Pulmonary Hypoplasia - RSNA Publications Online

2 downloads 0 Views 585KB Size Report
Osamu Nishimura, MD. Kazuhiro Minegishi, MD. Hitoshi Ishimoto, MD. Hiroshi Shinmoto, MD. Kazushige Ikeda, MD. Yasunori Yoshimura, MD. Index terms:.
Radiology

Obstetric Imaging Shinji Tanigaki, MD Kei Miyakoshi, MD Mamoru Tanaka, MD Yoshihisa Hattori, MD Tadashi Matsumoto, MD Kazunori Ueno, MD Katsuhiko Uehara, MD Osamu Nishimura, MD Kazuhiro Minegishi, MD Hitoshi Ishimoto, MD Hiroshi Shinmoto, MD Kazushige Ikeda, MD Yasunori Yoshimura, MD Index terms: Fetus, growth and development Fetus, MR, 60.121411, 60.121412, 60.121416 Fetus, respiratory system Fetus, US, 60.12981, 60.12989 Lung, congenital malformation, 60.14 Published online 10.1148/radiol.2323030359 Radiology 2004; 232:767–772 Abbreviations: FBW ⫽ fetal body weight FLV ⫽ fetal lung volume 1

From Depts of Obstetrics and Gynecology (S.T., M.T., Y.H., T.M., K. Ueno, K. Uehara, O.N., K. Miyakoshi, K. Minegishi, H.I., Y.Y.), Diagnostic Radiology (H.S.), and Pediatrics (K.I.), School of Medicine, Keio University, Tokyo, Japan. Received Mar 6, 2003; revision requested May 23; final revision received Nov 17; accepted Jan 5, 2004. Address correspondence to K. Miyakoshi, Div of Reproductive Sciences, Oregon National Primate Research Center, 505 NW 185th Ave, Beaverton, OR 97006 (e-mail: [email protected]).

Authors stated no financial relationship to disclose. Author contributions: Guarantors of integrity of entire study, S.T., Y.Y.; study concepts, M.T., H.S.; study design, S.T., M.T., K. Miyakoshi, H.S.; literature research, S.T., M.T., K. Miyakoshi, H.S.; clinical studies, S.T., Y.H., T.M., K. Ueno, O.N., K. Minegishi, H.I., H.S., K.I.; data acquisition, S.T., Y.H., T.M., K. Uehara, O.N., H.I., H.S.; data analysis/interpretation, S.T., M.T., K. Miyakoshi, H.S.; statistical analysis, S.T., M.T., K. Miyakoshi; manuscript preparation, definition of intellectual content, editing, revision/ review, and final version approval, S.T., M.T., K. Miyakoshi, H.S., Y.Y. ©

RSNA, 2004

Pulmonary Hypoplasia: Prediction with Use of Ratio of MR Imaging–measured Fetal Lung Volume to US-estimated Fetal Body Weight1 PURPOSE: To determine the ratio of fetal lung volume (FLV) to fetal body weight (FBW) by using ultrasonography (US) and magnetic resonance (MR) imaging and to evaluate the usefulness of this ratio in predicting pulmonary hypoplasia (PH) in fetuses at high risk. MATERIALS AND METHODS: MR imaging lung volumetry and US biometry were performed in 90 fetuses at 25–39 weeks gestation. In the control group of 73 fetuses, normal lung development was confirmed at neonatal follow-up and the normative ratio of MR imaging–measured FLV to US-estimated FBW (FLV/FBW) was determined. The high-risk group included 17 fetuses at risk for PH. The FLV/FBW was compared between the control and high-risk groups and with US parameters for predicting the development of PH in the high-risk group. Measurements 2 or more standard deviations below the mean control group measurement were considered abnormal. Comparisons of the FLV/FBW between groups were made by using the Student t test. The association between development of PH and measurement of each parameter was analyzed by using the Fisher exact probability test. RESULTS: In the control group, the FLV/FBW decreased with gestational age during the third trimester and had a normal distribution (mean ratio, 0.028 mL/g; range, 0.015– 0.444 mL/g). The mean FLV/FBW for the nine fetuses with PH (0.012 mL/g ⫾ 0.008) was significantly lower (P ⬍ .001) than that for the control group (0.028 mL/g ⫾ 0.007). Fetuses with abnormal FLV/FBW values were at significantly greater risk (P ⬍ .05) for PH development. Abnormal FLV/FBW values had higher diagnostic accuracy than abnormal US parameters. Sensitivity of the FLV/FBW was 89% (eight of nine fetuses); specificity, 88% (seven of eight fetuses); positive predictive value, 89% (eight of nine fetuses); negative predictive value, 88% (seven of eight fetuses); and accuracy, 88% (15 of 17 fetuses). CONCLUSION: The FLV/FBW reflects the adequacy of intrauterine lung growth and can help predict PH. ©

