sequences as needed, including fat-suppressed and opposed-phase imaging if a fat containing lesion is suspected, and time-of-flight imaging if further evaluation of a vascular structure is indicated.

Parallel imaging reconstruction algorithms GRAPPA with iPAT factor 2 are used to decrease the MR data acquisition time of the sequences therefore reducing fetal and maternal motion artifacts.

To minimize the deposition of radiofrequency energy in the pregnant patient and optimize temporal resolution, a 256 × 224 matrix is used with a partial-phase field of view of 0.75 in applicable rectangular geometries, such as the axial plane.

An attempt is made to confirm all suspected abnormalities in at least two imaging planes because the normal curvature of the uterus can potentially lead to a false-positive examination in a single imaging plane. When higher-resolu- tion imaging is required to maintain a satisfactory signal-to-noise ratio, additional images can be obtained in the desired plane using a T2-weighted fast spin-echo sequence. This sequence can be performed over a limited area during a breath-hold using some type of flip back pulse to shorten the repetition and acquisition times.

The use of fat suppression in conjunction with T1-weighted sequences improves the conspicuity of blood products.

Some investigators have advocated the use of gadolinium-based contrast agents to improve the specificity of MRI for diagnosis of placenta accreta by better defining the outer placental surface and myometrium and distinguishing placenta accreta from percreta (Palacios Jaraquemada and Bruno 2000; Tanaka et al. 2001).

Therefore, in clinical practices, gadoliniumbased contrast agents are not used in pregnancy, except when the potential risks to the patient are outweighed by the potential benefits of contrastenhanced imaging.

Clinical experience with diffusion-weighted placental imaging is likewise limited (Bonel et al. 2010; Morita et al. 2009), but this sequence has been recently demonstrated to be very useful in the detection of placental hematoma (Masselli et al. 2011b).

4\ Normal Appearance

Before interpreting images for pathologic findings, it is necessary to understand the normal anatomy and the normal findings of the placenta at multiplanar MR imaging (Nguyen et al. 2012).

The gravid uterus should be pear shaped, with the fundus and body being wider than the lower uterine segment. The uterine contour is usually smooth, and focal bulging should not be present.

Typically, the placenta is located along the anterior or posterior uterine wall, extending onto the lateral walls.

Placental size is expressed in terms of thickness in the midportion of the organ and should be between 2 and 4 cm. Placental thinning has been described in systemic vascular and hematologic diseases that result in microinfarctions. Thicker placentas (>4 cm) are seen in fetal hydrops, antepartum infections, maternal ­diabetes, and ­maternal anemia. Placental thickening can be simulated by myometrial contractions and underlying fibroids (Victoria et al. 2011).

The MR signal of the placenta varies according to the utilized imaging sequence (Leyendecker et al. 2012; Levine and Pedrosa 2005). With the most commonly utilized fetal pulse sequence, HASTE, the placenta demonstrates intermediate signal, hypoor isointense with respect to the surrounding myometrium. A fine line of separation between the myometrium and the placenta may be visualized, most likely representing the placental-myometrial interface. The placenta is predominantly homogeneous in signal in the early second trimester and has a relatively flat and smooth surface (Fig. 1).

On steady-state free-procession or true FISP images, the placenta is hypoto isointense with respect to the myometrium and homogeneous in appearance during the second trimester, becoming more heterogeneous as maturation occurs. The placental-myometrial interface may be seen, although it was less distinct than in HASTE images.

On T1 FLASH images, the placenta demonstrates a homogeneous signal, isointense to myometrium.

As the placenta matures, particularly in the third trimester, cotyledons become easier to discern as round, high-signal structures seen in fluid-sensitive sequences, delineated by a subtle peripheral low signal line, likely representing the normal placental septa (Fig. 2). The placenta also

becomes more complex appearing, with gentle lobulations seen on its fetal surface and fine vascular channels becoming more distinct as they traverse the placental tissue.

Placental septa and the cotyledons are more often seen when imaging with a 3 T system (Fig. 3).

The normal subplacental vascularity can be seen as numerous flow voids just under the placenta. A few flow voids can also be seen within the placenta and are usually in the region of the insertion point of the umbilical cord.

The myometrium has a variable thickness and thins as the pregnancy progresses.

It can be seen as three distinct layers of signal intensity; the inner and outer layers of the myometrium are seen as thin bands of decreased T2 signal intensity (Fig. 4). The middle layer is thicker, has intermediate T2 signal intensity, and often contains multiple flow voids representing the normal myometrial vascularity.

As the pregnancy progresses, the myometrium can become quite thin and should be visualized as a continuous band of soft tissue low intensity signal surrounding the placenta (Fig. 5).

