Pure nongestational uterine choriocarcinoma in postmenopausal women: a case report with literature review

ABSTRACT

Nongestational choriocarcinoma is rare, especially in postmenopausal women. It may be derived from the transformation of germ cells or dedifferentiation of other tumor tissue cells. It is usually found in the ovaries but rarely in the uterus. Pure nongestational uterine choriocarcinoma in postmenopausal women is exceptional. Thirty-four cases of choriocarcinoma have been found during the past 25 y at the Tianjin Medical University General Hospital. We retrospectively reviewed the medical records of these 34 cases and found only two cases of nongestational uterine choriocarcinoma, both of which were postmenopausal women. We also reviewed 19 other cases previously reported during the past 50 y (from 1970 to 2018) that were identified as uterine choriocarcinoma in postmenopausal women. Analysis of these combined data indicates that nongestational choriocarcinoma is a rare neoplasm in postmenopausal patients. It genetically originates entirely in the patient, and short tandem repeat (STR) analyses are usually required to differentiate gestational and nongestational choriocarcinoma. Although nongestational choriocarcinoma shows some response to chemotherapy, sensitivity is much poorer than that in gestational choriocarcinoma. The prognosis for nongestational choriocarcinoma is poor and the long-term survival rate is low.

Introduction

Choriocarcinoma is a rare and highly malignant tumor originating from the trophoderm, with pathological features showing biphasic proliferation of the cytotrophoblast and syncytiotrophoblast (which produce human chorionic gonadotropin) without chorionic villi and invasion of the myometrium or blood vessels. Choriocarcinoma can metastasize to the lungs mainly through the blood, as well as to the vagina, pelvic cavity, liver, brain, gastrointestinal tract, and kidney. Lurain¹ has reported that the incidence of choriocarcinoma ranges from 1 to 9 per 40,000 pregnancies and has declined over the past 30 yin all populations. Choriocarcinoma primarily occurs in women of reproductive age and is rarely observed in the postmenopausal period.² Choriocarcinoma can be classified as either gestational or nongestational according to its pathogenetic origin. Gestational choriocarcinoma (GC) occurs in association with antecedent gestational events, including normal deliveries, abortions, ectopic pregnancies, and hydatidiform moles. Nongestational choriocarcinoma (NGC) originates from trophoblastic differentiation of germ cell tumors, endometrioid adenocarcinoma or residual germ cells.³ Primary nongestational choriocarcinoma has been described in the ovaries, but it is very rare in the uterus, especially in postmenopausal women.⁴ It is clinically important to clearly distinguish between gestational choriocarcinoma and nongestational choriocarcinoma because the prognosis and treatment of these two kinds of choriocarcinomas are different. Recently, DNA polymorphism analyses have been used in their differential diagnosis. To further understand choriocarcinoma, we retrospectively reviewed the medical records of a total of 34 cases of choriocarcinoma at the Tianjin Medical University General Hospital over the past 25 y and report herein two cases of pure nongestational uterine choriocarcinoma in Chinese postmenopausal women that were confirmed with genetic analysis.

Case report

Wang et al.⁵ had previously reported one of the cases, a 62-y-old woman, G3P2 (induced abortion 1), who was diagnosed with pure nongestational uterine choriocarcinoma through DNA polymorphism analyses. The patient was assessed with International Federation of Gynecology and Obstetrics (FIGO) stage IV cancer (with lung, spleen, bilateral adrenal gland, and brain metastases) and assigned a World Health Organization (WHO) score of 15. The patient underwent a total abdominal hysterectomy and bilateral salpingo-oophorectomy after three cycles of preoperative chemotherapy (5-fluorouracil, actinomycin D, and vincristine, FAV). Postoperation, she received a total of four cycles of chemotherapy, and she died of respiratory failure 8 mo later.

We herein report another case of pure nongestational uterine choriocarcinoma, a 63-y-old Chinese woman (G4P2) whose last menstrual period (LMP) had been 8 y prior, presented with postmenopausal bleeding in May 2017. She had used an intrauterine device (IUD) for 18 y. She had experienced two term vaginal deliveries and two abortions. The antecedent pregnancy was medical abortion, 18 y prior to presentation. Blood tests showed a low platelet count. In July 2017, the patient underwent bone marrow biopsy and was diagnosed with myelodysplastic syndrome (MDS), subtype refractory cytopenia with multilineage dysplasia (RCMD). Then, she received two cycles of chemotherapy with a decitabine regimen.