RSNA, 2004

Pulmonary hypoplasia is defined as a reduction in the numbers of cells, airways, and alveoli in the lung that results in a decrease in the size and weight of the lung (1,2). Pulmonary hypoplasia characteristically causes severe respiratory failure that leads to neonatal death and is associated with congenital diaphragmatic hernia, intrathoracic mass, prolonged oligohydramnios, and skeletal dysplasia (1,2). In the clinical setting, the ability to predict the presence or development of fetal pulmonary hypoplasia is important for perinatal management and parental counseling because neonates with pulmonary hypoplasia require intensive respiratory therapy immediately after birth. The prediction of 767

Radiology

pulmonary hypoplasia is particularly useful for selecting fetuses with congenital diaphragmatic hernia who will need prenatal intervention. Various parameters measured at ultrasonography (US) have been proposed for use in predicting pulmonary hypoplasia. Some authors (3,4) have found that the lung area, the ratio of lung area to thoracic area, and the ratio of thoracic circumference to abdominal circumference are useful for evaluating possible pulmonary hypoplasia. However, unfavorable imaging conditions, including oligohydramnios and maternal obesity, can compromise spatial resolution at US. Owing to advances in magnetic resonance (MR) imaging techniques, fast MR imaging modalities have been widely used to evaluate fetal abnormalities. Several authors (5–7) have demonstrated the use of fetal lung volume (FLV) measurement at fast MR imaging for prenatal assessment of pulmonary hypoplasia. In previous studies of MR imaging lung volumetry, the relative lung volume, defined as a percentage of the expected lung volume, was used to estimate fetal lung growth (8,9). However, expected lung volumes based on gestational age are not applicable to growth-restricted or macrosomic fetuses. Currently, the ratio of lung volume to body weight is the most widely used parameter for diagnosing pulmonary hypoplasia (1,2). To the best of our knowledge, however, analyses of the correlation between FLV and the occurrence of pulmonary hypoplasia have been limited. In addition, to our knowledge, no comparative studies have been performed to determine whether US parameters or MR imaging parameters are better predictors of pulmonary hypoplasia. The purposes of our study were to determine the ratio of FLV to fetal body weight (FBW) by using US and MR imaging and to evaluate the use of this ratio for predicting pulmonary hypoplasia in fetuses at high risk.

MATERIALS AND METHODS

measurements performed during the first trimester. Informed consent was obtained from the mothers for all aspects of this study, which was approved by our institutional review board. All fetuses were examined with US within 1 week before singleshot fast spin-echo MR imaging was performed at 25–39 weeks gestation. The fetuses were assigned to two groups. One group consisted of 73 fetuses at low risk of developing pulmonary hypoplasia, who served as control subjects so that nomographs of FLV could be created. The characteristics of the pregnancies for the control group, with parameters expressed as mean values ⫾ standard deviations, were as follows: maternal age, 29.9 years ⫾ 5.0; gestational age at birth, 37.1 weeks ⫾ 2.0; and birth weight, 2624 g ⫾ 538. The control group included 26 fetuses with cerebral and neural tube abnormalities, 29 fetuses with intraabdominal and abdominal wall abnormalities, four fetuses with cardiac abnormalities, and 14 fetuses without abnormalities. These fetuses had no clinical manifestations of pulmonary hypoplasia, such as a bell-shaped thorax or the requirement for high ventilatory pressures, during postnatal follow-up. The other group consisted of 17 fetuses who were at risk of developing pulmonary hypoplasia. These fetuses were considered the high-risk group and included nine fetuses with thoracoabdominal anomalies, four fetuses with renal anomalies associated with oligohydramnios, and four fetuses with musculoskeletal dysplasia. The clinical features of neonates with pulmonary hypoplasia are characterized by the immediate onset of severe respiratory insufficiency leading to neonatal death, with small lung capacity and the requirement of high ventilatory pressures in the absence of obstruction or atelectasis (1,2). A pathologic (ie, postmortem) diagnosis of pulmonary hypoplasia was made when the ratio of lung weight to body weight was less than 0.012 (10). In the high-risk group, nine neonates received a diagnosis of pulmonary hypoplasia on the basis of clinical or pathologic findings.

