Prenatal diagnosis of fetal intraabdominal extralobar pulmonary sequestration: a 12-year 3-center experience in China

To provide useful information for diagnosing and predicting fetal intraabdominal extralobar pulmonary sequestration (IEPS), a retrospective review of diagnostic approaches was conducted. Ultrasonography was performed serially in 21 fetuses with IEPS from 2005 to 2017. Prenatal sonographic features, treatment, and outcomes of each case were evaluated and collected. These cases of IEPS were also compared to 43 cases previously reported by other researchers from 1986 to 2017. Of the 21 sonographic features, 14 (67%) were hyperechoic, 21 (100%) were well circumscribed, and 17 (81%) depicted a mass that shifted with fetal breaths/hiccups non-synchronized with adjacent organs (sliding sign). Feeding arteries were detected prenatally in 18 patients (86%). The lesion volume was 10.17 ± 4.66 cm3, the congenital cystic adenomatoid malformation volume ratio and cardiothoracic ratio were in normal range. The gestational age at diagnosis, location and echotexture of the lesion, and rate of surgical treatment were similar to previous studies, but with a significantly higher rate of detected feeding arteries (P < 0.01), and associated anomalies (P < 0.01). All infants who underwent surgery after birth had satisfactory outcomes. The sliding sign and feeding artery are essential features of IEPS in prenatal diagnosis.


Results
General conditions of the present study. Our 3 prenatal sonographic diagnostic centers performed 210,237 prenatal ultrasound exams. Of these, there were 21 cases of IEPS, confirmed by autopsy or surgery pathology (Table 1).

Postmortem findings or surgery findings in the present study. Abortion was performed in 43%
(9/21) of the patients with IEPS, according to the preferences of the parents who did not want to suffer potential adverse outcomes. Seven of the 9 aborted fetuses had concomitant anomalies and 2 had a mass size that had grown larger since the ultrasound examination. Postmortem autopsy confirmed all findings on prenatal ultrasound. There were soft, solid masses with no adherence to the surrounding structures and no communication to the gastrointestinal tract (Fig. 1a,b). The feeding arteries were from the abdominal aorta or its branch such as the celiac trunk (Fig. 1c). Final histological evaluation confirmed a diagnosis of IEPS in all cases (Fig. 1d). The 12 fetuses that were not aborted (isolated IEPS) survived through birth. After birth, the neonates received imaging exams. Postnatal ultrasound exam was performed in all of these patients. Some lesions showed no change in size and others grew proportionally. The sonographic features of the prenatal and postnatal ultrasound (Fig. 2a) exams were similar. Postnatal computed tomography (CT; Fig. 2b) or magnetic resonance imaging (MRI) was performed in some patients, which were consistent with the prenatal ultrasound findings.
The mass was successfully removed by surgery in all 12 infants, with 5 and 7 by open and laparoscopic surgery, respectively. Ten of these lesions could be clearly separated from the adjacent organs and others had mild fibrous connection to the adrenal gland. These tumors could be easily excised after finding the tissue stalk containing the supplying vessels arising from the abdominal aorta. Postoperative pathological diagnosis revealed bronchial-alveolar tissue, with the surface covered with mesothelial cells (Fig. 2c,d), consistent with all prenatal diagnoses. All of the patients recovered uneventfully and with favorable outcomes (that is, good recovery, and lack of complications or infection or respiratory or digestive symptoms during the follow-up period). The laparoscopic approach resulted in less postoperative pain and smaller scars. The mean age at surgery was 21.3 ± 17.1 months (range 2-55 mo) and the mean duration of postoperative follow-up was 10.2 ± 3.1 months (range 1-13 mo).
Essential prenatal ultrasound feature of IEPS in the present study. IEPS was more common in the left suprarenal region and appeared as a homogenous and hyperechoic solid mass with ellipsoidal shape (Table 2). Mostly lesions were well circumscribed and displayed a sliding sign (defined as a mass that shifts during fetal breath movements or hiccups, and movements are not synchronized with that of adjacent organs) and feeding artery (Fig. 3). The heart size, cardiothoracic ratio (CTR), and congenital cystic adenomatoid malformation volume ratio (CVR) were in normal range. One-third (33%) of the patients also had other anomalies and none developed hydrops.
Variability analysis. The intra-observer and inter-observer variability is reflected by the mean percentage error (Table 3). In the 3 centers, the intra-observer mean percentage error was less than 8.0% for volume, 9.0% for heart area, 8.0% for CVR, and 9.0% for CTR. Inter-observer mean percentage error was less than 12.0% for volume, 11.0% for heart area, 12.0% for CVR, and 12.0% for CTR.