Physical examination revealed that the liver and spleen were palpable 2 cm under the rib, with no sign of ascites. No percussive pain of the liver and spleen were reported. Pelvic examination revealed that the uterus was enlarged to the size of 12-week pregnancy. A pelvic ultrasound scan revealed a bulky uterus with a 7.7 × 8.4 × 7.0 cm intrauterine mass invading the myometrium, with the IUD present (Figure 1(a)). The level of serum β-human chorionic gonadotropin (β-hCG) was 92202.96 mIU/ml. A chest radiograph and brain CT scan were negative. A chest high-resolution computed tomography (HRCT) suggested tiny metastatic lesions in the lung. A dilatation and curettage were performed, and pathology demonstrated uterine choriocarcinoma (immunohistochemical analysis of tumor cells revealed positive immunostaining for hCG and CK, Figure 1(b–d)). The FIGO stage was III: 10.

Figure 1.

(a) Pelvic ultrasound scan revealed a bulky uterus with a 7.7 × 8.4 × 7.0 cm intrauterine mass. Pathologic pictures of the intrauterine mass (b, c, and d). (b) Microscopic findings showed choriocarcinoma tumor cells comprising cytotrophoblasts and syncytiotrophoblasts without other germ cell components (× 100). (c) Positive immunohistochemical staining for human chorionic gonadotropin in the syncytiotrophoblasts (× 200). (d) Positive immunohistochemical staining for CK in the tumor cell (× 200).

The patient underwent a total abdominal hysterectomy and bilateral salpingo-oophorectomy after two cycles of EMA-CO (etoposide, methotrexate, actinomycin D, cytoxan, and oncovin) preoperative chemotherapy. A dark brown tumor with obvious necrosis and hemorrhage found in the endometrium and convex to the uterine cavity measured 8 × 6 × 6 cm in diameter (Figure 2(a)). The mass had originated from the posterior wall of the uterus. There were no specific findings in the bilateral adnexa (Figure 2(b)).The tumor largely comprised endometrial necrosis without other germ cell components and had invaded less than half of the myometrium. A small number of heterocysts with relatively consistent morphology could be seen in the center of the tumor. Immunohistochemical analysis of these heterocysts revealed positive immunostaining for placental alkaline phosphatase (PLAP) (Figure 2(c)), HNF-1β and P53 and negative immunostaining for hCG and human placental lactogen (HPL) (Figure 2(d)). The postoperative serum β-hCG level, which had decreased to 165.2 mIU/ml, returned to normal after two cycles of postoperative chemotherapy. After that, an additional three cycles of chemotherapy were given according to FIGO guidelines for the treatment of high-risk cases. The patient had a regular follow-up examination at our clinic for 11 mo without evidence of recurrence.

Figure 2.

Gross and microscopic pathological findings. (a) Gross appearance of the uterus measuring approximately 12 × 11 × 7 cm, with no specific findings in the bilateral adnexa. (b) Cut section of the uterus with stainless steel ring IUD located in the anterior wall and fundus. (c) Positive immunohistochemical staining for placental alkaline phosphatase in the tumor cell (× 100). (d) Negative immunohistochemical staining for human placental lactogen in the tumor cell (× 100).

We compared the short tandem repeat (STR) patterns of the tumor with those from blood samples of the patient and her husband to discriminate the genetic origin of the tumor. Genomic DNA was extracted from the tumor tissue specimens and peripheral blood leukocytes of the patient and her husband with informed consent. DNA was amplified by a GoldenEye DNA ID System 20A kit (Beijing, China), with 19 STR markers including D19S433, D5S818, D21S11, D18S51, D6S1043, D3S1358, D13S317, D7S820, D16S539, CSF1PO, PentaD, vWA, D8S1179, TPOX, PentaE, TH01, D12S391, D2S1338, and FGA. Electrophoresis was performed by an ABI 3130XL Genetic Analyzer platform and was genotyped automatically with GeneMapper ID-X software (version1.4; Applied Biosystems). The STR genotype of the tumor tissues matched the patient’s blood sample at all loci and differed from her husband’s sample for multiple probes, which confirmed the tumor’s nongestational origin (Figure 3).

Figure 3.

Comparative genotyping analysis of tumor DNA and peripheral blood of the patient and her husband. (a-d) Tumor tissues match the patient at all loci but differ from paternal alleles in 11informativeloci (D18S51, D6S1043, D3S1358, D13S317, D16S539, CSF1PO, D8S1179, PentaE, D12S391, D2S1338, and FGA). Y-axis: fluorescence intensity of the labeled product. X-axis: allelic sizes in base pairs, such that the number stands for repeat size in each short tandem repeat locus and can be highly variable among individuals.