Patients and Groups We retrospectively examined the clinical and imaging data for a total of 90 fetuses whose mothers were referred for US and MR imaging assessment of fetal and placental abnormalities. The perinatal care of these fetuses was performed at our institution between June 1998 and April 2002. In all cases, the gestational age had been confirmed by using crown-to-rump length 768



Radiology



September 2004

MR Imaging Examination and Image Interpretation All fetuses were examined at MR imaging with a 1.5-T unit (Signa; GE Medical Systems, Milwaukee, Wis) and a torsoarray coil. After a scout acquisition was performed with a fast spoiled gradientecho sequence, a series of fetal MR images in the transverse, sagittal, and coro-

nal planes were obtained for the depiction of abnormalities. To avoid misregistration caused by fetal movement, each new acquisition was performed according to the images obtained in the immediately preceding acquisition. MR imaging parameters were ⬁/98 (repetition time msec/effective echo time msec), a 31.2-kHz bandwidth, and a total imaging time of 18 –25 seconds to acquire 15 contiguous sections. This acquisition time allowed maternal breath holding. The supervising radiologist (H.S.), who had a special interest in fetal MR imaging, optimized the field of view. After acquisition of the series of images depicting the abnormalities, transverse images were obtained through the chest of the fetus. The section thickness and intersection gap were 4 and 0 mm, respectively. In this study, a series of MR images obtained in eight patients were degraded by fetal movement, so imaging was repeated. The cross-sectional area of the lung was measured on each image section by manually tracing the outline of the lung edge (Fig 1a). The masses and herniated viscera were excluded from MR lung volumetry in cases of congenital cystic adenomatoid malformation of lung, bronchopulmonary sequestration, and congenital diaphragmatic hernia (Fig 1b). The area was then multiplied by the thickness of the section, and the individual volumes were summed for each lung. It took 15–20 minutes to perform MR image segmentation. The MR imaging– measured FLV was determined by combining the volumes of both lungs. To normalize the lung volume to the FBW, the ratio of the FLV measured at MR imaging to the estimated FBW measured at US (FLV/FBW) was obtained.

US Examination and Image Interpretation Detailed obstetric US examination of all fetuses was performed by one of four sonographers (including S.T.), who had at least 5 years experience performing fetal US. The US image interpretations were checked by board-certified radiology fellows (K. Miyakoshi, M.T.). US was performed by using a real-time US system (ProSound, Aloka, Tokyo, Japan; or Sonovista Color, Mochida, Tokyo, Japan) with a 3.5-MHz transducer. Freeze-frame capabilities were available, and on-screen calipers were used for measurement. The US-estimated FBW (in grams) was derived by using a formula for Japanese fetuses that is based on the biparietal diameter Tanigaki et al

Radiology

sia and the measurement of each parameter was analyzed by using the Fisher exact probability test. All calculations were performed by using JMP software (SAS Institute, Cary, NC). P ⬍ .05 was considered to indicate statistical significance.

RESULTS Evaluation of FLV/FBW in Fetuses at Low Risk for Pulmonary Hypoplasia

Figure 1. Transverse MR images obtained with single-shot fast spin-echo sequence (⬁/98, 4-mm section thickness). * ⫽ Normal lung parenchyma, H ⫽ heart. (a) MR image of thorax in 25-weekold fetus shows outlined areas of high signal intensity in lungs. Lung edges have been traced to measure the lung area; the numbers correspond to the order of the analysis (ie, order of outline tracing). (b) MR image of thorax in 30-week-old fetus with congenital cystic adenomatoid malformation. The mass (M) was excluded from the lung volume measurement.

