http://informahealthcare.com/jmf ISSN: 1476-7058 (print), 1476-4954 (electronic) J Matern Fetal Neonatal Med, Early Online: 1–6 ! 2014 Informa UK Ltd. DOI: 10.3109/14767058.2014.927862
ORIGINAL ARTICLE
Assessment of fetal lung maturity by ultrasound: objective study using gray-scale histogram
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Ryerson University on 12/03/14 For personal use only.
Ana Paula Avritscher Beck1, Edward Araujo Ju´nior1, Ana Teresa Figueiredo Stochero Leslie2, Luiz Camano1, and Antonio Fernandes Moron1 1
Department of Obstetrics and 2Discipline of Neonatal Pediatrics, Department of Pediatrics, Paulista School of Medicine – Federal University of Sa˜o Paulo (EPM-UNIFESP), Sa˜o Paulo – SP, Brazil Abstract
Keywords
Objective: To evaluate lung maturity using ultrasound (US), comparing the subjective and gray-scale histogram (GSH) techniques. Methods: A total of 77 single pregnancies were evaluated and divided into the following two groups: 11 women of gestational age 28 to 35 + 6 weeks and 66 women 36 weeks. The women underwent to emergency or planned cesarean section, according to fetal–maternal indications. The US was performed on the mean sagittal plane of the fetal torso, in order to observe the lung and hepatic areas. Fetal lung maturity was evaluated subjectively and through GSH. After delivery, the incidence of respiratory distress in the newborn was evaluated. The analyses were considered to be correct or incorrect, and comparisons were made using the McNemar test. In order to compare lung/hepatic echogenicity using GSH in groups with and without respiratory distress, the Student’s t-test was used. Results: The subjective evaluation identified 41 cases (53.2%) correctly and 36 (46.8%) incorrectly, while GSH found 58 (75.3%) correctly and 19 (24.7%) incorrectly (p ¼ 0.006). There was a significant difference in mean lung/hepatic echogenicity between the groups with and without respiratory distress (1.05 versus 1.26; p ¼ 0.002). In the group of 28 to 35 + 6 weeks, GSH presented sensitivity, specificity and accuracy in predicting respiratory distress of 61.9%, 89.1% and 81.6%, respectively. Conclusion: The evaluation of fetal lung maturity through GSH was more effective than the subjective method in predicting respiratory distress among newborns.
Fetus, histogram, lung, maturity, respiratory distress, ultrasound
Introduction Premature neonates are defined as newborns less than 37 weeks gestation, according to American College of Obstetrics and Gynecology and the American Academy of Pediatrics. This group of infants is heterogeneous and constitutes a wide range of gestational ages, which are associated with different outcomes. The lung development begins early in human gestation, and growth extends into childhood [1]. The successful transition between the fetal life to extrauterine environment depends on various changes occurring throughout the last weeks of gestation, in especial the production of the surfactant [2]. Pulmonary surfactant is essential to prevent the alveolar collapse; it is a complex mixture of phospholipids, neutral
Address for correspondence: Prof. Edward Araujo Ju´nior, PhD, Department of Obstetrics, Paulista School of Medicine – Federal University of Sa˜o Paulo (EPM-UNIFESP), Rua Carlos Weber, 956, apto. 113 Visage, Alto da Lapa, Sa˜o Paulo – SP CEP 05303-000, Brazil. Tel/Fax: +55-11-37965944. E-mail: [email protected]
History Received 19 March 2014 Revised 30 April 2014 Accepted 21 May 2014 Published online 19 June 2014
lipids and proteins that are produced by alveolar type 2 cells from the 20th week of gestation. The surfactant deficiency due to lung immaturity is associated to respiratory distress syndrome (RDS) or hyaline membrane syndrome, and the prematurity constitutes the main risk factor for RDS [3]. The major variable for RDS is the state of lung maturation at birth. The diagnosis of RDS is based on clinical diagnosis of respiratory distress and progressive increase in oxygen requirement after birth, added by characteristic radiographic findings of reticulogranular pattern (ground-glass) and air bronchograms. Differential diagnosis includes: transient tachypnea of the newborn, sepsis, pneumonia and air leak syndrome. RDS incidence decreases after 32 weeks gestation; however, it is still present among the late preterm infants [4]. Despite all the advances in the treatment of RDS, it represents the greatest cause of respiratory morbidity among premature neonates in the neonatal intensive care units (NICUs) [5], thus reinforcing the need for fetal lung evaluation, especially in planned cesarean sections, to avoid iatrogenic prematurity. Before the use of ultrasound (US), iatrogenic prematurity was the main cause of RDS, corresponding to 10–20% of all
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Ryerson University on 12/03/14 For personal use only.