Discussion

We performed a retrospective analysis on a total of 19 cases that were identified as uterine choriocarcinoma in postmenopausal women and reported from 1970 to 2018^2,5–21 (Table 1) by reviewing the English-language literature. The resultant 19 patients ranged in age from 50 to 73 y (mean, 58.4 y; median, 58 y). The duration of menopause ranged from 1 to 23 y (mean, 8.2 y; median, 6 y). Among the 19 patients, one was nulligravida, three were not described, and the latent period of the remaining 15 cases ranged from 12 to 38 y (mean, 25.4 y; median, 25 y). Most patients presented with postmenopausal bleeding. Among the 19 patients, one was diagnosed with gestational choriocarcinoma, five with nongestational choriocarcinoma and the others were not described. Only four cases were reported to have undergone molecular genetic analysis. Fifteen patients (seven stage I, one stage II, four stage III, and three stage IV) were assessed to be high risk (score≥6) according to the FIGO 2000 staging and risk factor scoring system, and 80% of them (12 of 15 cases) chose EMA-CO, which is considered to be the first-line regimen in the treatment of high-risk choriocarcinoma. In the remaining three high-risk cases, one patient (FIGO stage I:WHO score ≧8)⁷ chose monochemotherapy (MTX) because the prognostic scoring system seems to apply mainly to premenopausal women, but postmenopausal women are different; one patient (I:12)¹⁵ chose an MEA regimen because it had once been regarded as one of the most effective combination chemotherapies, with equivalent effectiveness to EMA-CO and fewer side effects; and the final patient (IV:15) was given a FAV regimen. In China, the fluorouracil-based combination chemotherapy regimen has yielded a complete remission rate of 80% for high-risk gestational trophoblastic neoplasia. Follow-up data were available in 17 patients: four patients (23.5%) were dead due to disease (DOD) at an average death time of 5.425 mo; three patients (17.6%) showed a recurrence (REC) with a mean interval of recurrence of 11.333 mo; one (5.9%) was alive with disease (AWD) at 12-mo follow-up; and nine (52.9%) were without evidence of disease (NED) at an average follow-up of 26.778 mo. Among the nine patients with stage I or II disease, one was DOD, two were REC, one was not described (ND) and all others were NED. In contrast, four of eight patients with stage III or IV disease were DOD or REC at follow-up.

Table 1.

All reported cases of pure uterine choriocarcinoma in postmenopausal women in the medical literature.

Case	Age (y)	Gravidity /parity	Menopausal duration (y)	Initial symptom	Antecedent pregnancy	Latent period (y)	2000 FIGO stage and risk factor score	Type	Surgery	First chemotherapy	DNA analysis	Follow-up status/duration (mo)
Bazinet et al.6	56	0/0	2.5	PMB	–	–	IV:ND	NGC	TAH, BSO	NO	No	DOD/1.5
Massenkeil et al.7	58	4/3	6	PMB	Normal delivery	29	I:≧8	ND	TAH, BSO	MTX	No	DOD/12
Ramondetta et al.8	60	5/5	5	Cough, hemoptysis, and weight loss	Normal delivery	38	III:13	ND	TAH,BSO, LND,OMTX	EMA-CO	No	REC/6
Baykal et al.9	54	5/4	1	PMB	Normal delivery	17	I:≧9	ND	TAH, BSO,LND	EMA-CO	No	ND
Marcu et al.10	62	4/4	10	PMB	Normal delivery	ND	ND	ND	TAH, BSO	ND	No	ND
Mukherjee et al.11	54	5/5	7	Abdominal pain	Normal delivery	25	I:≧8	ND	TAH,BSO,LND, OMTX	EMA-CO	No	REC/2
O’Neill et al.12	57	3/2	5	PMB	Abortion	22	III:10	GC	No surgery	EMA-CO	Yes	NED/>3
Yildiz et al.13	65	12/10	16	Abdominal pain and PMB	ND	25	ND	NGC	TAH, RSO	EMA-CO	No	NED/>20
Chittenden et al.14	62	2/2	14	PMB	Normal delivery	37	IV:17	ND	No surgery	EMA-CO	No	DOD/6 d
Desai et al.2	73	4/4	23	PMB	Normal delivery	38	IV:17	ND	No surgery	EMA-CO	No	AWD/>12
Hirataet al.15	58	6/3	4	PMB	ND	19	I:12	NGC	TAH, BSO	MEA	Yes	NED/>36
Samal et al.16	52	7/7	5	PMB	Normal delivery	12	I:11	ND	TAH, BSO	EMA-CO	No	NED/>2
Kaabia et al.17	63	7/7	14	PMB	ND	ND	I:<6	ND	TAH, BSO	EPA	No	NED/>60
Wang et al.5	62	3/2	10	PMB, cough, chest pain, hemoptysis, and fever	Normal delivery	36	IV:15	NGC	TAH, BSO	FAV	Yes	DOD/8
Yousefi et al.18	56	11/ND	6	PMB	Normal delivery	13	III:8	ND	No surgery	EMA-CO	No	NED/>12
Rafanan et al.19a	50	2/2	1	Abdominal pain and PMB	Normal delivery	15	I:11	ND	TAH, BSO	EMA-CO	No	REC/26
Rafanan et al.19b	51	2/2	12	PMB	Cesarean section	25	II:9	ND	TAH, BSO	EMA-CO	No	NED/>48
Guo et al.20	61	2/2	11	PMB	Normal delivery	30	III:10	ND	No surgery	EMA-CO	No	NED/>48
Wu et al.21	56	5/4	4	PMB	Abortion	ND	I:9	NGC	TAH, BSO,LND	EMA-CO	Yes	NED/>12