(BPD), in centimeters; the fetal trunk area (FTA), in square centimeters; and the femur length (FL), in centimeters: FBW ⫽ (1.25647 䡠 BPD3) ⫹ (3.50665 䡠 FTA 䡠 FL) ⫹ 6.30994 (11). In addition, in the high-risk group, the bone thoracic circumference, thoracic area, heart area, and lung area were determined by using a cross-section of the fetal chest at the four-chamber-view level, with the heart in ventricular diastole. The abdominal circumference was measured in the transverse plane at the level of the junction of the umbilical vein and the portal sinus. To assess the clinical usefulness of US in predicting the development of pulmonary hypoplasia, three parameters were measured: the thoracic circumference–to–abdominal circumference ratio, the (thoracic area ⫺ heart area)–to–thoracic area ratio, and the lung area.

Data and Statistical Analyses The ratio of MR imaging–measured FLV to US-estimated FBW, or FLV/FBW, was assessed throughout gestation in the control group. Regression analysis of the relationship between MR imaging–measured FLV and either gestational age or US-estimated FBW was performed. In addition, the distribution of the FLV/FBW during pregnancy was examined. The relationship between MR imaging–measured FLV and either gestational age or US-estimated FBW was also examined in the high-risk group fetuses. The FLV/FBW values in the control group Volume 232



Number 3

were compared with those in the highrisk group. In the high-risk group, the difference in FLV/FBW between the fetuses with and those without pulmonary hypoplasia also was analyzed. We then evaluated the effectiveness of the FLV/ FBW and of the US parameters, including the thoracic circumference–to–abdominal circumference ratio, the (thoracic area ⫺ heart area)–to–thoracic area ratio, and the lung area, in predicting pulmonary hypoplasia in the high-risk group. All FLV/FBW measurements that were 2 or more standard deviations below the mean were considered abnormal. All US parameter measurements that were 2 or more standard deviations below the mean also were considered abnormal, as previously reported (3). We calculated the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of each parameter for predicting pulmonary hypoplasia. Data are presented as means ⫾ standard deviations. The Pearson product moment correlation coefficient (r) was used to assess the strength of associations between MR imaging–measured FLV and either gestational age or US-estimated FBW. Regression analysis was performed to evaluate the FLV/FBW according to gestational age. A Shapiro-Wilks test was used to evaluate the distribution of FLV/ FBW values. Comparisons of FLV/FBW values between the control and high-risk groups were made by using the unpaired Student t test. The association between the development of pulmonary hypopla-

In the fetuses in the control group, the MR imaging FLV ranged from 20 mL at 25 weeks to 97 mL at 37 weeks. MR imaging FLV correlated with both gestational age and US FBW, and the correlation coefficient for US FBW (r ⫽ 0.59) was higher than that for gestational age (r ⫽ 0.45) (Figs 2, 3). The FLV/FBW decreased gradually during the third trimester (Fig 4). Shapiro-Wilks test results indicated a normal distribution of FLV/FBW values (P ⬍ .05; Fig 5). The mean FLV/FBW was 0.028 mL/g ⫾ 0.007; the value 2 standard deviations below the mean was 0.013 mL/g.

