Objective
The purpose of this study was to develop a simple and accurate approach for risk stratification of fetal lung lesions that are associated with respiratory compromise at birth.
Study Design
We conducted a retrospective review of 64 prenatal lung lesions that were managed at a single fetal care referral center (2001-2011). Sonographic data were analyzed and correlated with perinatal outcomes.
Results
Hydrops occurred in only 4 cases (6.3%). Among fetuses without hydrops, the congenital pulmonary airway malformation volume ratio (CVR) was the only variable that was associated significantly with respiratory compromise and the need for lung resection at birth ( P < .01). Based on a maximum CVR >1.0, the sensitivity, specificity, positive predictive value, and negative predictive value for respiratory morbidity were 90%, 93%, 75%, and 98%, respectively.
Conclusion
Nonhydropic fetuses with a maximum CVR >1.0 are a subgroup of patients who are at increased risk for respiratory morbidity and the need for surgical intervention. These patients should be delivered at a tertiary care center with pediatric surgery expertise to ensure optimal clinical outcomes.
Congenital pulmonary airway malformations (CPAMs), formerly referred to as congenital cystic adenomatoid malformations, are the most common lung lesions to be identified by fetal ultrasound imaging. Historically, CPAMs were thought to be rather infrequent, with a reported incidence between 1 in 25,000 and 1 in 35,000 live births. However, with ongoing improvements in the resolution of sonographic imaging, smaller CPAMs that are now being diagnosed with increasing frequency. As a result, the current incidence of CPAMs may now be as low as 1 in 12,000 live births.
Given our ability to diagnose these smaller less symptomatic lesions before birth, the overall prognosis of prenatally diagnosed CPAMs has improved considerably in recent years. Compared with older reviews that cited high morbidity and rates of hydrops that approach 40% among fetal lung lesions, hydrops and fetal death now are encountered infrequently in most maternal-fetal medicine practices. Nevertheless, there remain a number of important, yet undefined, areas in their management. For example, the recommended frequency of prenatal follow-up scans in these patients remains unknown, and the prognostic value of various sonographic characteristics that include mediastinal shift, microcystic disease, and indicators of mass size that include the CPAM volume ratio (CVR) continue to be debated. We therefore conducted this study to evaluate carefully the ultrasound characteristics and growth patterns of fetal lung lesions in an effort to identify fetuses who are at increased risk for respiratory morbidity and the need for lung surgery in the immediate newborn period.
Materials and Methods
We retrospectively reviewed all prenatal lung lesions referred to the Fetal Diagnosis and Treatment Center at the C.S. Mott Children’s and Von Voigtlander Women’s Hospital between August 2001 and July 2011. Approval for this study was granted by the University of Michigan Institutional Review Board (#51132). All expectant mothers who were carrying a fetus with a newly diagnosed thoracic mass underwent an initial comprehensive 2-dimensional ultrasound imaging (Philips iU22 xMatrix; Philips Healthcare, Andover, MA; Voluson 730; General Electric, Waukesha, WI; or ATL HDI 2000; Advanced Technology Laboratories, Seattle, WA) performed by one of our staff maternal-fetal medicine specialists. During this evaluation, the mass was characterized fully based on location, cyst characteristics (ie, microcystic vs macrocystic disease), and blood supply. Any nonimmune hydrops, mediastinal shift, cardiac dysfunction, or extrapulmonary fetal anomalies were noted. Hydrops was defined as the presence of fluid in ≥2 spaces, which included ascites, pleural effusion, pericardial effusion, or skin edema. In each case, the lesion was measured prospectively in centimeters across 3 different dimensions to calculate a CVR, a volumetric index of mass size that allows for comparisons of fetuses at different gestational ages as originally described elsewhere. To calculate the CVR, the length, width, and depth of the mass were multiplied by a 0.52 correction factor (with the assumption that its shape was a prolate ellipse) and then divided by the head circumference to normalize for gestational age. In an effort to minimize any potential intra- and interoperator variability over time, the maximal height of the lesion was first measured in the sagittal plane ( Figure 1 ). The maximal dimensions of the lesion were then measured in the plane perpendicular to this axis.