Choriocarcinoma is separated into two groups, gestational or nongestational, according to pathogenetic origin and genetic features. Gestational choriocarcinomas (GC) mostly occur in women of reproductive age (the average age is 27 y²²), usually over 12 mo following a preceding pregnancy. GC contains components of the paternal allele that confirm it as gestational in origin. Sixty percent of gestational trophoblastic neoplasias follow a molar pregnancy, 30% follow abortions, and 10% follow ectopic or term pregnancies.¹⁴ In contrast, nongestational choriocarcinomas (NGC) are exclusively derived from the individual in which they arise. NGC has been most commonly observed in the ovaries. Extragenital choriocarcinoma is similar to germ cell tumors and has a tendency to distribute along the central axis of the body, such as the brain-pineal gland, mediastinum, retroperitoneum, esophagus, stomach, jejunum, colon, pancreas, kidney, bladder, breast, etc. There are two types of NGC: one that arises from a germ cell tumor (such as teratoma, embryonic carcinoma, yolk sac tumor, and dysgerminoma) and the other that involves dedifferentiation of a tumor from another histologic type. NGCs are typically observed with these neoplastic components; the same is true in uterine cancer. To date, 28 cases of uterine carcinoma with trophoblastic components have been reported in the literature; the mean age of the patients was 62.8 y; most were multigravida, and the associated histologic types were endometrioid adenocarcinoma (60%), clear cell carcinoma (10%), endometrial serous carcinoma (10%), and carcinosarcoma (7%).²³ In our case, the lack of a nonchoriocarcinomatous neoplastic component with thorough sampling suggested that the tumor might not be derived from preexisting carcinoma. It is rare to observe pure NGC.

NGC has been reported in males, children, and postmenopausal women. However, the exact pathogenesis of these tumors remains unknown. There are several hypotheses that attempt to explain their origin: One hypothesis assumes that retained totipotent germ cells have an abnormal migration during embryonic development, fail to undergo apoptosis, and subsequently transform into choriocarcinoma. Another hypothesis posits dedifferentiation or metaplasia of adult tissue cells into cancer cells.^24,25

In the present case, the diagnosis of choriocarcinoma was made based on an endometrial biopsy specimen with immunohistochemistry analysis (positive for hCG and CK) combined with a high level of serum β-hCG. Furthermore, AE1/AE3 (a combination of two pan cytokeratin antibodies, AE1 and AE3), HLP, and PLAP staining are usually positive in choriocarcinomas.²⁶ OCT-3/4, CD-30 and AFP are markers of various germ cell tumors.²⁷ While the immunohistochemical staining of our patient’s postoperative specimens revealed that the heterocysts were negative for hCG and HPL, this is not the typical immunohistochemical staining pattern of choriocarcinoma. This finding may be the result of preoperative chemotherapy that induced the death of trophoblast cells. Various serum tumor markers (beta-hCG, AFP and CA-125) are also useful in the differential diagnosis of choriocarcinoma. Elevated AFP and CA-125 levels are known to be present in non-seminomatous germ cell tumors and ovarian carcinomas, respectively.²⁸ However, it is still difficult to discriminate whether a case is gestational or nongestational choriocarcinoma based on histopathologic features alone. Recently, DNA polymorphism analyses have been used to differentiate gestational (presence of both paternal and maternal alleles in equal amounts of tumor cells) from nongestational choriocarcinoma (maternal allele only).²⁹ Fisher et al.³⁰ pioneered the use of DNA analysis to identify a nongestational choriocarcinoma that was initially diagnosed as gestational choriocarcinoma on the basis of evaluation of the clinical history and presentation. In our patient’s case, the result of STR confirmed the diagnosis of NGC.