Use of FLV/FBW for Predicting Development of Pulmonary Hypoplasia in Fetuses at High Risk The clinical and imaging data for the 17 fetuses in the high-risk group are listed in Table 1. As already noted, nine fetuses developed pulmonary hypoplasia, which resulted in neonatal death. Seven of the nine neonates were confirmed to have pulmonary hypoplasia at autopsy with use of lung weight–to– body weight ratio measurements. In the cases of fetuses 5 and 9 (Table 1), at birth, the neonates had a bell-shaped thorax and required high ventilatory pressures; they died, respectively, 2 and 3 hours later. The parents in both these cases declined the autopsy. Thus, the two neonates received a diagnosis of pulmonary hypoplasia solely on the basis of clinical findings. In the cases of fetuses 4 and 8 (Table 1), the neonates had a bell-shaped thorax and respiratory distress syndrome. They needed extensive ventilatory support immediately after birth and had pneumothorax that led to death at about 3 hours of age. A marked reduction in lung volume was confirmed at autopsy in both of these neonates. The MR imaging FLVs measured in the fetuses at high risk are shown in Figures 2 and 3. The mean FLV/FBW for the highrisk group (0.017 mL/g ⫾ 0.010) was significantly lower than that for the control group (0.028 mL/g ⫾ 0.007) (P ⬍ .001, Fig 6). The mean FLV/FBW for the fetuses Prediction of Pulmonary Hypoplasia



769

Radiology Figure 2. Scatterplot shows relationship between MR imaging FLV and gestational age. Dots represent 73 fetuses at low risk for pulmonary hypoplasia; open triangles, eight fetuses at high risk but without pulmonary hypoplasia; and gray triangles, nine fetuses at high risk and with pulmonary hypoplasia. MR imaging FLV increased with gestational age in fetuses at low risk (r ⫽ 0.45).

with pulmonary hypoplasia in the highrisk group (0.012 mL/g ⫾ 0.008) was significantly lower than that for the fetuses without pulmonary hypoplasia in the high-risk group (0.023 mL/g ⫾ 0.009) (P ⬍ .05, Fig 7). A significant difference in FLV/FBW was noted between the fetuses with pulmonary hypoplasia and those in the control group (0.012 mL/g ⫾ 0.008 vs 0.028 mL/g ⫾ 0.007, P ⬍ .001). There were no significant differences in FLV/ FBW between the male and female fetuses in the control or high-risk group. The performances of the FLV/FBW and the three US parameters in predicting the development of pulmonary hypoplasia are compared in Table 2. Among the US parameters, the lung area had higher diagnostic accuracy than did the thoracic circumference–to–abdominal circumference ratio and the (thoracic area ⫺ heart area)–to–thoracic area ratio. Of all the parameters examined, the FLV/FBW had the highest diagnostic accuracy: The sensitivity was 89%; the specificity, 88%; the positive predictive value, 89%; the negative predictive value, 88%; and the accuracy, 88% (Table 2). Abnormal US parameters were not significantly associated with clinically or pathologically diagnosed pulmonary hypoplasia, whereas the prevalence of pulmonary hypoplasia was significantly higher in the fetuses with abnormal FLV/FBW values than in those with normal values (P ⬍ .05).

DISCUSSION Our study findings indicate an increase in FLV with gestational age. The 4-mm770



Radiology



September 2004

Figure 3. Scatterplot shows relationship between MR imaging FLV and US FBW. Dots represent 73 fetuses at low risk for pulmonary hypoplasia; open triangles, eight fetuses at high risk but without pulmonary hypoplasia; and gray triangles, nine fetuses at high risk and with pulmonary hypoplasia. MR imaging FLV correlated with US FBW in fetuses at low risk (r ⫽ 0.59).

Figure 4. Scatterplot shows relationship between FLV/FBW and gestational age. Dots represent 73 fetuses at low risk for pulmonary hypoplasia; open triangles, eight fetuses at high risk but without pulmonary hypoplasia; and gray triangles, nine fetuses at high risk and with pulmonary hypoplasia. The FLV/FBW decreased gradually during third trimester.

thick transverse MR image sections that we used are, to our knowledge, the thinnest of the sections currently used in studies of fetal MR imaging lung volumetry, and the intersection gap was 0 mm. We speculate that the MR imaging FLV determined by summing these sections might represent a more precise measurement than do the MR imaging FLVs obtained in previous studies. Several authors have reported racebased differences in gestational age–specific birth weights. Wang et al (12) found that the gestational age–specific mean birth weights of Japanese babies were lower than those of white American babies. The MR imaging FLVs measured in

Figure 5. Graph illustrates distribution of FLV/ FBW values in 73 fetuses at low risk for pulmonary hypoplasia. The FLV/FBW had a normal distribution. The mean FLV/FBW was 0.028 mL/ g ⫾ 0.007, and the value 2 standard deviations below the mean control value was 0.013 mL/g. Tanigaki et al

Radiology

TABLE 1 Postnatal Outcomes and FLV/FBW Values in Fetuses at High Risk for Pulmonary Hypoplasia Fetus No.