2
A. P. A. Beck et al.
cases [6]. With the introduction of US in obstetrics, there was a significant reduction of this incidence, especially due to proper dating of the pregnancy. The incidence of respiratory distress diminishes as gestational age advances, but it may be present up to the 39th week of pregnancy. In a study conducted in Denmark in 2008, the incidences of respiratory morbidities in elective caesarians performed after the 37th week were evaluated and found to be non-negligible [7]. The most accepted way for evaluating fetal lung maturity requires amniocentesis in order to obtain and analyze amniotic fluid. Among the most common techniques are: counting of lamellar bodies [8], lecithin/sphingomyelin ratio [9] and phosphatidylglycerol level [10]. In addition, a recent study showed that performing invasive procedures in pregnant women with more than 36 weeks of pregnancy is unjustifiable [11]. US is considered to be a noninvasive alternative for evaluating lung maturity, although subjective evaluation has not been shown to be an effective method for this purpose [12,13]. Recently, US has been shown to be useful for the diagnosis of lung disease for newborns at bedside in the NICU [14]. Gray-scale histogram (GSH) is a tool present in nearly all US devices. It is a graphical representation of the quantity, reflectivity and arrangement of pixels present in a delimited image. Evaluation of the GSH reduces intraobserver error because it is performed on the same image. The analysis provides the number of pixels, a mean and a standard deviation, which presents the advantage of not being subject to attenuation caused by the patient’s weight, the depth and the gain [15]. Maeda et al. [16] were the pioneers in evaluating lung and hepatic echogenicity through GSH, and observed that lung echogenicity increased with advancing gestational age, while hepatic echogenicity remained unaltered. However, there are no studies correlating fetal lung maturity evaluated by means of GSH and the incidence of respiratory distress. The objective of this study was to evaluate fetal lung maturity through US using both the subjective and GSH methods and correlate them with the prediction of respiratory distress.
Methods A prospective observational study was performed on 77 pregnant women divided into two groups: 11 women of gestational age 28 to 35 + 6 weeks and 66 women 36 weeks. These women underwent to either emergency or planned cesarean sections not more than 48 h after the US. This study was approved by the Research Ethics Committee of the Federal University of Sa˜o Paulo (UNIFESP), and those who consented to participate as a volunteer signed a consent statement. Gestational age was based on the last menstrual period and was confirmed by means of US performed by the 13th week. The exclusion criteria were: hypertension, diabetes mellitus, estimated fetal weight below the 10th percentile for the gestational age [17], congenital abnormalities or chromosome defects in the newborn, use of sedative drugs, signs of acute fetal distress from baseline antepartum cardiotocography and Apgar score57 in the first minute.
J Matern Fetal Neonatal Med, Early Online: 1–6
US examinations were performed by a single examiner (A. P. A. B.) using the Voluson Pro device (General Electric Medical System, Zipf, Austria), equipped with a convex transducer (RAB 4-8L). A mean sagittal plane of the fetal thorax was observed in order to view the liver and lungs with clear separation by the diaphragm. Whenever possible, the acoustic shadow of the fetal ribs was avoided by moving the transdutor to the right and to the left in order to get the best quality image without artifacts. Initially, a subjective evaluation of fetal lung maturity was made: this was considered hyperechogenic, if lung echogenicity was lower than that of the liver; isoechogenic, if the lung echogenicity was similar to that of the liver; and hypoechogenic, if the lung echogenicity was greater than that of the liver. Following this, an objective evaluation was made using GSH, with channel 1 standardized for the lung and the region of interest (ROI) delimited manually, so to not include the bronchial tree. Channel 2 was standardized for the liver, and the ROI was manually delimited, so as not to include the hepatic vascularization. The device automatically provided the graphs with the values of the means and standard deviations and the number of pixels (Figure 1). Newborns were evaluated daily in the Obstetrics Ward or the NICU of Hospital Sa˜o Paulo, UNIFESP, throughout their hospitalization. The birth weight and Apgar score at the first and fifth minutes were evaluated, and gestational age was determined using the New Ballard method. Newborns were considered mature, regarding their lungs, if they did not need orotracheal intubation or an oxygen mask. Respiratory distress was considered to be present among the newborns that developed transient tachypnea of newborn or RDS. Newborns showing increased work of breathing soon after birth were assessed at the delivery room regarding the need of mechanical ventilation. All patients admitted to the NICU requiring any type of respiratory support had a chest X-ray, in especial premature infants. Term infants showing progressive improvement on the respiratory distress within the first 4–6 h of life did not have a chest X-ray. The decision about the therapy with surfactant replacement was based on the signs of respiratory distress, the progressive oxygen requirement and chest X-ray findings. The therapy with prophylactic exogenous surfactant for preterm infants is not a routinely practice in our NICU. Initial continuous positive airway pressure has been used in the delivery room for premature neonates showing early signs of respiratory distress. If started in the delivery room, it may help the newborn to establish functional residual capacity, stabilize the chest wall and reduce the air resistance [18]. In order to classify the subjective and GSH method for evaluating lung maturity, the categories of correct and incorrect were used. For the mature newborns, if the subjective evaluation showed a hyperechogenic lung and the newborn did not need ventilation support, and the lung GSH presented a higher mean than the hepatic GSH, then the classification was correct. For the mature newborns, if the subjective evaluation showed an isoechogenic or hypoechogenic lung and the newborn was mature regarding the lungs and the lung GSH presented a mean lower than the hepatic GSH, then the classification was incorrect. For the immature newborns, if the subjective evaluation showed an
DOI: 10.3109/14767058.2014.927862
Fetal lung maturity using gray-scale histogram
3
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Ryerson University on 12/03/14 For personal use only.
Figure 1. Mean sagittal plane of the fetal trunk showing lung and hepatic areas, with channel 1 positioned in the lung area and channel 2 in the hepatic area and automatic calculation of gray-scale histograms.
isoechogenic or hypoechogenic lung and the newborn needed ventilatory support and the lung GSH presented a mean lower than the hepatic GSH, the classification was correct. For the immature newborns, if the subjective evaluation showed a hyperechogenic lung and the newborn needed ventilatory support and the lung GSH presented a mean higher than the hepatic GSH, the classification was incorrect. The data were transferred to a spreadsheet in the Excel 2007 software (Microsoft Corp., Redmond, WA) and were analyzed using the SPSS software, version 15.0 for Windows (SPSS Inc., Chicago, IL). In order to evaluate the correlation of hepatic and lung echogenicity between each other using GSH, as well as gestational age, Pearson’s correlation coefficient (r) was used. In order to compare the subjective and GSH methods, in terms of correct and incorrect, the McNemar test was used. In order to compare the lung and liver GSH measurements, as well as the lung/liver relationship, in the groups with and without respiratory distress, the paired Student’s t-test was used. For the subjective and GSH methods, the accuracy, specificity, sensitivity and positive and negative predictive values were calculated for predicting respiratory distress and lung maturity. In all analyses, the significance level used was p50.005.
