Case of the Month October 2016

Placenta Accreta

Submitted by Yaniv Raphael, MD with Rola Saouaf, MD and approved by Thomas J. Learch, MD.

SM is a 33-year-old G5P3 female 32 week 0 day gestation transferred from outside hospital with recurrent and intractable vaginal bleeding, requiring multiple blood transfusions. At 20 weeks gestation, during routine prenatal trans-vaginal ultrasound, the patient was diagnosed with marginal placenta previa; however, due to recurrent vaginal bleeding, a repeat scan was performed, confirming the diagnosis of complete posterior previa. Additional findings included an abnormally thickened placenta with multiple lacunae and disorganized hyper-vascularity at the utero-placental interface. Overall, placenta percreta was suspected. At 31 week 5 day gestation, the patient presented to outside facility with heavy vaginal bleeding, and was subsequently transfused 2 units packed RBCs. The bleeding eventually resolved, and the patient was transferred to Cedars-Sinai for higher level of care.

At time of examination, the patient reported feeling well with no recurrent vaginal bleeding, though endorsed intermittent cramping. Obstetric history is significant for three prior pregnancies via vaginal delivery in addition to one spontaneous abortion. The patient had a dilatation and curettage (D&C) in the past. She denied illicit drug use, cigarette smoking, or alcohol consumption. Vital signs were within normal range, and physical examination revealed a soft gravid uterus without tenderness to palpation. Laboratory data was pertinent for a normocytic anemia (Hb 9.2, MCV 90). Otherwise, remainder of laboratory values were within normal range. The patient was admitted to Cedars-Sinai and received two units packed RBCs, magnesium for neuroprotection, and a course of antenatal corticosteroids.

A non-contrast MRI of the soft tissue pelvis was ordered for elucidation of the diagnosis of placenta accreta. The examination revealed a gravid uterus with a posterior placenta and no subchorionic hemorrhage or abruption. The placenta was well attached with maximal thickness measuring 10 cm in anteroposterior dimension. Paracentral three-vessel umbilical cord insertion was noted. Marked heterogeneity of the placenta was identified, predominantly along the left posterolateral uteroplacental margin, extending approximately 14 cm in craniocaudad dimension (Figure 1). Additional loss and thinning of the normal trilayered myometrium with focal bulging and indistinct margins past the serosal surface were also demonstrated (Figure 2). Importantly, disorganized intra-placental bands and lacunae were recognized (Figure 3). No direct invasion of adjacent bowel loops of viscera was appreciated. A complete placenta previa, covering the entirety of the internal cervical os, was depicted, in addition to cervical effacement; the cervical canal measured 1.8 cm in length (Figure 4).
Regarding the fetus, a singleton intrauterine pregnancy was noted with vertex presentation. No definite neural or visceral anatomic abnormalities were identified. Visceral situs solitus was demonstrated. The remainder of the examination was pertinent for bilateral non-obstructive hydro-ureteronephrosis, likely physiologic in the context of gravid uterus.

For larger images, please click on Figure 1 , Figure 2 , Figure 3   and Figure 4 .

Figure 1: Sagittal single-shot fast spin echo demonstrates marked placental heterogeneity, abnormal uterine bulging contour (arrows), and loss of normal trilaminar myometrium.

Figure 2: Coronal single-shot fast spin echo reveals marked placental heterogeneity, abnormal uterine bulging contour, and indistinct margins past the uterine serosa (arrows). No frank visceral invasion was noted.

Figure 3: Axial single-shot fast spin echo reveals randomly oriented bands in the anterior portion of the placenta (arrows). In contradistinction, thinner placental septae are identified posteriorly. Lacunae are typically hyperintense on T2-weighted sequences (not shown).

Figure 4: Parasagittal single-shot fast spin echo reveals complete placenta previa with coverage of the entire internal cervical os. Cervical length is shorted to 1.8 cm with a “funneled” configuration, increasing risk of pre-term labor. Also note smooth and maintained vesico-uterine interface.