Diagnosis*

Gestational Age at Delivery (wk)

Birth Weight (g)

Pulmonary Hypoplasia

LW/BW†

FLV/FBW (mL/g)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Congenital diaphragmatic hernia Type 2 osteogenesis imperfecta Potter syndrome Bronchopulmonary sequestration Potter syndrome Pena-Shokeir syndrome Potter syndrome Potter syndrome Type 2 osteogenesis imperfecta Congenital diaphragmatic hernia Congenital diaphragmatic hernia Hypochondroplasia Hydrothorax CCAML CCAML Hydrothorax CCAML

39 36 37 25 34 34 35 28 33 35 37 39 39 40 37 29 27

2980 1488 2358 842 1650 1856 1962 1144 1436 2200 3084 3004 2942 2824 3187 2058 1590

Yes Yes Yes Yes Yes Yes Yes Yes Yes No No No No No No No No

0.006 0.003 0.006 0.010 NA 0.008 0.009 0.007 NA ND ND ND ND ND ND ND ND

0.003‡ 0.005‡ 0.007‡ 0.010‡ 0.009‡ 0.012‡ 0.013‡ 0.013‡ 0.032 0.007‡ 0.017 0.020 0.022 0.023 0.028 0.033 0.034

* CCAML ⫽ congenital cystic adenomatoid malformation of lung. † LW/BW ⫽ lung weight–to– birth weight ratio, as measured at autopsy. Fetuses 5 and 9, respectively, died 2 and 3 hours after birth; their parents declined autopsy, and, thus, LW/BW data are not available (NA). ND ⫽ no data: Pulmonary hypoplasia was not diagnosed in fetuses 10 –17, who survived; thus, no LW/BW data were obtained. ‡ Value 2 standard deviations below the mean or less.

TABLE 2 Use of FLV/FBW and US Parameters in Predicting Pulmonary Hypoplasia in Fetuses at High Risk

Figure 6. Graph illustrates comparison of mean FLV/FBW values (⫾ standard deviations) between fetuses in the control and high-risk groups. * ⫽ P ⬍ .001 for difference in FLV/FBW between the two groups. The FLV/FBW was significantly lower in the high-risk group than in the control group.

Figure 7. Graph illustrates comparison of mean FLV/FBW values (⫾ standard deviations) between fetuses in the high-risk group with and fetuses in the high-risk group without pulmonary hypoplasia (PH). * ⫽ P ⬍ .05 for difference in FLV/FBW between the two groups. The FLV/FBW was significantly lower in the fetuses in the high-risk group with pulmonary hypoplasia than in those in the high-risk group without the abnormality. Volume 232



Number 3

Parameter*

Sensitivity

Specificity

PPV

NPV

Accuracy

FLV/FBW Lung area TC/AC (TA ⫺ HA)/TA

8/9 (89) 7/9 (78) 4/9 (44) 7/9 (78)

7/8 (88) 6/8 (75) 6/8 (75) 3/8 (38)

8/9 (89) 7/9 (78) 4/6 (67) 7/12 (58)

7/8 (88) 6/8 (75) 6/11 (54) 3/5 (60)

15/17 (88) 13/17 (76) 10/17 (59) 10/17 (59)

Note.—Data are numbers of fetuses, with percentages in parentheses. NPV ⫽ negative predictive value, PPV ⫽ positive predictive value. * TC/AC ⫽ thoracic circumference–to–abdominal circumference ratio. (TA ⫺ HA)/TA ⫽ ratio of (thoracic area ⫺ heart area) to thoracic area.