Results A total of 77 single pregnancies were evaluated, of which 11 (14.3%) were between 28 and 35 + 6 weeks and 66 (85.7%) were 36 weeks. The mean gestational age was 37.06 (ranging from 28 to 40 weeks). The mean birth weight was 2847 g (ranging from 1258 to 3915 g). None of the newborns developed RDS, but 20 newborns (25.9%) developed some degree of respiratory distress, which was classified as transient tachypnea of the newborn. A direct and significant correlation was observed between the means for lung and hepatic echogenicity using GSH
(r ¼ 0.75; p50.001). A non-significant positive correlation was observed between the mean lung echogenicity using GSH and the gestational age (r ¼ 0.13; p ¼ 0.258). A nonsignificant negative correlation was observed between the mean hepatic echogenicity using GSH and the gestational age (r ¼ –0.11; p ¼ 0.361) (Figure 2). The subjective evaluation classified 41 cases (53.2%) correctly and 36 (46.8%) incorrectly. The evaluation using GSH classified 58 (75.3%) correctly and 19 (24.7%) incorrectly. In comparing the two evaluations, 26 (33.8%) of the correct classifications from GSH corresponded to incorrect subjective evaluations; and nine (11.7%) correct classifications in the subjective evaluation corresponded to incorrect GSH evaluations. Therefore, the discordance corresponded to 35 (45.5%) of the total. A statistically significant discordance (p ¼ 0.006) was observed between the two evaluations and, consequently, there was a greater proportion of correct classifications in the evaluation using GSH (Table 1). In the group from 28 to 35 + 6 weeks, nine newborns (81.8%) developed transient tachypnea of the newborn. This group consisted of 22 cases, among which 14 (63.6%) were classified correctly and 8 (36.4%) incorrectly using the subjective technique; while GSH presented 9 correct classifications (40.9%) and 13 incorrect classifications (59.1%). There was no statistical difference between the two techniques (p ¼ 0.180). Taking one as the cutoff value for the relationship between the means for lung/hepatic echogenicity, GSH presented sensitivity, specificity and accuracy for predicting respiratory distress of 61.9%, 89.1% and 81.6%, respectively (Table 2). In the group 36 weeks, nine newborns (81.8%) developed transitory tachypnea of the newborn. This group consisted of 55 patients, with 27 correct classifications (49.1%) and 28 incorrect classifications (50.9%) through the subjective technique; while GSH presented 49 correct
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Ryerson University on 12/03/14 For personal use only.
4
A. P. A. Beck et al.
J Matern Fetal Neonatal Med, Early Online: 1–6
Figure 2. (a) Scatter plot of correlation between the means for the lung and liver echogenicity through GSH; (b) scatter plot of correlation between mean lung echogenicity and gestational age through GSH; (c) scatter plot of correlation between mean liver echogenicity and gestational age through GSH. GSH: gray-scale histogram.
Table 1. Distribution of the results from the whole sample. GSH evaluation Subjective evaluation Correct Incorrect Total
Correct n (%)
Incorrect n (%)
Total n (%)
32 (41.6) 26 (33.8) 58 (75.3)
9 (11.7) 10 (13.0) 19 (24.7)
41 (53.2) 36 (46.8) 77 (100.0)
McNemar test: p ¼ 0.006*; n ¼ absolute number; GSH ¼ gray-scale histogram. *Statistical significance.
respiratory distress, showing a statistical difference for lung echogenicity (p ¼ 0.005). Table 4 presents the relationship of the means for lung and hepatic echogenicity between the groups with and without respiratory distress, showing a statistically significant difference (p ¼ 0.002). Taking one as the cutoff for the relationship of the means for lung/hepatic echogenicity, GSH presented sensitivity, specificity and accuracy for predicting lung maturity of 89.1%, 40.9% and 75.3%, respectively (Table 5).
Discussion Table 2. Comparative results from the methods for predicting respiratory distress. Test Subjective GSH
Sensitivity Specificity PPV NPV PLR NLR Accuracy % 63.6 61.9
% 52.7 89.1
% % 35 78.4 68.4 86
1.3 5.7
0.7 0.4
% 55.8 81.6
GSH, gray-scale histogram; PPV, positive predictive value; NPV, negative predictive value; PLR, positive likelihood ratio; and NLR, negative likelihood ratio.