At the time, the decision was made to proceed with cesarean section and hysterectomy with the gynecology-oncology service present. The operation was complicated by significant peripartum maternal hemorrhage, and the patient received 12 units of packed RBCs, 8 units FFP, 1 unit platelets, and 500 mL albumin. A cesarean hysterectomy and bilateral salpingectomy was ultimately performed. The preserved ovaries were noted to be adherent to the uterus due to surrounding decidualized tissue. The abdomen was otherwise unremarkable without ascites or mass. Bowel surfaces were smooth and uninterrupted.

Upon pathologic analysis, the uterus was enlarged, measuring 20.6 x 14.8 x 6.5 cm in dimensions. The attached placenta was implanted within the posterior uterine corpus and lower uterine segment with possible extension into the anterior lower uterine segment, consistent with complete previa. Sectioning revealed focal area of placental tissue with possible superficial invasion of the myometrium at the posterior uterine corpus, consistent with placenta accreta. There was no evidence for placenta percreta.

The patient’s post-operative course was uncomplicated. She was subsequently discharged on post-operative day #4 with standard precautions and routine follow-up.

Placenta accreta (PA) encompasses a spectrum of placental adhesive disorders. The underlying pathogenesis is related to abnormal placentation, namely failure of decidua basalis membrane formation beneath the endometium, thus allowing direct trophoblastic (chorionic villi) contact and invasion to the myometrium. Within the PA spectrum, Accreta connotes the mildest form, referring to placental villi Attachment without invasion. Placenta Increta is defined as placental trophoblastic Invasion; whereas, placenta Percreta—the most severe—describes frank penetration and extension past the uterine serosa. Possible invasion of adjacent structures (i.e. bowel loops, bladder) may be present. The disorder affects 1 in 2,000 pregnancies, but incidence has recently been increasing due to growing number of cesarean sections performed in the Western world. Other acknowledged risk factors include: prior dilatation and curettage (as in this case), myomectomy, endometritis, multi-gestational pregnancy, and advanced maternal age. Symptoms generally include painless vaginal bleeding in the second and third trimesters. Diagnosis may be established on routine prenatal trans-vaginal ultrasound.

Ultrasound remains the primary initial diagnostic modality for placenta accreta with sensitivity and specificity approaching 90%. MRI possesses similar accuracy in diagnosing PA, though excels in evaluation of depth of invasion or in certain circumstances (i.e. posterior PA, prior myomectomy). Ultrasound examination should always be performed with Color Doppler to assess vascularity at the uterine-serosa margin. In addition, serial scans are recommended as placenta accreta vera may progress to percreta. Ultrasonographic findings are variable, but include loss of the normal hypoechoic retroplacental myometrial (clear) zone with thinning/disruption of the uterine serosal-bladder interface. A well reported finding of lacunae within the placenta, conferring a “moth-eaten” or “swiss cheese” appearance, refers to numerous disorganized vessels that violate placental borders and extend past the serosa, oftentimes involving adjacent structures (Figure 5). This finding has a sensitivity 80% and specificity 92%, underscoring the value of Color Doppler. Additional findings include interruption of bladder wall - uterine interface and thinning of the myometrium. However, the diagnosis is often confirmed on dedicated MRI to ascertain depth of invasion and involvement of viscera for optimal pre-operative planning.

For a larger image, please click on Figure 5 .

Figure 5: Gray-scale sagittal trans-abdominal ultrasound demonstrating large anechoic spaces (yellow arrows) within the inferior placenta at the bladder interface, consistent with vascular lacunae. These vascular lakes are seen to approximate and invade through the serosa into the bladder wall (red arrow). Case from Core Radiology, Jacob Mandell.