Subsequent ultrasound assessments were performed based on the presence of hydrops and the most recent CVR measurement. Lesions with a CVR >1.6 and/or associated with hydrops were followed by weekly ultrasound imaging. All other lesions were evaluated by ultrasound images every 2 weeks until the lesion began to regress in size. If there were an interval increase in the CVR or new onset of hydrops, ultrasound assessment was conducted more frequently, usually at least once a week. Once patients were documented to have a stable lesion over the course of 4 weeks, monthly assessments were performed until delivery.
A multidisciplinary team of maternal-fetal medicine specialists, pediatric surgeons, and other providers was involved in the counseling of all expectant mothers. An amniocentesis for karyotype and microarray analyses was recommended in the presence of multiple fetal anomalies. Fetal treatment varied based on ultrasound findings and available practice guidelines at the time of the evaluation. For example, after 2009, we routinely administered maternal betamethasone for fetuses with microcystic lesions with hydrops or a CVR >1.6 at <26 weeks’ gestation, as detailed elsewhere. Near-term fetuses with very large lesions (eg, CVR >2.0) in association with hydrops were considered for the ex utero intrapartum treatment (EXIT) procedure with lung resection while on placenta bypass, as previously described. All other fetuses were treated expectantly and evaluated at birth for signs of respiratory morbidity (defined as tachypnea, the need for supplemental oxygen, or requirement of mechanical ventilation). Asymptomatic neonates were observed in the newborn nursery and discharged after approximately 48 hours with outpatient pediatric surgery follow up at 3 months of age. At that time, chest computed tomography scans were performed to determine the need for elective resection. Lesions in neonates with respiratory morbidity were evaluated by chest computed tomography scan and echocardiogram followed by open thoracotomy and pulmonary resection during the newborn period when indicated.
Prenatal sonographic variables were correlated with respiratory morbidity and the need for neonatal lung resection. In addition to the initial CVR, the maximum CVR (defined as the highest CVR measured at any time during the pregnancy) and the final CVR (defined as the last CVR measured before delivery) were evaluated as prognosticators of perinatal outcome. Statistical analyses were performed by the Mann-Whitney U test or the Fisher exact test with the SPSS software package (SPSS Inc, Chicago, IL). Receiver operating characteristic curves were calculated based on the CVR in relationship with outcome. Significance was defined as a probability value of < .05. All values were reported as the mean, unless otherwise indicated.
Results
Initial evaluation
Sixty-four fetal lung lesions were identified. Forty-nine of these (76.6%) were referred from outside institutions. The mean gestational age at initial evaluation was 24.3 weeks (range, 18.7–36.0 weeks). Forty fetuses (62.5%) had a left-sided lesion; 23 fetuses (35.9%) had a right-sided lesion, and 1 fetus (1.6%) had bilateral disease. Fifty lesions (78.1%) were microcystic, and 14 lesions (21.9%) were macrocystic. Forty-four lesions (68.8%) were associated with mediastinal shift, as evidenced by moderate-to-severe displacement of the heart towards the contralateral hemithorax. Based solely on sonographic features, fifty-four lesions (84.4%) were thought to be CPAM; 8 lesions (12.5%) were considered bronchopulmonary sequestrations (BPS), and 2 lesions (3.1%) were diagnosed as a hybrid of the 2 aforementioned lesions. Four fetuses(6.25%) had major extrapulmonary anomalies that included congenital heart disease (n = 2), renal agenesis with contralateral multicystic dysplastic kidney (n = 1), and bilateral hydronephrosis (n = 1). All of these major anomalies were confirmed shortly after birth. One fetus had hydrops on initial presentation at 22.1 weeks’ gestation. The mean initial CVR for the entire group was 0.66 (range, 0.02–3.8).