classifications (89.1%) and six incorrect classifications (10.9%). There was a statistical difference between the two techniques (p50.001). Table 3 presents the means for lung and hepatic echogenicity through GSH in the groups with and without
The most accurate diagnosis regarding fetal lung maturity is always made after birth. Evaluation of immaturity, however, generates greater difficulty because respiratory distress may result from various factors, thus complicating correct diagnosis. Today, the greater number of elective interventions in pregnancies has created a growing need for evaluation of fetal lung maturity [6,7]. However, for this situation, there has been a lack of any effective, harmless and easily reproducible method. The increase in lung echogenicity over the course of pregnancy occurs because of the great proliferation of capillaries, which produce a greater number of interfaces, called the canalicular phase. The hepatic tissue, on the other hand, undergoes a linear increase in its echogenicity, without great changes, thus favoring, in this case, studies on the lung/ liver relationship. However, human vision has difficulty in
Fetal lung maturity using gray-scale histogram
DOI: 10.3109/14767058.2014.927862
Table 3. Distribution of the means for lung and hepatic echogenicity using the gray-scale histogram in both groups. Group Without RD (n ¼ 57)
With RD (n ¼ 20)
Student’s t test
Mean echogenicity–lung Mean ± standard deviation Minimum–maximum
28.0 ± 6.0 17.2–42.1
23.2 ± 7.3 12.1–41.9
p ¼ 0.005*
Mean echogenicity–liver Mean ± standard deviation Minimum–maximum
23.0 ± 6.4 12.7–37.0
22.5 ± 7.2 11.3–35.2
p ¼ 0.760
Parameter
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Ryerson University on 12/03/14 For personal use only.
RD: respiratory distress. *Statistical significance. Table 4. Distribution of the relation between the means for lung and hepatic echogenicity using the gray scale histogram in both groups. Group Lung/liver ratio–mean echogenicity
Without RD (n ¼ 57)
With RD (n ¼ 20)
Student’s t test
Mean ± standard-deviation Minimum–maximum
1.26 ± 0.25 0.69–1.89
1.05 ± 0.20 0.71–1.49
p ¼ 0.002*
RD: respiratory distress. *Statistical significance.
Table 5. Comparative results from the methods for predicting lung maturity. Test Subjective GSH
Sensitivity Specificity PPV NPV PLR NLR Accuracy % % % % % 52.7 89.1
63.6 40.9
78.4 35.0 79.0 60.0
1.5 1.5
0.7 0.3
55.8 75.3
GSH, gray-scale histogram; PPV, positive predictive value; NPV, negative predictive value; PLR, positive likelihood ratio; and NLR, negative likelihood ratio.
differentiating between the various shades of gray [19], which justifies the need for objective evaluation. The histogram is the most commonly used type of graph, and this shows the distribution of frequencies for continuous or discrete data. In the US field, it has always been used to calibrate gray scales. However, it could aid in determining the echotexture of tissues. Because of the easy application and reproducibility of GSH, its use in other tissues such as caesarian scars [20] and the uterine cervix during pregnancy [21] has been described. The technique to obtain the GSH is easy and reproducible [20,21], and in our experience is necessary performing few US exams to obtain the ideal expertise. Since there is no consensus in the literature regarding the best method for studying lung maturity through US [22–24], two types of easily performed evaluation were compared in this study: subjective analysis and analysis through GSH. In our sample, no cases of RDS were found. We believe that this was due to the use of antenatal steroids in pregnancies less than 34 weeks, which is routine in our service. This practice has been shown to be effective in reducing the risk of RDS. Respiratory distress was found in 20 newborns, mostly from the group of mothers with gestations
5
of less than 36 weeks, which were already expected to present some degree of lung immaturity. In a pioneering study on the evaluation of lung and hepatic echogenicity through GSH, Maeda et al. [16] evaluated 52 normal fetuses between 24 and 38 weeks and observed that lung echogenicity was greater between 30 and 38 weeks than between 24 and 29 weeks, but with no difference in hepatic echogenicity. The ratio of the means for lung/hepatic echogenicity was less than one between 24 and 29 weeks and greater than one between 30 and 35 weeks. However, these authors did not correlate lung echogenicity with development of respiratory distress among newborns. In a more recent study, Serizawa and Maeda [25] evaluated predictions of lung maturity through GSH among 22 newborns with respiratory distress and 25 newborns with mature lungs. They observed values for lung echogenicity, hepatic echogenicity and the lung/hepatic echogenicity ratio in the group with respiratory distress of 32.21%, 37.03% and 0.865, respectively. In the visual comparison of this study, the means were lower. However, the different gestational period evaluated in this study (28–40 weeks) and the fact that we did not exclude pregnant women who received corticoid at a gestational age younger than 34 weeks need to be emphasized. In this study, the evaluation through GSH obtained better results than the subjective evaluation, especially in the group in which maturity was already expected ( 36 weeks), due to the aforementioned difficulty in distinguishing between subtle differences in shades of gray [19]. GSH was effective for predicting lung immaturity, especially if the relationship between the means for lung and liver echogenicity was used, without requiring any other propaedeutic resource. In the group in which immaturity was expected, GSH produced specificity of 89.1% and sensitivity of 61.9%. One point that should be noted is that this related to a specific test, thus demonstrating its importance concerning decision-making in weighing up the risk of prematurity in situations of elective resolution of pregnancies. In summary, evaluation of fetal lung maturity through GSH was more effective than the subjective method for predicting respiratory distress in newborns.
Declaration of interest The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.
References 1. Burri PH. Structural aspects of postnatal lung development: alveolar formation and growth. Biol Neonate 2006;89:313–22. 2. Hillman NH, Kallapur SG, Jobe AH. Physiology of transition from intrauterine to extrauterine life. Clin Perinatol 2012;39:769–83. 3. Gleason CA, Devaskar SU. Avery’s diseases of the newborn. 9th ed. Philadelphia (PA): Elservier Saunders; 2012. 4. Jobe AH. What is RDS in 2012? Early Hum Dev 2012;88:S42–4. 5. Rubaltelli FF, Dani C, Reali MF, et al. Acute neonatal respiratory distress in Italy: a one-year prospective study. Acta Paediatr 1998; 87:1261–8. 6. Hansen AK, Wisborg K, Uldbjerg N, Henriksen TB. Elective caesarean section and respiratory morbidity in the term and nearterm neonate. Acta Obstet Gynecol Scand 2007;86:389–94.