the present study were smaller than those obtained by using US and MR imaging in other fetal lung volumetry studies. These findings indicate that there may be racebased differences in lung volumes (5–7). We observed a correlation between MR imaging FLV and US FBW in the fetuses in the control group; the correlation with US FBW was higher than the correlation with gestational age. Shepard et al (13) similarly observed a strong correlation between the lung weight and the body weight of human fetuses at postmortem analysis. In their study, the FLV/FBW appeared to decrease with gestational age during the third trimester in fetuses in the control group. Several authors have studied the lung weight–to– body weight ratio at autopsy of human fetuses (13,14). Wigglesworth et al (14) found that the lung weight–to– body weight ratio in fetuses with normal

lung growth decreased during the second half of pregnancy. Duncan et al (6) reported that the median postmortem lung volume was 0.0317 of the fetal volume measured at MR imaging; these data are similar to our current study results. The use of FLVs in the control group was a potential limitation of our study. The control group was not a strictly healthy population: It included fetuses with structural abnormalities. However, all of the control group fetuses had normal chest findings at US and diagnoses not known to be associated with pulmonary hypoplasia. Indeed, no control group fetuses had pulmonary hypoplasia at birth. These findings suggest that the measurements obtained in the control subjects examined in the present study represent reasonable estimates of normal lung development. The FLV/FBW measurements in the Prediction of Pulmonary Hypoplasia



771

Radiology

high-risk group were significantly lower than those in the control group (P ⬍ .001). In addition, significantly low FLV/ FBW values were observed in the fetuses with pulmonary hypoplasia as compared with the values observed in the fetuses who were at high risk for but did not have this abnormality (P ⫽ .019). The present study results indicate that fetuses who are at risk for pulmonary hypoplasia have a small lung volume relative to their body weight. Pathologic parameters used to diagnose pulmonary hypoplasia include the lung weight–to– body weight ratio, the radial alveolar count, and the amount of lung DNA relative to body weight (1,2). The radial alveolar count is the number of alveolar septa demarcated by a line starting at the center of a respiratory bronchiole and ending at the angles to the right of the nearest connective tissue septum. A radial alveolar count of less than 4.1 (ie, 75% of the mean normal count) has been shown to be one of the criteria used to diagnose pulmonary hypoplasia (1,2). Of these diagnostic parameters, the lung weight–to– body weight ratio is the simplest and most widely used. Pulmonary hypoplasia is characterized by a substantial reduction in lung volume in the clinical setting. We hypothesized that the FLV/FBW might be useful for predicting pulmonary hypoplasia at birth. We found that fetuses with abnormal FLV/FBW values had a significantly (P ⬍ .05) increased risk of developing pulmonary hypoplasia compared with fetuses with normal values. In our study, an FLV/ FBW that was 2 or more standard deviations below the control mean value was considered abnormal. We observed a normal distribution of FLV/FBW values in the fetuses in the control group, in support of this definition. In previous studies (8,9) of fetal volumetry at MR imaging, in cases of congenital diaphragmatic hernia, the relative lung volume ratio— defined as the FLV relative to the expected value— could be used to predict pulmonary hypoplasia and postnatal outcome. Mahieu-Caputo et al (8) found that neonatal death occurred in all fetuses with a relative lung volume ratio of less than 0.35. Paek et al (9) concluded that a relative lung volume ratio of less than 0.4 was suggestive of a poor postna-

