Differential diagnosis of ovarian tumors in pregnant women currently remains one of the unsolved problems in obstetrics. According to some authors, ovarian tumors and tumor-like formations are detected, on average, in 2–3% of pregnant women; the malignant ovarian tumors detection rate in pregnant women is 1 per 10,000–50,000 patients [1–6].

Ovarian germ cell tumors (OGCTs) account for 20–30% of all ovarian tumors; 2–5% of OGCTs are malignant [7, 8]. The most common benign tumor is mature teratoma. The incidence of mature teratoma reaches 12% of all ovarian tumors. Rare teratomas include struma ovarii (1–2.7% of all ovarian teratomas). According to literary sources, in 5–20% of cases struma ovarii can transform into carcinoma [9–12]. Malignant germ cell tumors (malignant OGCTs) are divided into two groups based on the clinical and histological features: dysgerminomas and non-dysgerminomas (yolk sac tumor, embryonal carcinoma, choriocarcinoma, immature teratoma, polyembryoma, and combinations of all listed types). Dysgerminomas account for up to 50% of all malignant OGCTs, this morphological type has more favorable prognosis compared to non-dysgerminoma [7, 13–16].

Currently, the problem of ovarian tumors differential diagnosis is solved using comprehensive ultrasound examination at different levels of efficiency. However, ultrasound imaging is an operator-dependent technique due to subjective interpretation of results. Therefore, to standardize the ultrasound imaging results, improve the ovarian tumor diagnosis accuracy and stratify the risk of malignant OGCT in order to determine the strategy of the further patients’ management and treatment, the algorithms and complex predictive models have been constructed (RMI, IOTA (simple rules, LR1, LR2), ROMA, Kentucky morphology index (MI), O-RADS). According to ESGO guidelines (2017), in case of initial stage malignant OGCT detection, peritoneal surgical staging, and surgical removal of affected adnexa with preservation of contralateral ovary and pregnancy are recommended. However, biopsy of intact ovary and retroperitoneal lymph node dissection are considered non-viable options [17]. Cytoreductive surgery with maximum removal of tumor nodes and preservation of contralateral ovary without total lymphadenectomy is indicated to patients with advanced stage tumors. Chemotherapy is not recommended to patients with stage IA immature teratoma G1 or stage IA dysgerminoma. Chemotherapy is required in patients with advanced stage tumors. The platinum etoposide combination is considered the standard regimen (ВЕР or ЕР) [6, 18]. Organ-preserving surgery makes it possible to preserve reproductive function and is in line with treatment concept for improved quality of life. In most cases, menstrual function is restored after chemotherapy, therefore, normal pregnancy might be expected [16, 19].

The study was aimed to assess the ultrasound signs of benign and malignant OGCTs in pregnant women in order to determine the feasibility of surgical treatment and chemotherapy.

METHODS

In 2000–2020 a study was carried out involving 199 patients with histologically verified OGCTs. Inclusion criteria: consent to participate in the study; pregnancy; ultrasound findings confirming the pregravid or gravid ovarian tumor, subsequent surgical treatment and morphological verification of OGCT. Exclusion criteria: the pregnant woman’s refuse to participate in the study; threatened miscarriage; intrauterine infection; defects in a fetus diagnosed before the study.

Ultrasonographic examination was performed with the Voluson Е8 system (General Electric; USA) combined with transabdominal and transvaginal Doppler color flow mapping and pulse wave Doppler measurements. The tumor ultrasound features were assessed using the earlier proposed models [20].

In 192 pregnant women, the logistic regression models were used to assess the ultrasound signs of benign and malignant OGCTs, as well as to determine the feasibility of surgical treatment and chemotherapy. Of them, a group of 183 patients (95.3%) had benign tumors: 180 patients had mature teratomas, and three patients had struma ovarii (tab. 1).