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Ryerson University on 12/03/14 For personal use only.
6
A. P. A. Beck et al.
7. Hansen AK, Wisborg K, Uldbjerg N, Henriksen TB. Risk of respiratory morbidity in term infants delivered by elective caesarean section: cohort study. Br Med J 2008;336:85–7. 8. Ashwood ER, Palmer SE, Taylor BA, Pangree S. Lamellar body counts for rapid maturity testing. Obstet Gynecol 1993;81: 619–26. 9. Gluck MD, Kulovich MV, Borer RC, et al. Diagnosis of the respiratory distress syndrome by amniocentesis. Am J Obstet Gynecol 1971;109:440–5. 10. Hallman M, Gluck L. Phosphatidilglicerol in lung surfactant II, subcelular distribution and mechanism of biosynthesis in vitro. Biochim Biophys Acta 1975;409:172–91. 11. Luo G, Norwitz E. Revisiting amniocentesis for fetal lung maturity after 36 weeks gestation. Obstet Gynecol 2008;1:61–8. 12. Cayea PD, Grant DC, Doubilet PM, Johns TB. Prediction of fetal lung maturity: inaccuracy of study using conventional ultrasound instruments. Radiology 1985;155:473–5. 13. Feingold M, James SR, Cetrulo CL, Koza D. Fetal lung to liver reflective ratio and lung maturity. J Clin Ultrasound 1987;15: 384–7. 14. Liu J. Lung ultrasonography for the diagnosis of neonatal lung disease. J Matern Fetal Neonatal Med 2014;27:856–61. 15. Allison JW, Barr LL, Massoth RJ, et al. Understanding the process of quantitative ultrasonic tissue characterization. Radiographics 1994;14:1009–108. 16. Maeda K, Utsu M, Yamamoto N, Serizawa M. Echogenicity of fetal lung and liver quantified by the grey-level histogram width. Ultrasound Med Biol 1999;25:201–8.
J Matern Fetal Neonatal Med, Early Online: 1–6
17. Hadlock FP, Harrist RB, Martinez-Poyer J. In utero analysis of fetal growth: a sonographic weight standard. Radiology 1991;181: 129–33. 18. de Klerk AM, de Klerk RK. Use of continuous positive airway pressure in preterm infants: comments and experience from New Zealand. Pediatrics 2001;108:761–3. 19. Julesz B, Gilbert EM, Frish HL. Inability of humans to discriminate between visual features that agree in second-order statistics – revisited. Perception 1973;2:391–405. 20. Santos Filho Ode O, Nardozza LM, Arau´jo Junior E, et al. [Cesarean uterine scar evaluation by the grey-level histogram]. Rev Assoc Med Bras 2010;56:99–102. 21. Furtado MR, Pires CR, Araujo Ju´nior E, et al. Transvaginal grey scale histogram of the cervix at 20–25 weeks of pregnancy. Aust N Z J Obstet Gynaecol 2010;50:444–9. 22. La Torre R, Cosmi E, Anceschi MH, et al. Preliminary report on a new and noninvasive method for the assessment of fetal lung maturity. J Perinat Med 2003;31:431–4. 23. Piazze JJ, Anceschi MM, Picone G, et al. Association between maternal-fetal Doppler velocimetry and fetal lung maturity. J Perinat Med 2003;31:484–8. 24. Cobo T, Bonet-Carne E, Martı´nez-Terro´n M, et al. Feasibility and reproducibility of fetal lung texture analysis by Automatic Quantitative Ultrasound Analysis and correlation with gestational age. Fetal Diagn Ther 2012;31:230–6. 25. Serizawa M, Maeda K. Noninvasive fetal lung maturity prediction based on ultrasonic gray level histogram width. Ultrasound Med Biol 2010;36:1998–2003.
ORIGINAL ARTICLE
Assessment of fetal lung maturity by ultrasound: objective study using gray-scale histogram
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Ryerson University on 12/03/14 For personal use only.
Ana Paula Avritscher Beck1, Edward Araujo Ju´nior1, Ana Teresa Figueiredo Stochero Leslie2, Luiz Camano1, and Antonio Fernandes Moron1 1
Department of Obstetrics and 2Discipline of Neonatal Pediatrics, Department of Pediatrics, Paulista School of Medicine – Federal University of Sa˜o Paulo (EPM-UNIFESP), Sa˜o Paulo – SP, Brazil Abstract
Keywords
Objective: To evaluate lung maturity using ultrasound (US), comparing the subjective and gray-scale histogram (GSH) techniques. Methods: A total of 77 single pregnancies were evaluated and divided into the following two groups: 11 women of gestational age 28 to 35 + 6 weeks and 66 women 36 weeks. The women underwent to emergency or planned cesarean section, according to fetal–maternal indications. The US was performed on the mean sagittal plane of the fetal torso, in order to observe the lung and hepatic areas. Fetal lung maturity was evaluated subjectively and through GSH. After delivery, the incidence of respiratory distress in the newborn was evaluated. The analyses were considered to be correct or incorrect, and comparisons were made using the McNemar test. In order to compare lung/hepatic echogenicity using GSH in groups with and without respiratory distress, the Student’s t-test was used. Results: The subjective evaluation identified 41 cases (53.2%) correctly and 36 (46.8%) incorrectly, while GSH found 58 (75.3%) correctly and 19 (24.7%) incorrectly (p ¼ 0.006). There was a significant difference in mean lung/hepatic echogenicity between the groups with and without respiratory distress (1.05 versus 1.26; p ¼ 0.002). In the group of 28 to 35 + 6 weeks, GSH presented sensitivity, specificity and accuracy in predicting respiratory distress of 61.9%, 89.1% and 81.6%, respectively. Conclusion: The evaluation of fetal lung maturity through GSH was more effective than the subjective method in predicting respiratory distress among newborns.