Comparison with prior studies. A comparison was conducted between the prenatal ultrasound characteristics, treatment, and outcomes of affected fetuses in the present study with that of prior studies ( Table 4). The gestational ages at diagnosis of the present and prior studies were similar (24.2 ± 2.3 cf. 24.6 ± 6.1 weeks), as was the male-to-female ratio (3.3:1 cf. 4.3:1). In addition, there were no significant differences between the present study and prior ones regarding the ratio of lesions in the left suprarenal region and those in the right suprarenal region (Fig. 4). In terms of echotexture, the present study and prior reports were similar with regard to the ratio of homogenous and hyperechoic lesions, and heterogeneous masses with cysts.
It is notable that associated anomalies were detected in 33% (7/21) of fetuses with IEPS in the present study (Table 1), but only 5% (2/43, one with diaphragmatic hernia, gastric duplication cyst and bilateral choroid plexus cysts; and another with gastric duplication cyst) in prior studies (P < 0.01). In the present study, the prenatal rate of feeding arteries detected by color Doppler (86%) was significantly higher than that of prior studies (7%, P < 0.01). Two of the 43 patients (5%) in prior studies were followed late into childhood, with spontaneous regression of the lesion after birth in serial imaging studies. One patient in a prior study had several severe anomalies and died immediately after birth. In the present study, spontaneous regression was not found in any of the 21 patients.

Discussion
Before 1986 there was no prenatal report of IEPS 9 , and it remains challenging to diagnose accurately due to the absence of clinical signs and symptoms in utero 10 . Our study investigated whether prenatal ultrasound imaging may improve the accuracy of IEPS diagnosis, and in addition identified the special characteristics of IEPS on ultrasonogram. This type of lesion was found in 0.01% of scanned patients from 2005 to 2017. We conclude that accurate prenatal diagnosis of IEPS relies on correct interpretation of sonographic features.
In gray-scale imaging, most IEPS lesions affected the left suprarenal region and involved homogenous hyperechoic solid masses with an ellipsoidal shape. Notably, the sliding sign was a unique feature of IEPS, defined as a mass that shifts during fetal breath movements or hiccups (both can be noticed as early as 10 weeks gestation 11,12 ) and is non-synchronized relative to adjacent organs, such as the liver, stomach, spleen, kidney, and adrenal gland. To the best of our knowledge, there has been no previous report of the sliding sign associated with IEPS in the fetus. In the present study, most lesions had well-defined margins, and the sliding sign could be observed. Thus, it was easier to differentiate IEPS from other tumors or organs during fetal breath movements or hiccups. Although the IEPS mass is supplied by the abdominal aorta, it is usually stable and rarely grows aggressively. Large masses in the fetal chest, such as an intrathoracic ELS, often lead to fetal hydrops and maldevelopment of the lungs 13 . In the present study, the heart area, CVR, and CTR were usually normal, and no ascites or hydrops were observed. This may be because IEPS had not progressed enough to cause high-output cardiac failure, or most lesions were located on the left side, far away from the inferior vena cava, which ensures a good venous return. Additionally, although some IEPS cases involved diaphragmatic hernia, the mass was under the diaphragm. This seems to block the diaphragmatic defect and prevent abdominal organs from further entering the thorax and compressing the fetal heart and lungs.
Another important essential ultrasound feature of IEPS in color Doppler imaging is the presence of feeding arteries arising from the abdominal aorta. Prenatal identification of the feeding artery by color Doppler was not reported until 1994 14 . Since then, feeding arteries have been detected more frequently, as the sensitivity of ultrasound has increased. Some studies reported that although systemic arterial flow is the most distinctive feature of IEPS on Doppler scan, it has high specificity but low sensitivity 8,15 . However, feeding arteries were identified in 86% of patients by using prenatal color Doppler in our study. The higher rate of detection of feeding arteries was due to the adoption of a more advanced ultrasound system, the higher sensitivity of Color Doppler, and thus greater ability to recognize the feeding arteries of IEPS. Our 3 medical centers have much experience in diagnosing IEPS. When IEPS is suspected, we focus on the relationship between the feeding artery and the abdominal aorta or its branches in multiple views, which are adjusted as necessary and not limited to the standard coronal, horizontal, and sagittal sections. Such observations allow more support for a correct diagnosis. Our study shows that feeding arteries can be identified prenatally by color Doppler, with convenience, safety, and low cost.