Follow up evaluations
A mean of 4.3 ultrasound images per patient were performed. Although there was a subpopulation of larger lung lesions for which the CVR tended to increase in size and peak at approximately 26 weeks’ gestation ( Figure 2 , A), most lesions either remained stable or decreased slightly through the reminder of pregnancy. As a result, the growth pattern for the entire cohort of lung lesions was fairly constant ( Figure 2 , B). The average maximum CVR increased to 0.89 (range, 0.02–3.80) at a mean gestational age of 26.3 weeks; the average final CVR decreased slightly to 0.58 (range, 0.00–3.00) at a mean gestational age of 30.0 weeks. Five lesions (7.8%), all of which were microcystic, were undetectable by ultrasound imaging before delivery.
In addition to the 1 fetus who initially presented with hydrops, 3 others had hydrops at 28.3, 32.4, and 35.1 weeks’ gestation. Fetal right heart dysfunction was confirmed by echocardiography in all cases. Two of these patients had microcystic lesions that were treated with maternal steroids with subsequent hydrops resolution. The initial, maximum, and final CVR among hydropic fetuses were 2.20, 2.27, and 1.87, respectively.
Perinatal outcomes
Complete outcomes data were available in 62 fetuses (96.9%). There were no pregnancy terminations or attempts at in utero lung resection. There were 60 live births, 51 of which (79.7%) were inborn. Two babies (3.1%) were delivered elsewhere and were lost to follow up. The mean gestational age at birth was 38.3 weeks (range, 33.4–41.4 weeks). The mean birthweight was 3.19 kg (range, 1.96–4.68 kg). There was 1 death, which resulted in a mortality rate of 1.7%, that occurred shortly after birth in a nonhydropic neonate who had profound pulmonary hypoplasia that was the result of unilateral renal agenesis with a contralateral multicystic dysplastic kidney.
Excluding the patient with pulmonary hypoplasia and 2 other patients with major congenital heart disease, respiratory morbidity at birth was evident in 8 of 58 cases (13.8%). The average initial CVR among nonhydropic fetuses who had respiratory morbidity at birth was 0.82. Although none of these patients had an initial CVR >1.6, the average maximum CVR in this cohort rose to 2.05 on subsequent imaging. All fetuses who experienced hydrops at any time in gestation had respiratory morbidity, even if the hydrops had resolved before delivery. At the other end of the spectrum, all lesions that were undetectable by ultrasound imaging before delivery (7.8%) were asymptomatic at birth. The average maximum CVR for this latter subgroup was 0.37 (range, 0.03–0.67).
Perinatal lung resection (eg, neonatal resection, EXIT-to-resection) was performed in 7 of 58 patients (12.1%). Surgical procedures in those who experienced hydrops included the EXIT procedure with lung resection and extracorporeal membrane oxygenation cannulation in the setting of severe pulmonary hypertension (n = 1), neonatal lung resection (n = 2), and extracorporeal membrane oxygenation cannulation followed by delayed lung resection (n = 1). One of these cases has been published previously as a case report. One additional EXIT procedure with lung resection was performed on a nonhydropic fetus who had a final CVR of 3.0. Two additional nonhydropic cases with respiratory morbidity required lung resection within the first week of life. Another nonhydropic neonate with a small microcystic lesion presented with a pneumothorax that was managed with a chest tube. That child had a thoracoscopic lung resection at 6 months of age. In contrast to all asymptomatic neonates who were discharged within 3 days of delivery, symptomatic neonates generally required lung resection and therefore had prolonged hospitalizations. In the symptomatic group, the mean ventilator duration was 12.7 days (range, 3–30 days), and the mean length of hospitalization was 32.8 days (range, 17–51 days). Of all resected lesions, 45% were CPAMs; 20% were BPSs, and 25% were hybrid lesions.