Fetus, histogram, lung, maturity, respiratory distress, ultrasound
Introduction Premature neonates are defined as newborns less than 37 weeks gestation, according to American College of Obstetrics and Gynecology and the American Academy of Pediatrics. This group of infants is heterogeneous and constitutes a wide range of gestational ages, which are associated with different outcomes. The lung development begins early in human gestation, and growth extends into childhood [1]. The successful transition between the fetal life to extrauterine environment depends on various changes occurring throughout the last weeks of gestation, in especial the production of the surfactant [2]. Pulmonary surfactant is essential to prevent the alveolar collapse; it is a complex mixture of phospholipids, neutral
Address for correspondence: Prof. Edward Araujo Ju´nior, PhD, Department of Obstetrics, Paulista School of Medicine – Federal University of Sa˜o Paulo (EPM-UNIFESP), Rua Carlos Weber, 956, apto. 113 Visage, Alto da Lapa, Sa˜o Paulo – SP CEP 05303-000, Brazil. Tel/Fax: +55-11-37965944. E-mail: [email protected]
History Received 19 March 2014 Revised 30 April 2014 Accepted 21 May 2014 Published online 19 June 2014
lipids and proteins that are produced by alveolar type 2 cells from the 20th week of gestation. The surfactant deficiency due to lung immaturity is associated to respiratory distress syndrome (RDS) or hyaline membrane syndrome, and the prematurity constitutes the main risk factor for RDS [3]. The major variable for RDS is the state of lung maturation at birth. The diagnosis of RDS is based on clinical diagnosis of respiratory distress and progressive increase in oxygen requirement after birth, added by characteristic radiographic findings of reticulogranular pattern (ground-glass) and air bronchograms. Differential diagnosis includes: transient tachypnea of the newborn, sepsis, pneumonia and air leak syndrome. RDS incidence decreases after 32 weeks gestation; however, it is still present among the late preterm infants [4]. Despite all the advances in the treatment of RDS, it represents the greatest cause of respiratory morbidity among premature neonates in the neonatal intensive care units (NICUs) [5], thus reinforcing the need for fetal lung evaluation, especially in planned cesarean sections, to avoid iatrogenic prematurity. Before the use of ultrasound (US), iatrogenic prematurity was the main cause of RDS, corresponding to 10–20% of all
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Ryerson University on 12/03/14 For personal use only.
2
A. P. A. Beck et al.
cases [6]. With the introduction of US in obstetrics, there was a significant reduction of this incidence, especially due to proper dating of the pregnancy. The incidence of respiratory distress diminishes as gestational age advances, but it may be present up to the 39th week of pregnancy. In a study conducted in Denmark in 2008, the incidences of respiratory morbidities in elective caesarians performed after the 37th week were evaluated and found to be non-negligible [7]. The most accepted way for evaluating fetal lung maturity requires amniocentesis in order to obtain and analyze amniotic fluid. Among the most common techniques are: counting of lamellar bodies [8], lecithin/sphingomyelin ratio [9] and phosphatidylglycerol level [10]. In addition, a recent study showed that performing invasive procedures in pregnant women with more than 36 weeks of pregnancy is unjustifiable [11]. US is considered to be a noninvasive alternative for evaluating lung maturity, although subjective evaluation has not been shown to be an effective method for this purpose [12,13]. Recently, US has been shown to be useful for the diagnosis of lung disease for newborns at bedside in the NICU [14]. Gray-scale histogram (GSH) is a tool present in nearly all US devices. It is a graphical representation of the quantity, reflectivity and arrangement of pixels present in a delimited image. Evaluation of the GSH reduces intraobserver error because it is performed on the same image. The analysis provides the number of pixels, a mean and a standard deviation, which presents the advantage of not being subject to attenuation caused by the patient’s weight, the depth and the gain [15]. Maeda et al. [16] were the pioneers in evaluating lung and hepatic echogenicity through GSH, and observed that lung echogenicity increased with advancing gestational age, while hepatic echogenicity remained unaltered. However, there are no studies correlating fetal lung maturity evaluated by means of GSH and the incidence of respiratory distress. The objective of this study was to evaluate fetal lung maturity through US using both the subjective and GSH methods and correlate them with the prediction of respiratory distress.
Methods A prospective observational study was performed on 77 pregnant women divided into two groups: 11 women of gestational age 28 to 35 + 6 weeks and 66 women 36 weeks. These women underwent to either emergency or planned cesarean sections not more than 48 h after the US. This study was approved by the Research Ethics Committee of the Federal University of Sa˜o Paulo (UNIFESP), and those who consented to participate as a volunteer signed a consent statement. Gestational age was based on the last menstrual period and was confirmed by means of US performed by the 13th week. The exclusion criteria were: hypertension, diabetes mellitus, estimated fetal weight below the 10th percentile for the gestational age [17], congenital abnormalities or chromosome defects in the newborn, use of sedative drugs, signs of acute fetal distress from baseline antepartum cardiotocography and Apgar score57 in the first minute.
