjuvant setting for chemoresistant disease in the uterus, its use in a neoadjuvant setting may be a novel approach in selected cases. In some patients presenting with choriocarcinoma of the uterus, who would have required multiagent chemotherapy in view of a high WHO score, a primary, neoadjuvant hysterectomy may reduce the need for this toxic multiagent chemotherapy; instead, cure may be achieved with single-agent chemotherapy post-operatively. The following case illustrates this point:

A 36-year-old para 1 presented with persistent vaginal bleeding following her delivery 3 months previously. Clinical examination revealed a 14-week-sized uterus. The hCG level was 5 364 000 IU/L. Ultrasound and CT of the pelvis showed a bulky heterogeneous uterus with thickened endometrium (53 mm) of mixed densities (hemorrhagic and necrotic areas) within, and evidence of deep myometrial extension. No evidence of significant adenopathy or distant metastases was seen. There were no other signs of metastatic disease either clinically or radiologically. The patient had a WHO score of 8 (term antecedent pregnancy, 2; hCG level ≥100 000 IU/L, 4; primary tumor ≥5 cm; 2). The patient was advised that the standard of care would be multiagent chemotherapy. She was further advised that if she did not respond to primary chemotherapy, escalation to more toxic chemotherapy and hysterectomy would be indicated. As the patient was desirous of future fertility, she underwent a total hysterectomy with ovarian preservation. The histology of the uterus showed choriocarcinoma involving the myometrium with extensive lymphovascular space invasion. (Figure 5). Post-operatively, her hCG levels dropped to less than 50 000 IU/L 2 weeks later. She was started on single-agent actinomycin-D every 2 weeks. Her hCG level normalized after two cycles, and she was given a further two cycles. She continues to be disease-free after 8 years.

Figure 5

Gross specimen of uterus with choriocarcinoma and low power micrograph of choriocarcinoma with extensive lymphovascular space invasion.

In pre-menopausal women opting for hysterectomy, totally normal ovaries may be conserved as illustrated in the case above. If theca lutein cysts are noted during hysterectomy in pre-menopausal women, immediate surgical intervention is not necessary. Theca lutein cysts will resolve with time, lagging the fall in hCG levels. Surgery is typically only necessary for sequelae such as torsion or cyst rupture.

Hysterectomy is also the cornerstone of treatment for placental site trophoblastic tumor and epithelioid trophoblastic tumor. These two entities have low levels of hCG and are resistant to conventional chemotherapy. In cases where the disease is confined to the uterus, hysterectomy should be the primary treatment. When isolated metastases occur, especially in accessible sites such as the lungs, these should be resected.

Thoracotomy

Pulmonary surgery in the management of gestational trophoblastic neoplasia has been done to establish a diagnosis in clinically suspicious lesions, more commonly before the 1980s, and in contemporary practice, more commonly to remove chemoresistant foci.23 In a single-institution series of 122 patients treated for gestational trophoblastic neoplasia, the most important prognostic factors for patients undergoing pulmonary surgery were pre-operative hCG levels and the extent of pulmonary disease. Prognostic factors for favorable outcomes in patients who have to undergo pulmonary surgery for gestational trophoblastic neoplasia include surgical fitness, definitive control of the primary pelvic malignancy, malignancy confined to one lung, and hCG level <1500 IU/L.24

Craniotomy

Craniotomy s generally employed to control hemorrhage from metastatic gestational trophoblastic neoplasia. These tumors are very vascular, and hemorrhage is a common complication, either spontaneously or more often when chemotherapy is started. Currently, the standard of care for these high-risk patients is to start induction chemotherapy, with a lower dose to reduce the risk of catastrophic hemorrhage. The other option, in selected patients with solitary metastases in accessible brain sites, would be to remove these lesions surgically before chemotherapy.

SURGICAL OUTCOMES AND COMPLICATIONS

In general, the response to first-line chemotherapy in gestational trophoblastic neoplasia is excellent, with an overall response rate of 60–90% in low-risk disease and an overall survival approaching 100%.25 Data from a single reference center over five decades showed that 96.8% of patients with high-risk gestational trophoblastic neoplasia achieved complete remission.26 Data from the Charing Cross Hospital gestational trophoblastic disease database suggest that overall survival for patients with low-risk and high-risk disease at presentation, who subsequently relapsed was 100% (n=35), and 84% (n=25) (95% CI 66% to 96%: p<0.05), respectively. Disease that was refractory to primary treatment predicted poorer overall 5-year survival.27

As mentioned before, surgery in the management of gestational trophoblastic neoplasia has been used mainly to remove chemoresistant foci or to improve clinical outcomes in patients who have a partial or incomplete response to chemotherapy. Patients with chemoresistant disease confined to the uterus and/or lung are more likely to benefit from salvage surgery.28 The role of surgery in a neoadjuvant setting to downgrade the ‘risk’ has not been studied, but is interesting in principle; what might the effect of fairly simple surgical procedures be on the overall survival in patients with high-risk disease?