772



Radiology



September 2004

tal outcome. However, the use of the FLV/FBW to predict pulmonary hypoplasia was not investigated in these studies. Our study results suggest that an abnormally low FLV/FBW is a useful predictor of pulmonary hypoplasia. Given the limited number of cases in our study, further investigations should be performed to assess the clinical applicability of the FLV/FBW for the prediction of pulmonary hypoplasia. The FLV/FBW had better diagnostic accuracy in the prediction of pulmonary hypoplasia in the fetuses who were at high risk than did the US parameters that we studied. Various US parameters have been proposed for use in assessing fetal lung development (3,4). Yoshimura et al (3) found that the thoracic circumference–to–abdominal circumference ratio and the lung area were useful for evaluating pulmonary hypoplasia. Vintzileos et al (4) reported that of the predictive indexes that they evaluated, the (thoracic area ⫺ heart area)–to–thoracic area ratio had the most value for predicting pulmonary hypoplasia. However, two-dimensional US parameters are not necessarily representative of total lung volume. Three-dimensional US lung volumetry has been investigated for use in assessing lung development, although the use of this examination to predict pulmonary hypoplasia remains to be evaluated (15– 17). US evaluation of fetal lung growth is limited by poor acoustic differentiation between fetal lung tissue and surrounding structures. Our study results show that the combination of MR imaging– measured lung volume parameters and US-estimated body weight parameters— specifically, the FLV/FBW—was more useful for predicting pulmonary hypoplasia than any US parameter alone. Acknowledgment: The authors are grateful to Naoki Shimada, MD, assistant professor of the Department of Preventive Medicine and Public Health, Keio University, for his valuable statistical assistance. References 1. Lauria MR, Gonik B, Romero R. Pulmonary hypoplasia: pathogenesis, diagnosis, and antenatal prediction. Obstet Gynecol 1995; 86:466 – 475. 2. Laudy JA, Wladimiroff JW. The fetal lung. II. Pulmonary hypoplasia. Ultrasound Obstet Gynecol 2000; 16:482– 494. 3. Yoshimura S, Masuzaki H, Gotoh H, Fukuda S, Ishimaru T. Ultrasonographic

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

prediction of lethal pulmonary hypoplasia: comparison of eight different ultrasonographic parameters. Am J Obstet Gynecol 1996; 175:477– 483. Vintzileos AM, Campbell WA, Rodis JF, Nochimson DJ, Pinette MG, Petrikovsky BM. Comparison of six different ultrasonographic methods for predicting lethal fetal pulmonary hypoplasia. Am J Obstet Gynecol 1989; 161:606 – 612. Coakley FV, Lopoo JB, Lu Y, et al. Normal and hypoplastic fetal lungs: volumetric assessment with prenatal single-shot rapid acquisition with relaxation enhancement MR imaging. Radiology 2000; 216:107–111. Duncan KR, Gowland PA, Moore RJ, Baker PN, Johnson IR. Assessment of fetal lung growth in utero with echo-planar MR imaging. Radiology 1999; 210:197– 200. Rypens F, Metens T, Rocourt N, et al. Fetal lung volume: estimation at MR imaging—initial results. Radiology 2001; 219: 236 –241. Mahieu-Caputo D, Sonigo P, Dommergues M, et al. Fetal lung volume measurement by magnetic resonance imaging in congenital diaphragmatic hernia. BJOG 2001; 108:863– 868. Paek BW, Coakley FV, Lu Y, et al. Congenital diaphragmatic hernia: prenatal evaluation with MR lung volumetry— preliminary experience. Radiology 2001; 220:63– 67. Askenazi SS, Perlman M. Pulmonary hypoplasia: lung weight and radial alveolar count as criteria of diagnosis. Arch Dis Child 1979; 54:614 – 618. Aoki M. Fetal weight calculation: Osaka University method. In: Chiba Y, ed. Ultrasound in obstetrics and gynecology. Kyoto, Japan: Kinpodo, 1990; 67–77. Wang X, Guyer B, Paige DM. Differences in gestational age-specific birthweight among Chinese, Japanese and white Americans. Int J Epidemiol 1994; 23:119 – 128. Shepard TH, Shi M, Fellingham GW, et al. Organ weight standards for human fetuses. Pediatr Pathol 1988; 8:513–524. Wigglesworth JS, Desai R, Aber V. Quantitative aspects of perinatal lung growth. Early Hum Dev 1987; 15:203–212. Pohls UG, Rempen A. Fetal lung volumetry by three-dimensional ultrasound. Ultrasound Obstet Gynecol 1998; 11:6 –12. Osada H, Iitsuka Y, Masuda K, et al. Application of lung volume measurement by three-dimensional ultrasonography for clinical assessment of fetal lung development. J Ultrasound Med 2002; 21:841– 847. Laudy JA, Janssen MM, Struyk PC, Stijnen T, Wladimiroff JW. Three-dimensional ultrasonography of normal fetal lung volume: a preliminary study. Ultrasound Obstet Gynecol 1998; 11:13–16.

Tanigaki et al