The classic sonographic appearance of CHM described above may be lacking in patients who present with bleeding early in the first trimester, making the sonographic diagnosis more difficult. The hydropic villi are smaller at earlier gestational ages and the molar tissue may appear as a complex echogenic intrauterine mass with several anechoic or cystic spaces [10, 13] (Fig. 18.3). In early gestations it may also be more difficult to differentiate PHM from CHM. One characteristic that may help to differentiate the two types of molar pregnancy is the presence of a gestational sac in PHM that is either empty or contains small fetal echoes surrounded by a large rim of placental echoes with cystic spaces [14]. Additionally, the earlier the presentation the more difficult it is to differentiate molar pregnancy from a hydropic nonmolar abortion. Ultrasonic descriptions of early histologically confirmed molar pregnancies include: an empty gestational sac or intrauterine anechoic fluid collection, fluid collection in association with an echogenic mass, thickened endometrium and echogenic fluid-filled levels within the endometrium [15, 16].

Ultrasound can also be used to assess the volume of molar tissue present in the uterus which has been shown to be a risk factor for GTN [17–20]. A three-dimensional assessment of volume is felt by some experts to be superior to a single measurement. However, in one study the size of the lesion measured sonographically was not predictive of need for chemotherapy [18]. Although myometrial invasion can be assessed with ultrasound, MRI is often a better modality to demonstrate the presence of invasion of molar tissues into the uterine wall [19, 20].

Several investigators have examined the accuracy of ultrasound in diagnosis of molar pregnancy. Fowler et al. [21] published the largest study evaluating the role of ultrasound in detection of molar pregnancy. The authors reviewed 1053 consecutive cases of molar pregnancy with early ultrasound evaluation and found that the sensitivity for detecting either a CHM or PHM was 44 %. The mean gestational age at diagnosis was 10 weeks. They found that ultrasound was better at detecting CHM versus PHM (79 % vs. 29 %, p < 0.0001) and there was a false-positive rate of 10 % whose final pathology demonstrated hydropic degeneration of nonmolar pregnancies. Other smaller reviews have shown sensitivity of ultrasound in detecting molar pregnancy to range from 34 to 57 % [15, 16, 21, 22]. Correlation with hCG is critical to the diagnosis. It has been shown that the ultrasonographic diagnosis of complete molar pregnancy is facilitated by clinical factors such as hCG [23]. When analyzing ultrasound sensitivity by type of molar pregnancy, ultrasound is consistently more sensitive in detecting CHM (sensitivity range 58–95 %) [13, 16, 21–24] as compared to PHM (sensitivity range 17–29 %) [16, 21, 22]. The sensitivity of ultrasound improves with increasing gestational age as the hydropic villi grow in size with advancing gestational age and are more easily seen by ultrasound.

There are other conditions in early pregnancy that mimic the sonographic appearance of molar pregnancy such as hydropic degeneration of the placenta, missed abortion, blighted ovum and retained products of conception. False-positive rates of ultrasound in molar pregnancy have been estimated anywhere from 4 to 10 % [21, 25]. Hydropic changes of the placenta in other gestations can appear similar to the hydropic villi of molar disease but tend to be less homogeneously distributed [21, 26]. Although it can be difficult to differentiate an early partial molar pregnancy from other abnormalities of early pregnancy, the presence of an echogenic rim around the sac may be more indicative of a missed abortion or blighted ovum [14].

Color Doppler ultrasound allows the clinician to assess vascularity. In trophoblastic tissues a high velocity, low resistance flow is consistent with increased vascularity. Thus, use of color Doppler imaging may help differentiate molar pregnancy from mimics such as missed abortion [10, 27].

Several studies have shown that uterine artery resistive indices correlate well with hCG levels and may therefore be helpful in diagnosis and monitoring of treatment [28–33]. Yalcin et al. followed 21 patients with molar pregnancy with hCG and uterine artery Dopplers and found significant negative correlations with hCG and Doppler indices, i.e., as the hCG declined demonstrating resolution of disease, the resistive indices rose [31]. In this study, the patients who ultimately developed postmolar GTN had significantly lower Doppler indices than those whose disease regressed spontaneously. Others have also demonstrated this inverse relationship of lower resistive indices indicating a greater risk of developing postmolar GTN [29, 30, 34–36]. Finally, one study of 25 patients with molar pregnancy undergoing surveillance with hCG and transvaginal ultrasound (TVUS) with Doppler also noted that surveillance with Doppler predicted GTN 1–3 weeks before routine hCG monitoring and that ultrasound findings tended to resolve about 8 weeks earlier than hCG normalization [32].

