Prenatal imaging and pathology of placental mesenchymal dysplasia: a report of three cases

Introduction

Placental mesenchymal dysplasia (PMD) is a rare vascular anomaly, which is estimated to occur in 0.02% of pregnancies, characterized by mesenchymal stem villous hyperplasia. The gross features of PMD include placentomegaly, dilation of chorionic vessels, and large areas of cystic villous change along with a normal placenta. The microscopic features include mesenchymal dysplasia with prominent thick-walled vessels, aneurysmal dilation of stem villous vessels, and possible chorangiomatoid changes. In the first trimester, ultrasonic findings of PMD are similar to those of molar pregnancy, such as enlarged placentas with multicystic and anechoic lesions. Although a normal fetus is typically present in patients with PMD, unlike molar pregnancies, it is important to distinguish PMD from partial hydatidiform mole or a twin pregnancy with a normal fetus and coexistent complete mole [1, 2].

Because various maternal and fetal complications may occur during pregnancy with PMD, strict clinical management, including evaluation of fetal well-being, fetal biometry, and fetal Doppler, such as umbilical artery, middle cerebral artery, and ductus venous is crucial. Approximately 20% of patients with PMD have Beckwith-Wiedemann syndrome (BWS), which exhibits several unique features, including macrosomia, exomphalos, macroglossia, omphalocele, craniofacial features, and ear anomalies. In patients with PMD without BWS, the rates of fetal growth restriction (FGR) and intrauterine fetal death (IUFD) are approximately 50% and 40%, respectively [1]. The rate of preterm birth reaches 53% (including spontaneous preterm birth, 33%), and the rates of preterm premature rupture of the membrane and artificial termination of pregnancy are 17% and 19%, respectively. Further, the incidence of maternal complications such as gestational hypertension, preeclampsia, and eclampsia is 9%. Neonatal or maternal complications are absent in only 9% patients with PMD [3]. Here we report three cases with PMD, focusing on imaging and pathological findings.

Presentation of the case

Three Japanese women with PMD were treated at our hospital between 2009 and 2013 (we treated 4210 pregnancies). All pregnancies were natural, and they were suspected as molar pregnancies at 10–13 weeks of gestation due to intrauterine cystic sonographic findings (Figure 1A, D). All patients had 46 somatic and XX chromosomes. The total serum hCG levels of patients 1 and 3 were higher than the normal range for gestational age [4]. Labor was induced in patient 1 at 36 weeks of gestation due to oligohydramnios and FGR. She delivered a 2010 g female with Apgar scores of 8 and 9 (at 1 min and 5 min, respectively). Patients 2 and 3 were diagnosed with IUFD at 28 and 24 weeks of gestation, respectively, despite normal fetal growth. The patients did not exhibit characteristic findings of BWS or fetal anomalies. The clinical features of these patients are summarized in Table 1.

Figure 1:

Placental imaging and cross sections of the placentas of patients 1 and 2.

Ultrasound imaging shows a first trimester, enlarged placenta totally occupied by multicystic areas (A, D). In patient 1, the placental vesicular lesions reduced gradually after 30 weeks of gestation (B). In patient 2, the entire placental area was filled with the cystic areas (E). In case 1, multicystic lesions accounted for 40–50% of the placental parenchyma (C), whereas in case 2, the thickened placenta was completely covered with multicystic lesions (F). The dilated chorionic vessels detected on the fetal side of the enlarged placenta are indicated by short arrows.

Imaging of patient 1 during the first trimester revealed an enlarged and thickened placenta that was fully occupied with multicystic areas. Serial imaging revealed a gradual size reduction of the multicystic areas after 30 weeks of gestation. Moreover, magnetic resonance imaging (MRI) at 31 weeks of gestation showed a decrease in multicystic areas in the enlarged and thickened placenta (Figure 1B). In contrast, most of the placentas of patients 2 and 3 were filled with the cystic areas throughout the pregnancy. MRI imaging of patient 2 at 23 weeks of gestation showed that the entire placenta area was occupied by multicystic formations (Figure 1E).