J Matern Fetal Neonatal Med, Early Online: 1–6
US examinations were performed by a single examiner (A. P. A. B.) using the Voluson Pro device (General Electric Medical System, Zipf, Austria), equipped with a convex transducer (RAB 4-8L). A mean sagittal plane of the fetal thorax was observed in order to view the liver and lungs with clear separation by the diaphragm. Whenever possible, the acoustic shadow of the fetal ribs was avoided by moving the transdutor to the right and to the left in order to get the best quality image without artifacts. Initially, a subjective evaluation of fetal lung maturity was made: this was considered hyperechogenic, if lung echogenicity was lower than that of the liver; isoechogenic, if the lung echogenicity was similar to that of the liver; and hypoechogenic, if the lung echogenicity was greater than that of the liver. Following this, an objective evaluation was made using GSH, with channel 1 standardized for the lung and the region of interest (ROI) delimited manually, so to not include the bronchial tree. Channel 2 was standardized for the liver, and the ROI was manually delimited, so as not to include the hepatic vascularization. The device automatically provided the graphs with the values of the means and standard deviations and the number of pixels (Figure 1). Newborns were evaluated daily in the Obstetrics Ward or the NICU of Hospital Sa˜o Paulo, UNIFESP, throughout their hospitalization. The birth weight and Apgar score at the first and fifth minutes were evaluated, and gestational age was determined using the New Ballard method. Newborns were considered mature, regarding their lungs, if they did not need orotracheal intubation or an oxygen mask. Respiratory distress was considered to be present among the newborns that developed transient tachypnea of newborn or RDS. Newborns showing increased work of breathing soon after birth were assessed at the delivery room regarding the need of mechanical ventilation. All patients admitted to the NICU requiring any type of respiratory support had a chest X-ray, in especial premature infants. Term infants showing progressive improvement on the respiratory distress within the first 4–6 h of life did not have a chest X-ray. The decision about the therapy with surfactant replacement was based on the signs of respiratory distress, the progressive oxygen requirement and chest X-ray findings. The therapy with prophylactic exogenous surfactant for preterm infants is not a routinely practice in our NICU. Initial continuous positive airway pressure has been used in the delivery room for premature neonates showing early signs of respiratory distress. If started in the delivery room, it may help the newborn to establish functional residual capacity, stabilize the chest wall and reduce the air resistance [18]. In order to classify the subjective and GSH method for evaluating lung maturity, the categories of correct and incorrect were used. For the mature newborns, if the subjective evaluation showed a hyperechogenic lung and the newborn did not need ventilation support, and the lung GSH presented a higher mean than the hepatic GSH, then the classification was correct. For the mature newborns, if the subjective evaluation showed an isoechogenic or hypoechogenic lung and the newborn was mature regarding the lungs and the lung GSH presented a mean lower than the hepatic GSH, then the classification was incorrect. For the immature newborns, if the subjective evaluation showed an
DOI: 10.3109/14767058.2014.927862
Fetal lung maturity using gray-scale histogram
3
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Ryerson University on 12/03/14 For personal use only.
Figure 1. Mean sagittal plane of the fetal trunk showing lung and hepatic areas, with channel 1 positioned in the lung area and channel 2 in the hepatic area and automatic calculation of gray-scale histograms.
isoechogenic or hypoechogenic lung and the newborn needed ventilatory support and the lung GSH presented a mean lower than the hepatic GSH, the classification was correct. For the immature newborns, if the subjective evaluation showed a hyperechogenic lung and the newborn needed ventilatory support and the lung GSH presented a mean higher than the hepatic GSH, the classification was incorrect. The data were transferred to a spreadsheet in the Excel 2007 software (Microsoft Corp., Redmond, WA) and were analyzed using the SPSS software, version 15.0 for Windows (SPSS Inc., Chicago, IL). In order to evaluate the correlation of hepatic and lung echogenicity between each other using GSH, as well as gestational age, Pearson’s correlation coefficient (r) was used. In order to compare the subjective and GSH methods, in terms of correct and incorrect, the McNemar test was used. In order to compare the lung and liver GSH measurements, as well as the lung/liver relationship, in the groups with and without respiratory distress, the paired Student’s t-test was used. For the subjective and GSH methods, the accuracy, specificity, sensitivity and positive and negative predictive values were calculated for predicting respiratory distress and lung maturity. In all analyses, the significance level used was p50.005.
Results A total of 77 single pregnancies were evaluated, of which 11 (14.3%) were between 28 and 35 + 6 weeks and 66 (85.7%) were 36 weeks. The mean gestational age was 37.06 (ranging from 28 to 40 weeks). The mean birth weight was 2847 g (ranging from 1258 to 3915 g). None of the newborns developed RDS, but 20 newborns (25.9%) developed some degree of respiratory distress, which was classified as transient tachypnea of the newborn. A direct and significant correlation was observed between the means for lung and hepatic echogenicity using GSH
(r ¼ 0.75; p50.001). A non-significant positive correlation was observed between the mean lung echogenicity using GSH and the gestational age (r ¼ 0.13; p ¼ 0.258). A nonsignificant negative correlation was observed between the mean hepatic echogenicity using GSH and the gestational age (r ¼ –0.11; p ¼ 0.361) (Figure 2). The subjective evaluation classified 41 cases (53.2%) correctly and 36 (46.8%) incorrectly. The evaluation using GSH classified 58 (75.3%) correctly and 19 (24.7%) incorrectly. In comparing the two evaluations, 26 (33.8%) of the correct classifications from GSH corresponded to incorrect subjective evaluations; and nine (11.7%) correct classifications in the subjective evaluation corresponded to incorrect GSH evaluations. Therefore, the discordance corresponded to 35 (45.5%) of the total. A statistically significant discordance (p ¼ 0.006) was observed between the two evaluations and, consequently, there was a greater proportion of correct classifications in the evaluation using GSH (Table 1). In the group from 28 to 35 + 6 weeks, nine newborns (81.8%) developed transient tachypnea of the newborn. This group consisted of 22 cases, among which 14 (63.6%) were classified correctly and 8 (36.4%) incorrectly using the subjective technique; while GSH presented 9 correct classifications (40.9%) and 13 incorrect classifications (59.1%). There was no statistical difference between the two techniques (p ¼ 0.180). Taking one as the cutoff value for the relationship between the means for lung/hepatic echogenicity, GSH presented sensitivity, specificity and accuracy for predicting respiratory distress of 61.9%, 89.1% and 81.6%, respectively (Table 2). In the group 36 weeks, nine newborns (81.8%) developed transitory tachypnea of the newborn. This group consisted of 55 patients, with 27 correct classifications (49.1%) and 28 incorrect classifications (50.9%) through the subjective technique; while GSH presented 49 correct
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Ryerson University on 12/03/14 For personal use only.