Ovarian theca lutein cysts develop in 25–65 % of complete molar pregnancies in association with markedly elevated hCG levels >100,000 mIU/ml. In a review of 386 patients with untreated hydatidiform mole, 102 patients (26.4 %) had concurrent theca lutein cysts [37]. Theca lutein cysts appear sonographically to be anechoic, multi-loculated ovarian cysts [38] (Fig. 18.4). Mean cyst diameters are reported around 7 cm but can range in size from 3 to 20 cm [26, 38]. Typically these cysts will be accompanied with the uterine sonographic findings detailed previously. Of note, theca lutein cysts are less likely to be seen in the first trimester complete or partial moles pregnancies where the hCG level is generally <100,000 mIU/ml [15]. The size of the theca lutein cysts has not been shown to correlate with persistent disease, although the presence of bilateral cysts is associated with an increased risk of GTN [37]. Serial ultrasound examination is useful to monitor the regression of theca lutein cysts which tend to regress slowly over 2–4 months following molar evacuation as the hCG level declines [39]. Although theca lutein cysts rarely rupture spontaneously or undergo torsion, prompt laparoscopic intervention can be used effectively. The use of ultrasound guidance during percutaneous drainage of massively enlarged theca lutein cysts may also provide considerable relief of abdominal discomfort.

Concurrent twin pregnancy with a hydatidiform mole and coexisting fetus is estimated to occur 1 in 22,000–100,000 pregnancies [40]. The diagnosis of a molar pregnancy with coexisting fetus is almost always made based on ultrasound findings and tends to be diagnosed at later gestational age than a singleton CHM [41]. Ultrasound findings show a live fetus with either a single enlarged placenta with the classic cystic changes and increased echoes or there may be two placentas, one normal and one molar [42] (Fig. 18.5). Ultrasound is also critical in the ongoing management of these pregnancies to help ensure the well-being of the coexisting fetus. Of those pregnancies described in the literature that were continued after diagnosis of a twin molar gestation, more than half continued beyond the 28th week of gestation with almost 70 % of children surviving [40]. The rate of GTN was not significantly different between the group who chose to continue the pregnancy and those that interrupted the pregnancy at diagnosis. However, these twin molar pregnancies are more likely to develop GTN as compared to singleton molar pregnancies [41].

The first step in management of a molar pregnancy is uterine evacuation, typically with suction curettage [43]. The technique for uterine evacuation is similar to that used for spontaneous and induced abortions. However, there is a greater concern for blood loss due to the increased vascularity of molar gestations. Intraoperative ultrasound may be a very useful tool during a suction curettage for molar pregnancy, particularly in cases with a large volume of intrauterine disease. When available, ultrasound should be used at the start of the procedure to examine the intrauterine disease and assess for pelvic extension. Ultrasound can be utilized to help prevent uterine perforation during the serial dilation of the cervix and placement of the suction curette. Finally, at the conclusion of the procedure, ultrasound allows the clinician to visualize the uterine cavity and confirm complete evacuation of molar tissues [43].

Gestational trophoblastic neoplasia (GTN) includes invasive mole, choriocarcinoma, PSTT, and ETT. Choriocarcinoma arises from both cyto- and syncytio-trophoblast and produces high levels of hCG. Choriocarcinoma is associated with early metastatic spread but is generally highly sensitive to chemotherapy. Unlike choriocarcinoma, PSTT and ETT arise from extravillous intermediate trophoblasts and produce low levels of hCG. Unfortunately and importantly, both PSTT and ETT are relatively resistant to chemotherapy unlike the other types of GTN. Ultrasound is the imaging modality of choice for the initial evaluation of the uterus and adnexa when a patient has been diagnosed with GTN. Ultrasound is not able to differentiate between types of GTN. Therefore, correlation with clinical history and hCG is critical. Betel et al. compared 17 cases of GTD to 14 cases of retained products of conception sonographically and found that GTD cases were more likely associated with a larger mass (>3.45 cm), thin endometrium (<12 mm), myometrial based mass and vascular lakes [44]. Non-gestational conditions that have been described to mimic GTD sonographically include uterine leiomyomas and an adenomyomatous polyp [26, 45]. In addition to the increased vascularity, Doppler can also illustrate focal areas of increased flow within the myometrium in cases of invasive molar pregnancy and choriocarcinoma [46]. Measurement of uterine artery Doppler indices including the resistive index (RI) and the pulsatility index (PI) has been studied in GTN. RI and PI tend to be very low in GTN and changes within these indices correlate with response to treatment and resolution of GTN over the course of follow-up with the indices increasing as the hCG levels decline [47, 48]. Nonetheless, some sonographic findings may be clues to the diagnosis.