The placentas of each patient were enlarged and thickened with dilated chorionic vessels. The placental weights were greater than the 95^th percentile, and those of patients 2 and 3 were heavier than that of patient 1 compared with the average week of gestation [6]. Cross sections of the placentas showed normal spongy parenchyma mixed with multicystic areas. The cystic lesions of patient 1 accounted for 40–50% of the placental parenchyma (Figure 1C). In contrast, cystic lesions were present throughout the placentas of patients 2 and 3 (Figure 1F). There were aneurismal and varicose dilations of chorionic vessels in the fetal growth plates of all placentas (Figure 1C). Each umbilical cord was inserted close to the middle and had two arteries. Rupture or thrombosis of these vessels was not detected.

Microscopy of the cystic lesions of each patient showed focally dilated intermediate stem villi with fibroblastic stroma and increased vascularization or formation of sinusoids (Figure 2A, D, and G). Some vascular walls in the stem villi were thickened, and chorionic plate vessels were dilated (Figure 2D). Villous chorangiosis, defined as the presence of at least 10 capillaries in at least 10 terminal villi in at least three low-power fields, was identified in several areas (Figure 2G and H). The vacuoles (cisternae), which were similar to lymphatic spaces, were detected in abnormal villi (Figure 2A).

Figure 2:

HE staining and immunohistological analysis of CD34 and D2-40 expression in PMD tissues.

The images show three patterns of staining of tissues from patients 1, 2, and 3 as follows: HE (A, D, and G), CD34 (B, E, and H; 1/100 dilution; QBEnd/10; Leica, Newcastle, UK), and D2-40 (C, F, and I; 1/1; Nichirei, Tokyo, Japan). Fetal vascular endothelial cells expressed CD34 but CD34 expression was undetectable around the cisternae (B, E). The cisternae were surrounded by a D2-40-positive single-cell layer containing a cellular matrix, and the chorionic pericellular/perivascular extracellular matrix was D2-40 positive, but the vascular wall was negative (C, F, and I). Villous chorangiosis is shown in panels G and H. (Patient 1 original magnifications, 200× and 40×) (Patients 2 and 3, original magnification: 100× and 40.) Long arrow, vessel in abnormal villi; asterisk, cistern; short arrow, thickened vascular wall.

In low-power fields, D2-40 expression was particularly positive in the stroma of abnormal villi (Figure 2F, I). Observations of high-power fields revealed that the cisternae were surrounded by a D2-40-positive single-cell layer containing a cellular matrix and that the chorionic pericellular/perivascular extracellular matrix was D2-40-positive (Figure 2C). Immunostaining using an anti-CD34 antibody readily detected multiple small vessels in the abnormal villi, and the lining of the cisternae did not originate from the endothelium (Figure 2B, C). Immunostaining using an anti-Ki-67 antibody did not detect excessive proliferation of trophoblasts in the placenta of the PMD tissue or normal placenta (Figure 3C, G), while increased proliferation was detected in partial and complete moles (Figure 3K, O). The cyclin-dependent kinase inhibitor 1C (P57, KIP2) is implicated in human imprinting disorders such as hydatidiform mole and BWS. Immunostaining using the antibody against p57^kip2 detected cytotrophoblasts and stromal cells in normal placenta; however, most abnormal villi in PMD tissue comprised p57^kip2-negative stromal cells, although positive cells were detected in some stem villi. Further, cytotrophoblasts expressed p57^kip2 (Figure 3D, H, L, P).

Figure 3:

Immunohistochemical analysis of CK7, Ki-67, and p57^kip2 expression in patient 1 (normal placenta at 37 weeks), partial mole, and complete mole (magnification: 100×).