Diseases similar to IEPS include neuroblastoma, adrenal hematoma, and teratoma 6,8,[16][17][18] , with neuroblastoma the most common. Neuroblastoma is not usually identified until the third trimester 17 , because sympathetic ganglion cells do not fully form until 18-20 weeks' gestation 19 , whilst IEPS is usually detected in the second trimester 8 . In addition, IEPS is usually on the left side, neuroblastoma on the right 17 ; IEPS appears as a homogenous and hyperechoic solid mass, while neuroblastoma is either a solid or a mixed-solid and cystic echotexture. Calcification is an essential feature of neuroblastoma 20 8,21 . The feeding artery is an important differentiating feature to diagnose IEPS. Color Doppler imaging of neuroblastoma can reveal high vascularity in a solid/mixed mass or peripheral vascularization and no blood flow in a cystic mass [22][23][24] . We recommend that any hyperechoic solid lesion detected in the upper abdomen (especially on the left) in the second trimester should prompt suspicion of IEPS. The sliding sign is important to differentiate IEPS from other tumors or organs. A clear feeding artery from the systemic circulation should confirm IEPS.
The natural course and consequences of IEPS is still not well known. Tuberculosis and nocardiosis have been reported in ELS cases 25 , but malignancy was rare (only 2 cases of squamous cell carcinoma had ever been reported 26,27 . ELS can also be complicated with other less common conditions such as hemothorax 28 and torsion 29 . As a type of ELS, none of the IEPS cases in the present study showed any of these rare complications. This may be related to the sample size and the age of the fetuses or infants of the study population. Some previous studies have reported cases of spontaneous regression in IEPS. In the studies of Danielson and Sherman 30 and Chowdhury et al. 31 , masses grew in utero and began to regress after birth, but the earliest to regress was 6 months after birth. However, in another case the mass continued to grow until surgical resection was performed at 2 years and 9 months of age 32 . In the present study, spontaneous regression was not noticed in any of the 21 cases. Long-term follow up may be required to observe the occurrence of spontaneous regression. The treatment of IEPS is controversial. Since spontaneous regression is possible 31,33 , conservative therapy has been advocated. Today, surgical resection remains the mainstay treatment, because complete excision can eliminate the potential risks of many complications, such as infection and malignant degeneration 32 . The postsurgical (open and laparoscopic) outcomes of isolated IEPS in our study were satisfactory.
This study also has some limitations. Firstly, although our series is the largest of its kind to date, 21 fetuses is a small sample. To estimate more accurately the occurrence rate of fetal IEPS would require the participation of more medical centers. Secondly, the follow-up period was also brief. A larger series with long-term follow-up is needed to confirm the outcomes of surgical resection, and to determine the long-term post-birth consequences such as spontaneous regression.  Data collection from our database. We retrospectively searched the database of our 3 centers for cases of IEPS that had been prenatally diagnosed between March 2005 and December 2017. Only cases confirmed by autopsy or surgery pathology were included in the present study. For each case, baseline demographics, ultrasound findings, antenatal findings, fetal karyotypes, treatment, and outcomes were collected for further analysis. For ultrasound findings, the heart size, CVR, CTR, sonographic features of the lesion (size, shape, boundary, echotexture, and vascularity), and presence of hydrops were noted. To determine the heart size, heart area was measured and used to obtain Z-scores (to avoid the influence of gestational age) using previously published normative data 34 . Prior studies revealed that CVR was a useful measurement in diagnosing fetal congenital cystic adenomatoid malformation and PS, because CVR was found associated with fetal hydrops and postnatal symptoms 3,35 . Since IEPS is a type of PS, it is reasonable to speculate that CVR could also be adopted to assess the severity and prognosis of IEPS. The CVR was calculated by dividing the volume of IEPS by head circumference (HC), so that CVR = (Length × Width × Height × 0.52/HC 3,35 . The diagnostic criteria for IEPS were the following: detected in the second trimester and continued to be present; homogenous and hyperechoic; with or without cystic components, and with no significant change in echotexture in follow-up serial exams; showed clear boundary and with the surrounding tissue with no obvious adhesion; a blood supply from the abdominal aorta or its branches; and a sliding sign (see Supplementary Video 1).   Repeatability test for the present data. Two observers independently assessed the routine obstetrical and echocardiography parameters. To test intra-observer variability, a single observer analyzed the data twice on occasions separated by an interval of 1 month. To test inter-observer variability, a second observer analyzed the