GC is considered the most curable gynecologic cancer with an overall survival rate from 82% to 100%, even in the presence of metastatic disease.³¹ Chemotherapy is the standard treatment.³² The management and treatment of choriocarcinoma are based on recommendations set forth by the FIGO. According to the FIGO, Gestational trophoblastic neoplasia can be divided into two groups: the low-risk group and the high-risk group. For low-risk groups, single-agent chemotherapy regimens can be used because the side effects are relatively minor. There are two main first-line chemotherapy options for the treatment of low-risk groups including actinomycin D and methotrexate. Comparing these regimens, Cochrane et al.³³ have reported that actinomycin D was associated with significantly higher rates of primary cure than methotrexate (five studies, 513 women; RR 0.64, 95% confidence interval (CI) 0.54 to 0.76). Multidrug therapy is the gold standard first-line therapy in metastatic forms of high-risk choriocarcinoma. Currently, EMA-CO and fluorouracil-based combination regimens are highly recommended for the treatment of high-risk groups. For the past two decades, the EMA-CO regimen has become the most widely accepted therapeutic regimen, positive responses to which have been obtained in 80% to 85% of patients, with an overall survival rate of approximately 100%.³⁴

In the present circumstances, the management and treatment of GC can be referred to the FIGO. However, the prognostic scoring system for malignant gestational trophoblastic disease seems not to apply to postmenopausal women. It stresses the interval from the last pregnancy to the development of disease as a prognostic factor. These prognostic differences are measured in months, yet, in postmenopausal women, the antecedent pregnancy took place many years before the choriocarcinoma occurred. In contrast to GC, most cases of NGC are discovered in the advanced stage, and the disease tends to progress rapidly and be widely metastatic in the short term. NGC has been associated with resistance to chemotherapy and a reduced survival rate.^35,36 Mello et al.²⁶ found that NGC had resistance to the first line of chemotherapy protocol (methotrexate and folinic acid) but exhibited a positive response to the second line of treatment using actinomycin-D at the 60-mo follow-up. Currently, there is no treatment standard for NGC, but instead, treatment generally follows that of GC. Surgical treatment is an important means of treatment for NGC because the tumor is exclusively derived from the individual in whom arose. However, considering the large tumor sizes, wide invasion sites and easy blood metastasis of NGC, radical surgery is usually performed after chemotherapy has reduced blood hCG to normal or near-normal levels to relieve the tumor burden, which can shorten the course of treatment and reduce the recurrence rate. For patients with NGC especially those with other germ cell tumors, BEP (bleomycin, etoposide, and cisplatin) or PVB (cisplatin, vincristine, and bleomycin) regimens are often used. A more effective combination of chemotherapy should be taken into account and applied to the treatment of extragenital NGC while simultaneously considering the characteristics of the origin site and the components of other tumors. EMA-CO, VIP (etoposide, ifosfamide, and cisplatin), or 5-Fu+KSM+ etoposide regimens may be selected. An EMA-CO regimen was selected in our case, considering the FIGO stage and WHO score. Our patient received two courses of preoperative chemotherapy to reduce the tumor burden, then underwent radical surgery, and finally received consolidated chemotherapy for three courses to reduce recurrence.

Conclusion

This case report describes a pure nongestational uterine choriocarcinoma in a postmenopausal woman that was confirmed by genetic analysis. The diagnosis of choriocarcinoma in women over the age of 50 is very rare, especially nongestational uterine choriocarcinoma. Practitioners should consider choriocarcinoma as a differential diagnosis when evaluating postmenopausal women with elevated serum β-hCG levels. It is difficult to identify the source of choriocarcinoma only through clinical presentation, pathomorphological analysis, and biochemical tests. Due to the resistance to chemotherapy and low survival rate of NGC, STR analyses should be applied to differentiate gestational from nongestational choriocarcinoma.

Funding Statement

This work was supported by the National Natural Science Foundation of China [81472761];Natural Science Foundation of Tianjin City [14JCYBJC25300].

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.