Fertility can usually be spared in gestational trophoblastic neoplasia because an excellent response to first-line chemotherapy can be expected.2 8 27

MINIMALLY INVASIVE SURGERY IN GESTATIONAL TROPHOBLASTIC NEOPLASIA

The idea of minimal access surgery is attractive in the surgical management of gestational trophoblastic neoplasia because the indications for surgery are largely confined to focal areas of primary disease or chemoresistant foci. When the surgical goal is a simple hysterectomy or wedge resection of the myometrium, minimally invasive surgery in the form of either robotic surgery or traditional laparoscopy would be appropriate. The traditionally cited benefits of minimally invasive surgery—namely, lower blood loss, shorter average length of stay, improved cosmesis, faster recovery, lower overall loss of function, and higher patient satisfaction, will apply to such surgery in the treatment of unifocal or limited gestational trophoblastic neoplasia. The primary limitation of any minimally invasive approach is that a comprehensive assessment of the peritoneal cavity is not easily or reproducibly accomplished. As surgical robots get smaller and more agile, the thoroughness of access and, therefore, assessment will continue to improve, and the impact of surgery will continue to decrease. Imaging technology and the algorithmic assessment of imaging data by deep-learning systems also promise to improve the identification of high-risk patients, who are likely to have refractory disease or relapse, enabling more aggressive treatment options upfront.29

MULTIDISCIPLINARY APPROACHES AND COLLABORATIVE CARE

The diagnosis of gestational trophoblastic neoplasia has implications across the entire spectrum of a patient’s functional capacity, life expectancy, and reproductive potential. Adopting a multidisciplinary approach to supporting the patient through diagnosis, treatment, and surveillance is considerate and good clinical practice.30 At our institution, any diagnosis of gestational trophoblastic disease or gestational trophoblastic neoplasia is managed by a multidisciplinary team within the Gynecological Cancer Program. At the time of diagnosis, women are referred for evaluation of their emotional state, their response to, and understanding of, the diagnosis and available support systems. They are longitudinally followed up by a case manager over the course of active treatment. In our experience, this has resulted in higher patient satisfaction and also improves patient compliance with treatment and surveillance regimens.

In young women who have fertility-sparing treatment for gestational trophoblastic neoplasia, a multidisciplinary approach facilitates the seamless delivery of care from cancer treatment to successful pregnancy outcomes.31 At our institution, the establishment of an onco-fertility program bringing together the specialty teams providing gynecological cancer care and reproductive medicine ensures that the women in our care do not ‘fall through the cracks’ in their care journey from one service to the next. This integrated approach also ensures a consistency in messaging when patients hear about plans for their care journey and their future functional and reproductive potential. Confidence in the care team is a cornerstone of compliance and good clinical outcomes.

EMERGING TECHNOLOGIES AND INNOVATIONS

When Vaitukaitis et al first described their assay for β-hCG in the early 1970s, a new age in the prognosis for gestational trophoblastic neoplasia dawned.6 The twin innovations of systemic agents active in gestational trophoblastic neoplasia and a sensitive assay for hCG completely changed the prognosis for women diagnosed with gestational trophoblastic neoplasia from uniformly fatal to curable. These early foundational changes continue to define the management of the disease. There are several developments that show significant promise in improving the way in which gestational trophoblastic neoplasia is diagnosed, our understanding of its malignant potential, and its treatment.

Circulating cell-free DNA is an analytical technique that allows small fragments of genetic material found in the blood to be analyzed reliably and with reproducible accuracy. Cell-free DNA from trophoblastic tumors can be detected and be a non-invasive method of early detection.32–34 The identification of circulating trophoblasts has been demonstrated and reported to be more precise in the identification of the genome as it does not have maternal DNA, which is a possible confounder.34 Having the genomic profile of gestational trophoblastic neoplasia early in the disease has the potential to significantly improve outcomes and minimize treatment-related morbidity.

Understanding the molecular bases for malignant behavior, such as metastasis and tissue invasion, can provide valuable insights into how gestational trophoblastic neoplasia spreads and causes morbidity and also provide therapeutic targets.35 36

Developments in immuno-oncology have also shown promise, especially in gestational trophoblastic neoplasia, which is refractory to traditionally used systemic agents.37–39

PSYCHOSOCIAL AND QUALITY OF LIFE CONSIDERATIONS

Patients who receive a diagnosis of gestational trophoblastic disease often report anxiety owing to the unusual nature of the diagnosis, its proximity in pathophysiology to pregnancy as well as malignancy, and the rarity of the diagnosis, reflected in the less confident way in which the diagnosis and treatment may be communicated. Moderate to severe adjustment problems are common after a diagnosis of gestational trophoblastic neoplasia, but having children or having experienced pregnancy loss tends to lessen the impact of adjustment.30 40 Therefore it is important for patients with a diagnosis of gestational trophoblastic neoplasia to be supported by a multidisciplinary team of experts with the necessary experience to provide information and sound advice on treatment options, especially if this information is communicated consistently, confidently, and with compassion. This sets the stage for patient compliance, satisfaction, and ultimately, better clinical outcomes.41

CONCLUSION

Gestational trophoblastic disease is a unique gynecologic malignancy. Its extreme chemosensitivity resulted in it being the first human solid tumor to be cured by chemotherapy, when Hertz et al cured a patient with choriocarcinoma with methotrexate in 1953.20 What was previously a uniformly fatal disease had become very curable. Since then, the primary mode of treatment of malignant gestational trophoblastic disease has been chemotherapy. The role of surgery has been mainly for treating chemoresistant disease and for complications such as hemorrhage. However, there is, now, a renewed interest in surgery. Fertility-sparing surgery is possible in selected cases and judicious use of surgery can reduce the need for the type and number of cycles of chemotherapy. A paradigm shift would be the concept of ‘neoadjuvant’ surgery in selected patients. Surgery before any chemotherapy can reduce the need for toxic combination chemotherapy. Patients can be cured with less toxic single-agent chemotherapy or need a smaller number of cycles of combination chemotherapy. The widespread adoption of minimally invasive surgery also helps to reduce surgical morbidity and improves patient acceptance. Finally, it is imperative that all patients with gestational trophoblastic disease are managed in a tertiary center with expertise in managing this rare condition. It is the only way to achieve optimal outcomes.

Ethics statementsPatient consent for publication

Not applicable.

Ethics approval

Not applicable.

Acknowledgments

The authors wish to acknowledge Associate Professor Jeffrey Low, Chief, Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, National University Hospital without whose support this would not have been possible.