Prenatal risk stratification
Fetuses with an initial CVR >1.6 were significantly more likely to experience hydrops ( P < .001). Whereas 3 of the 5 patients (60%) with an initial CVR >1.6 experienced hydrops, only 1 patient (1.8%) with an initial CVR <1.6 became hydropic. Based on an initial CVR >1.6, the sensitivity was 75%; the specificity was 97%; the positive predictive value (PPV) was 60%, and the negative predictive value (NPV) was 98% for the detection of hydrops.
Neither microcystic disease nor macrocystic disease were predictive of respiratory morbidity at birth ( P = .43). Mediastinal shift similarly was not associated with respiratory morbidity at birth ( P = .24). There was no association between prenatal diagnosis (ie, CPAM, BPS, or hybrid lesion) and the likelihood of respiratory morbidity. However, the CVR was found to be a strong predictor of respiratory morbidity ( Figure 3 , A). The maximum CVR in neonates who had respiratory morbidity after birth was significantly higher than those who were asymptomatic (2.16 vs 0.70, respectively; P = .01; Figure 3 , B). The final CVR in neonates who had respiratory morbidity after birth was significantly higher than those who were asymptomatic (1.89 vs 0.38, respectively; P < .001; Figure 3 , B). In an analysis that was limited to nonhydropic fetuses, the maximum and final CVR continued to be independent predictors of respiratory morbidity. Although there was also a trend towards respiratory morbidity among those with a high initial CVR, this did not reach statistical significance ( Figure 3 , B; P = .07).
The prenatal predictors for neonatal lung resection were virtually identical to those for respiratory morbidity. Neither microcystic disease nor mediastinal shift was significantly associated with the need for early lung surgery ( P = .61 and .08, respectively). The maximum CVR in fetuses who required resection was 2.77 compared with 0.63 in those who did not need resection ( P < .001). The final CVR for fetuses who required resection was 2.4 compared with 0.38 in those who did not need resection ( P < .001). The initial CVR was also associated significantly with the likelihood of neonatal lung surgery. The initial CVR in those fetuses who required resection was 1.90 compared with 0.51 in those who did not need resection ( P = .03).
Because of the strong association between the CVR and perinatal outcomes, receiver operating characteristic curves were generated in an attempt to stratify patients into high- and low-risk categories based on the likelihood for respiratory morbidity at birth and the need for neonatal lung surgery. Results of this analysis that was based on initial, maximum, and final CVR values set at 1.0 were the most significant. The area under the curve was greatest for the final CVR ( Table 1 ). Data on sensitivity, specificity, PPV, and NPV for initial, maximum, and final CVRs >1.0 are shown in Table 2 . Both the maximum CVR and the final CVR had the highest PPV and NPV for respiratory morbidity and the need for neonatal lung surgery ( Table 2 ). For fetuses with a maximum CVR >1.0, the likelihood of respiratory morbidity at birth and the need for neonatal lung resection were both 75%. Conversely, for fetuses with a maximum CVR <1.0, the likelihood of being asymptomatic at birth and avoiding neonatal lung resection was 98% and 100%, respectively. Although an initial CVR >1.0 had a high NPV, it had a relatively lower PPV when compared with both maximum and final CVR >1.0.
| Variable | Positive, n | Negative, n | Area under the receiver operating curve | P value |
|---|---|---|---|---|
| Respiratory symptoms | ||||
| Initial CVR >1.0 | 10 | 45 | 0.717 | .033 |
| Maximum CVR >1.0 | 12 | 43 | 0.727 | .017 |
| Final CVR >1.0 | 7 | 48 | 0.908 | .001 |
| Perinatal resection | ||||
| Initial CVR >1.0 | 10 | 45 | 0.739 | .019 |
| Maximum CVR >1.0 | 12 | 43 | 0.750 | .009 |
| Final CVR >1.0 | 7 | 48 | 0.929 | < .001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