Among them, 32 patients (17.5%) underwent surgery during pregnancy, 117 patients (63.9%) received surgical treatment during a cesarean section, and 34 patients (18.6%) received it postnatally: 11 patients (6%) 3–5 days, and 23 patients (12.6%) 3–6 months after giving birth. In most cases patients with small-sized benign OGCTs underwent resection of the ovaries, and patients with tumors larger than 10 cm in diameter underwent unilateral adnexectomy. In nine patients (4.7%), malignant OGCTs were detected during progressive pregnancy. Seven patients (77.8%) out of nine had stage I tumors, and two patients had stage III tumors. Pregnancy was preserved in all monitored patients. Six patients out of nine underwent surgery during different stages of gestation. Surgical treatment of three patients included the following procedures: cesarean section and adnexectomy with surgical staging. In four cases the tumor morphology was characteristic of pure dysgerminoma, and in one case the tumor was a combination of dysgerminoma with yolk sac tumor. Three pregnant women had immature teratomas, and one woman had a combination of immature teratoma with yolk sac tumor. Among nine patients with tumor process detected during pregnancy, three patients received drug treatment after surgery. Two patients received first-line BEP chemotherapy after giving birth because of stage III disease or yolk sac tumor, and one patient received two cycles of adjuvant carboplatin-based chemotherapy starting from the 18th week of gestation. In order to access menstrual function and fertility, medical records together with pregnancy and childbirth outcomes were studied in seven patients, who had received treatment for malignant OGCTs before pregnancy. Six out of seven patients have stage I tumors, and one patient had stage III tumor. All patients underwent organ-preserving surgical treatment (adnexectomy). In five cases, the tumor morphology was characteristic of pure dysgerminoma; one patient had immature teratoma, and another one had a yolk sac tumor. Six out of seven patients received BEP/E chemotherapy: one patient with immature teratoma G2, one patient with a yolk sac tumor, and four patients with IC stage or later stage dysgerminoma.

The surgical treatment extent in patients with malignant OGCTs was affected by each particular situation and by the urgent histological examination results (tab. 2).

Histological examination was performed by standard methods. The morphological diagnoses were provided in accordance with the WHO Classification of Tumors of Female Reproductive Organs issued in 2014 [21].

Statistical analysis was carried out using the SPSS 15 software package (IBM; USA). The differences were considered significant when p < 0.05.

RESULTS

The study showed that in pregnant women the scanograms of mature teratomas were characterized by polymorphous echographic pattern. In most cases mature teratomas were unilateral; bilateral lesion was identified in 29 patients (15.8%). The tumor diameter varied significantly in the range of 0.5–15.0 cm, and the small-sized tumors (0.5–3.0 cm) were detected in 38 cases (21%).

Based on the study results, in 75 cases (41%), the ultrasound semiotics of benign OGCTs included mixed structure with predominant cystic component. In 49 patients (27%), solid component prevailed over cystic component (cystic-solid structure). In 33 cases (18%), neoplasms were represented by solid component with clearly visible capsule. The tumors were completely anechogenic (cystic variant) only in 26 cases (14%).

The color and energy Doppler mapping revealed either single colored peripheral blood flow loci with resistance index (RI) of 0.4–0.6, or tumor avascularity. When assessing tumor vascularization using color and energy Doppler mapping, blood flow with peripheral zones of vascularization was detected in 53 patients (29%) with benign OGCTs.

The proposed logistic regression models for differential diagnosis of benign, borderline and malignant ovarian tumors in pregnant women possessed high reproducibility, sensitivity and specificity.

When diagnosing teratomas in pregnant women (tab. 3) using the model, the accuracy exceeded 90%, and the model sensitivity and specificity were 97% and 95%, respectively (fig. 1). All tumors were characterized by mixed echogenicity.

Based on the study results, the ultrasound semiotics of malignant OGCTs was represented by solid masses found in 5 patients (55.6%). Cystic and solid malignant OGCTs were detected in three cases (33.3%), and the mixed structure with predominant cystic component was revealed in one case (11.1%). In one case, the solid tumor component of higher echogenicity was represented by parenchymal papilla of more than 10 cm in diameter, and amounted to more than 80% of the total tumor volume. Hypervascularization was detected in 100% of cases. Ascitis was diagnosed in 3 patients (33.3%).

The distinctive ultrasound signs of dysgerminoma were as follows: solid (lobulated) tumor, clear margins and rough external contour, medium-level echogenicity and multiple centrally located blood flow loci. In all cases, in scanograms of pregnant women, immature teratomas were represented by hypervascular cystic and solid masses. On the ultrasound scan the yolk sac tumor, 12 cm in diameter, was represented by a solid mass with multiple cystic cavities of various size, which contained suspension. Color Doppler mapping revealed multiple loci of blood flow with low RI values. In cases of mixed germ cell tumors (dysgerminoma + yolk sac tumor, immature teratoma + yolk sac tumor) the obtained ultrasound features and Doppler measures were consistent with ultrasound signs characteristic of malignant neoplasm in 100% of patients (unclear, uneven contour, solid component amounting to 80–100% of tumor volume, hypervascularization together with high values of blood flow velocity and low values of RI, presence of ascitis).

When constructing the model for differential diagnosis of benign, borderline and malignant ovarian tumors, the Spearman's rank correlation coefficients were calculated and included in the regression model for diagnosis of borderline and malignant ovarian tumors in pregnant women (tab. 4).

When diagnosing borderline and malignant ovarian tumors in pregnant women, the model sensitivity was 100%, the specificity was 92.3%, and the overall accuracy was 92.8% (fig. 2).