However, the myometrium may blend into the placenta, and it can be difficult to visualize even at technically adequate examinations in patients with prior cesarean section.

One problem with MR imaging of placenta adhesive disorders is that distinction between the myometrium and the placenta can be difficult on the types of sequences typically used (Kim and Narra 2004; Lax et al. 2007).

If placenta accreta is suspected, additional imaging planes are chosen that best show the placenta-myometrium interface in the region of suspected abnormality or other structures of interest, such as the bladder dome. Such imaging is typically best accomplished in an angled scan plane perpendicular to the placenta-myometrium interface or myometrium-bladder interface (Masselli et al. 2008).

4.1\ Placenta Variants

Most placentas are round or discoid in shape, but other shapes should be described when present;

Fig. 1  Normal placenta in this 26-week-old fetus. (a) Sagittal T2-weighted half-Fourier RARE T2-weighted MR image shows a placenta (P) with intermediate signal intensity. (b) Sagittal T1-weighted fat saturation sequence

demonstrates a homogeneous placenta, isointense to myometrium. (c) Sagittal DWI sequence (B = 800) and ADC map (d) demonstrates a homogeneous placenta

Fig. 2  Coronal T2-weighted half-Fourier RARE (a) and true FISP (b) MR images show a homogeneous placenta with thin linear areas of decreased signal intensity in a

Fig. 3  Coronal T2-weighted half-Fourier RARE at 3 Tesla MR scanner shows the cotyledon structure of the placenta (arrows)

regular pattern (arrows) representing normal placental septa

Fig. 4  Axial T2-weighted MR image of healthy secondtrimester placenta shows the three layers of the normal myometrium. The hypointense outer (short arrows) and inner (arrows) layers surround the more hyperintense middle layer, which contains the vasculature

Fig. 5  Axial T2-weighted MR image of healthy thirdtrimester placenta shows the placenta has homogeneous intermediate signal intensity and the myometrium is visible as a low-signal-intensity line external to the placenta (arrows)

Fig.6  Sagittal T2-weighted half-Fourier RARE MR image shows a normal placenta with a succenturiate lobe. The main body of the placenta is located along the posterior uterine wall (arrow). A second soft tissue structure with similar signal intensity is seen along the anterior uterine wall and represents the succenturiate lobe (small arrow)

variant placental shapes include bilobed, succenturiate, circumvallate, and placenta membranacea (Huppertz 2008).

Usually discoid in shape, the placenta can exhibit various morphologies. The placenta can have a separate lobule that is not contiguous with the main placental body, which is called a succenturiate placenta (Fig. 6) (Elsayes et al. 2009).

A small lobule of placenta separate from the main bulk of the placenta is referred to as a succenturiate lobe. This is important to describe when present because of the risk of connecting vessel rupture or retention of the lobe at delivery, both potentially resulting in significant hemorrhage (Bernirschke and Kaufmann 2000).

If there are two lobes of placenta, similar in size, the placenta can be described as bilobed. Circumvallate placenta is best described as having a rolled-up edge. In a retrospective review of 7666 deliveries, the odds ratio of placental abruption­ in patients with circumvallate placenta was 13.10 (95% confidence limits: 5.65–30.20) (Elsayes et al. 2009). A placenta membranacea or “placenta diffusa” occurs when villous atrophy fails to occur early in gestation. As a result, fetal membranes remain covered with chorionic villi. This rare entity presents with a thinned diffuse placenta covering the uterine cavity, and is associated with placental invasion (Linduska et al. 2009). Annular placenta may be a variant of placenta membranacea, presents with a ring-shaped placenta, and has similar risks of hemorrhage and growth restriction (Derwig et al. 2011).

Lastly, in placenta fenestrata, the placenta may also demonstrate a central defect in which placental tissue is nonexistent, leaving only a membranous sheath.

The normal umbilical cord measures 50–60 cm long, contains two umbilical arteries and one vein, and typically inserts centrally within the placenta (Palacios Jaraquemada and Bruno 2000). A marginal cord insertion, also known as a battledore placenta, occurs within 1–2 cm of the placental edge. With a velamentous cord insertion, the placental vessels insert separate from the placenta and traverse between the amnion and chorion before entering the placenta (Tanaka et al. 2001). Umbilical vessels crossing the internal os of the cervix in the setting of velamentous insertion, a condition known as vasa previa, predispose to catastrophic hemorrhage of the fetal umbilical artery (Palacios Jaraquemada and Bruno 2000). Undiagnosed vasa previa has a fetal mortality rate nearing 60% (Bardo and Oto 2008).

Placenta previa refers to abnormal implantation of the placenta in the lower uterine segment,