Upon reviewing MRI, placental morphology, location (with respect to the uterus), and anomalies should be reported. Examination is typically performed in the left lateral decubitus position during the third trimester in order to avoid risk of impaired venous return from uterine compression of the inferior vena cava. Adequate bladder distention is advised to evaluate the vesico-uterine interface; an over-distended bladder, however, may push a low-lying placenta caudally, conferring the false appearance of complete or marginal placenta previa. In our case, the placenta entirely covered the internal cervical os, which is consistent with a complete placenta previa. Additionally, the distance between the internal and external oses was foreshortened to 1.8 cm (abnormal < 3 cm), which is compatible with cervical effacement, increasing the risk for pre-term labor. Anomalous umbilical cord insertion site and/or vasa previa should also be assessed; in our case, both parameters were within normal range. Of note, placental appearance and enhancement are dynamic and evolve throughout pregnancy, becoming more lobulated and heterogenous on T2-weighted sequences as the trimesters progress. The normal myometrium maintains a trilaminar appearance on T2-weighted sequences: heterogenous hyperintense core with enveloping hypointense bands, as assessed on T2-weighted sequences. T1-weighted sequences are typically reserved for evaluation of subchorionic hemorrhage or placental abruption due to the inherent magnetic characteristics of hemoglobin.

Many MRI findings of PA are documented, some of which are well illustrated in our case. The most predictive sign includes randomly oriented dark intra-placental bands (on T2-weighted sequences), which may extend from the placenta-myometrial interface and are typically thicker than normal placental septae. These are thought to represent abnormal intra-placental fibrin deposition. In contradistinction, anomalous tortuous vessels are usually hyperintense on T2-weighted sequences and correspond to the ultrasonographic finding of lacunae (Figure 6). Other signs depicted in our case include outward uterine bulging, loss of normal tri-layered myometrium, and placental heterogeneity (Figure 7). The distinction between accreta-increta and increta-perceta is not always apparent, unless frank invasion of adjacent viscera is evident (i.e. bladder, bowel). Additionally, some degree of placenta-myometrial interface lobulation and myometrial thinning may be seen in normal patients. Certain imaging planes may confer a false discontinuous contour of the uterus, owing to its natural curved shape. Care should be undertaken to not solely rely on a single finding, but rather maintain a high index of suspicion particularly in high-risk patients.

For larger images, please click on Figure 6 and Figure 7 .

Figure 6: Heterogeneous placenta with dark placental bands and vessels. (A) Axial single-shot fast spin-echo image through the placenta shows a heterogenous placenta (P). The distinction between placental bands and vessels is apparent on T2 weighted sequences as high intensity structures (B). Case from Placental Evaluation with Magnetic Resonance, Brian C. Allen MD.

Figure 7: Placenta Percreta. Coronal single-shot fast spin-echo of a 38-year old female at 35 weeks gestation demonstrates a lateral bulge (arrows) in the placenta (P) and lack of identifiable subjacent myometrium at the site of prior resection of a right-sided rudimentary uterine horn. Without the history of prior surgery in this area, this would be an unusual location for abnormal placentation. Case from Placental Evaluation with Magnetic Resonance, Brian C. Allen MD.

Treatment usually consists of a multi-specialty approach with close monitoring and planned surgical expertise. Transfer is often coordinated to a high level of care facility, equipped with appropriate ancillary services (i.e. intensive care, blood bank, etc). A cesarean hysterectomy is usually performed. In anticipation of high volume maternal blood loss (average 3-5 L), internal iliac artery balloon catheters may be prophylactically placed. In more dire circumstances, post-partum uterine artery embolization (UAE) may be undertaken to control hemorrhage. Conservative management implies cesarean section with uterine preservation, though remains controversial in the United States, and increases the risk of recurrent PA in subsequent pregnancies.

PA is an important diagnosis for the radiologist to consider as it portends significant risk of maternal and fetal demise, namely due to catastrophic hemorrhage at time of placental separation. Rarer complications include: cystostomy (15%), ureteral injury (2%), and bleeding diatheses (2%).

Visit our Case of the Month Archive