Surgical treatment of patients with benign and malignant OGCTs was performed after tumor detection (see tab. 2). In the group of patients with benign OGCTs, who underwent surgery during different stages of gestation (n = 32), the emergencies in the first–second trimesters (adnexal torsion resulting from the tumor) were considered indications in four cases (12.5%); tumor size and bilateral tumor in the second trimester determined the kind of surgical treatment in 28 cases (87.5%). In 13 patients (11.1%) the presence of neoplasm was considered an indication to surgery. The other 104 pregnant women (about 89%) underwent surgery because of the combination of indications (uterine scar, placenta accreta, preeclampsia, breech presentation, hemolytic disease of the fetus, narrow pelvis, fetal malposition, and the presence of a tumor). In the group of 106 patients (90.5%) with benign OGCTs the delivery was performed at term. Preterm operative vaginal delivery due to life-threatening birth defects, placental abruption, preeclampsia not amenable to therapy, and premature discharge of amniotic fluid was performed in 11 cases (9.4%) between 25–35 weeks of gestation. All infants were born alive; the newborns’ morphofunctional characteristics corresponded to their gestational age (full-term infants with Apgar score 7–9 weighted 2,500–4,150 g; premature infants with Apgar score 3–9 and Silverman Andersen respiratory severity score 3–5 weighted 650–2,850 g). Surgical treatment in the form of resection of the ovaries was performed postpartum in patients with small-sized tumors.

The median age of nine patients at the moment of malignant OGCT detection was 26 years (20; 32).

Three patients out of nine with malignant OGCTs received drug treatment. Two patients received BEP as first-line chemotherapy after delivery. In cases of stage IA dysgerminomas only surgical treatment was performed. Combined-modality therapy (surgery + chemotherapy) was used in patients with stage III immature teratomas (G2) and mixed germ cell tumors (stages I and III). It should be noted that one of these patients received two cycles of adjuvant carboplatin-based chemotherapy during pregnancy, starting from the 18th week. The mixed ovarian germ cell tumor (combination of immature teratoma with yolk sac tumor) was diagnosed at 12–13 weeks of gestation; surgery was performed 2 days later. Surgical treatment included unilateral adnexectomy, biopsy of the contralateral ovary, omentectomy, multiple peritoneal biopsy, and peritoneal washing.

 No fetal injury was detected after the combined-modality therapy. This full-term pregnancy ended in second livebirth. The full-term newborn girl weighted 3,850 g, and was 55 cm long. She was assigned the Apgar score 8–9. The patient is under observation with no signs of disease progression. The child grows and develops normally, matching her age.

Recurrent tumor was detected in one patient, who had previously undergone surgery because of malignant OGCT at 25 weeks of gestation. Surgical treatment included unilateral adnexectomy; histologic diagnosis of mixed germ cell tumor (combination of dysgerminoma with yolk sac tumor) was established. Preterm labor occurred at 30 weeks of gestation: the preterm cesarean section was performed together with surgical removal of anterior abdominal wall tumor and ovarian ligament stump, resection of the contralateral ovary, omentectomy, paracolic gutters biopsy, surgical removal of paraaortic lymph node (based on the histological assessment results, the mixed germ cell tumor with predominant yolk sac tumor structures was diagnosed). The newborn boy weighed 1420 g and was 40 cm long. He was assigned Apgar score 6, and Silverman Andersen respiratory severity score 4.

Among nine pregnant women with malignant OGCTs detected during pregnancy, six women (66.6%) underwent cesarean section (four women because of tumors, one woman because of postterm pregnancy, and one woman because of preterm labor); three patients (33.3%) delivered per vias naturalis.

The median observation time in patients with malignant OGCTs was 66 months (12–240 months), one patient with stage III mixed germ cell tumor (combination of dysgerminoma and yolk sac tumor) died.

The further combined-modality therapy (surgery + chemotherapy) outcome monitoring in six out of seven patients with malignant OGCTs, who had received treatment before pregnancy, revealed no menstrual function disorders. The median time between the end of chemotherapy and the end of the study was 150 months (48–216 months). The first pregnancy occurred in 1–3 years (two patients), 4–6 years (two patients), and 7–10 years (one patient) after chemotherapy. Only one patient became pregnant after 12 years with the help of IVF. A total of 11 pregnancies occurred after the malignant OGCT treatment. All pregnancies ended in birth of healthy babies. During the follow-up period four patients became pregnant and delivered second time. No pregnancy and labor complications were observed in these patients.