The analyses employed the antibodies as follows: anti-CK7 mouse monoclonal antibody (1/100 dilution; OV-TL 12/30; DAKO, Glostrup, Denmark), anti-Ki-67 mouse monoclonal antibody (1/50 dilution; Mib-1; DAKO), and an anti-p57^kip2 mouse monoclonal antibody (1/200 dilution; KP39; Thermo Fisher, CA, USA). CK7 expression was detected in cells or the matrix that form the cisternae, which were not of trophoblast origin (A and B). Analysis of Ki-67 expression showed no excessive proliferation of trophoblasts in the PMD tissue or normal placentas (C and G). In contrast, there was excessive proliferation of trophoblasts in partial and complete moles (K and O). Cytotrophoblasts and stromal cells were partially positive for p57 ^kip2 expression in normal placenta (H) and partial mole (L). Some abnormal villi composed of stromal cells were negative for p57^kip2 expression in PMD tissue (D), and the complete mole was negative.

Discussion

Complications associated with PMD, such as FGR and IUFD, may be explained by potentially chorionic hypoxia secondary to obstructive fetal vascular thrombosis and a decrease in maternal-fetal gas exchange caused by an insufficient number of normal chorionic villi [1]. Because the fetus develops an increasing demand for oxygen and nutrition during gestation, a related placental dysfunction may be more closely associated with fetal death in the last stage of pregnancy rather than at an earlier stage. Fetal outcomes of patients with PMD are predicted by the size and growth of the remaining functional normal placenta, not necessarily by the extent of abnormal cystic regions in PMD tissue [2].

Consistent with these reports, patient 1, in whom the multicystic areas of the placenta decreased in size as pregnancy progressed, gave birth to a living child. In contrast, stillbirths were the outcomes for patients 2 and 3, in whom the placentas were completely covered with multicystic tissue, and normal areas were not consistently detected from the first trimester. These outcomes were consistent with their respective pathological features.

In contrast, another study did not identify a relationship between placental weight and fetal or maternal outcome in 21 patients with PMD who delivered during the second half of pregnancy [7]. Moreover, there are favorable outcomes in patients with PMD with heavy placentas, although there were unfavorable outcomes with certain amounts of remaining normal placenta. One may argue that anomalies of the umbilical cord, including complete obstruction, rupture of an aneurysmal dilation, velamentous insertion, and hypercoiling may cause hypoxia or hypocirculation in the placenta [1].

The prediction and evaluation of fetal outcome in patients with PMD remains to be determined. Therefore, further efforts are required to develop improved diagnostic techniques. For example, careful follow-up of normal placental growth using imaging studies may improve the prediction of fetal outcomes of patients with PMD [2]. Using ultrasonography, it is difficult to view the entire placenta, and MRI may provide an alternative imaging modality for evaluating intrauterine formations. Moreover, ultrasonography or MRI can be used to measure placental blood flow and volume during fetal growth, and changes in fetal and placental oxygenation can be measured using blood oxygen level-dependent MRI [8].

D2-40 is used as a lymphatic endothelial marker because D2-40 reacts with an O-linked sialoglycoprotein present in the lymphatic endothelium but not with the blood-vessel endothelium. In the present study, we detected expression of CD34 by fetal vascular endothelial cells but not in the lining of the cisternae [9]. Although it is sometimes difficult to distinguish fetal vessels and cisternae using hematoxylin-eosin (HE) staining, immunohistochemical assays of CD34 and D2-40 expression may serve this purpose.

The reciprocal staining pattern using these antibodies may be unique to PMD. These findings further support the hypothesis that cells lining the cisternae are not of endothelial origin. Immunohistochemical analysis of PMD using an antibody against p57^kip2 detected positive stromal cells, which indicates biparental villi. Moreover, the absence of p57^kip2-positive stromal cells in abnormal villi suggests that the villi of affected lesions were androgenetic/biparental mosaics [10]. Although there was marked enlargement and hypervascularity of the stem villi, these villi proliferated at a high rate, which was demonstrated by lack of detection of Ki-67 expression.

In conclusion, if PMD is suspected antenatally then regular assessment of fetal growth, amniotic fluid volume, and Doppler should be instituted. The images presented here could suggest a role for placental MRI in evaluating PMD and one should try to assess the amount of “normal” placental morphology remaining. If clinicians notice a decrease in the percentage of normal placental areas in patients with PMD as the pregnancy progresses or find anomalies of the umbilical cord, the possibility of FGR and IUFD should always be considered.