 In five patients (71%), who had received treatment for malignant OGCTs before pregnancy, spontaneous labor at term was observed. In two cases (29%) cesarean section was performed because of uterine scar and macrosomic postterm delivery. The newborns’ condition matched their gestational age, they were assigned Apgar score 8–9.

DISCUSSION

OGCTs account for 20–30% of ovarian tumors. According to literary sources, these tumors are characterized by the greatest structural diversity. Mature teratomas account for 12% of all ovarian tumors and for up to 97% of all OGCTs [8, 22–24]. Seven types of benign and several types of malignant neoplasms have been reported [22]. In 2005, three major types of ultrasound mature teratoma appearance were distinguished: cystic type (47–60%), predominance of solid component (20–43%), and mixed structure (9–20%) [24]. A rare type of teratoma is struma ovarii, which accounts for 1–2.7% of all ovarian teratomas. In 75% of cases, ovarian strumae have a specific feature: the spongy texture area of medium-level echogenicity with single hyperechoic inclusions inside the tumor [23].

There is evidence that the ultrasound imaging sensitivity when examining teratomas is 92.3%, and the method specificity is 99.4%; when diagnosing borderline and malignant ovarian tumors in pregnant women, sensitivity and specificity are 88.9 and 98.2% respectively [25]. In our study, with the use of contemporary ultrasound imaging techniques for differential diagnosis of OGCTs, we managed to establish the correct diagnosis during the pre-operative assessment of patients in almost all cases despite the tumor structure variability. It can be concluded that tumors with increased risk of malignant transformation are characterized by solid or mixed structure with predominant solid component, as well as by hypervascularization.

Ovarian tumors contribute to pregnancy complications, as well as to high maternal and fetal morbidity. The most common obstetric complication is miscarriage (its frequency is 30–75%); miscarriage may result from underlying process, surgical treatment and chemotherapy [26–28]. According to literary sources, the incidence of ovarian tumor torsion is 6–9%, the incidence of tumor capsule rupture is 12–14.7%, and the incidence of birth canal obstructed by the tumor is 14–21% [25]. In our study, the threatened miscarriage rate was 18.2%, adnexal torsion was detected in 2.1% of patients, hematoma with abscess formation was a complication of one pregnancy, and the birth canal obstruction rate reached 8.9%. Two pregnancies ended in preterm labor 3 and 5 weeks after surgical treatment of malignant OGCTs.

Unilateral adnexectomy is the optimal primary surgical treatment regardless of the tumor stage. According to world’s literature, in case of malignant OGCT diagnosed during pregnancy, the treatment tactics are focused on prolongation of pregnancy with combined-modality therapy: surgery in the second trimester with subsequent chemotherapy or follow-up [4, 16, 29]. During our study, pregnancy was preserved in all patients, and all surgical procedures were organ-preserving (unilateral adnexectomy), which was in line with international guidelines. In a number of studies, 10 pregnant patients were monitored. Four pregnancies ended in abortions; six patients underwent adnexectomy at different stages of gestation [16, 30]. Currently, owing to high sensitivity of malignant OGCTs to chemotherapy, as well as to possibility to cure up to 80% of patients even with advanced stage tumors, the surgical intervention extent has become possible allowing us to preserve the patient’s reproductive function [16]. According to literature, chemotherapy is not indicated to patients with malignant OGCTs after surgery in case of stage IA immature teratoma GI or stage IA dysgerminoma. Chemotherapy is indicated to patients with advanced stage tumors. BEP or EP is the standard chemotherapy regimen. Three out of nine pregnant women with malignant OGCTs received BEP chemotherapy as firstline therapy after surgery: two patients received chemotherapy after giving birth, and one patient received it during pregnancy, starting from the 18th week of gestation (two cycles of adjuvant carboplatin-based chemotherapy).

It can be assumed that the most aggressive malignant neoplasms are found during pregnancy. Therefore, in case of stage II–III non-dysgerminoma detection during pregnancy, chemotherapy should be prescribed after primary surgical treatment and the implantation and organogenesis completion, without waiting for the baby to be born. Exposure to bleomycin, etoposide and cisplatin in the second and third trimesters is safe for the fetus [4]. In patients with malignant OGCTs who had received pregravid organ-preserving surgery and drug treatment, pregnancies occurred and progressed, and healthy babies were born. In patients with benign OGCTs, the absence of ultrasound signs of malignant transformation made it possible to avoid surgery and related perinatal complications during pregnancy, and to perform surgical treatment of tumors postpartum.

CONCLUSION

Comprehensive diagnosis of OGCTs, which comprised determining the ultrasound criteria of malignancy, made it possible to provide rational management of pregnancy and childbirth, surgical treatment and chemotherapy of malignant OGCTs, as well as to preserve patients’ health and reproductive function.