4
A. P. A. Beck et al.
J Matern Fetal Neonatal Med, Early Online: 1–6
Figure 2. (a) Scatter plot of correlation between the means for the lung and liver echogenicity through GSH; (b) scatter plot of correlation between mean lung echogenicity and gestational age through GSH; (c) scatter plot of correlation between mean liver echogenicity and gestational age through GSH. GSH: gray-scale histogram.
Table 1. Distribution of the results from the whole sample. GSH evaluation Subjective evaluation Correct Incorrect Total
Correct n (%)
Incorrect n (%)
Total n (%)
32 (41.6) 26 (33.8) 58 (75.3)
9 (11.7) 10 (13.0) 19 (24.7)
41 (53.2) 36 (46.8) 77 (100.0)
McNemar test: p ¼ 0.006*; n ¼ absolute number; GSH ¼ gray-scale histogram. *Statistical significance.
respiratory distress, showing a statistical difference for lung echogenicity (p ¼ 0.005). Table 4 presents the relationship of the means for lung and hepatic echogenicity between the groups with and without respiratory distress, showing a statistically significant difference (p ¼ 0.002). Taking one as the cutoff for the relationship of the means for lung/hepatic echogenicity, GSH presented sensitivity, specificity and accuracy for predicting lung maturity of 89.1%, 40.9% and 75.3%, respectively (Table 5).
Discussion Table 2. Comparative results from the methods for predicting respiratory distress. Test Subjective GSH
Sensitivity Specificity PPV NPV PLR NLR Accuracy % 63.6 61.9
% 52.7 89.1
% % 35 78.4 68.4 86
1.3 5.7
0.7 0.4
% 55.8 81.6
GSH, gray-scale histogram; PPV, positive predictive value; NPV, negative predictive value; PLR, positive likelihood ratio; and NLR, negative likelihood ratio.
classifications (89.1%) and six incorrect classifications (10.9%). There was a statistical difference between the two techniques (p50.001). Table 3 presents the means for lung and hepatic echogenicity through GSH in the groups with and without
The most accurate diagnosis regarding fetal lung maturity is always made after birth. Evaluation of immaturity, however, generates greater difficulty because respiratory distress may result from various factors, thus complicating correct diagnosis. Today, the greater number of elective interventions in pregnancies has created a growing need for evaluation of fetal lung maturity [6,7]. However, for this situation, there has been a lack of any effective, harmless and easily reproducible method. The increase in lung echogenicity over the course of pregnancy occurs because of the great proliferation of capillaries, which produce a greater number of interfaces, called the canalicular phase. The hepatic tissue, on the other hand, undergoes a linear increase in its echogenicity, without great changes, thus favoring, in this case, studies on the lung/ liver relationship. However, human vision has difficulty in
Fetal lung maturity using gray-scale histogram
DOI: 10.3109/14767058.2014.927862
Table 3. Distribution of the means for lung and hepatic echogenicity using the gray-scale histogram in both groups. Group Without RD (n ¼ 57)
With RD (n ¼ 20)
Student’s t test
Mean echogenicity–lung Mean ± standard deviation Minimum–maximum
28.0 ± 6.0 17.2–42.1
23.2 ± 7.3 12.1–41.9
p ¼ 0.005*
Mean echogenicity–liver Mean ± standard deviation Minimum–maximum
23.0 ± 6.4 12.7–37.0
22.5 ± 7.2 11.3–35.2
p ¼ 0.760
Parameter
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Ryerson University on 12/03/14 For personal use only.
RD: respiratory distress. *Statistical significance. Table 4. Distribution of the relation between the means for lung and hepatic echogenicity using the gray scale histogram in both groups. Group Lung/liver ratio–mean echogenicity
Without RD (n ¼ 57)
With RD (n ¼ 20)
Student’s t test
Mean ± standard-deviation Minimum–maximum
1.26 ± 0.25 0.69–1.89
1.05 ± 0.20 0.71–1.49
p ¼ 0.002*
RD: respiratory distress. *Statistical significance.
Table 5. Comparative results from the methods for predicting lung maturity. Test Subjective GSH
Sensitivity Specificity PPV NPV PLR NLR Accuracy % % % % % 52.7 89.1
63.6 40.9
78.4 35.0 79.0 60.0
1.5 1.5
0.7 0.3
55.8 75.3
GSH, gray-scale histogram; PPV, positive predictive value; NPV, negative predictive value; PLR, positive likelihood ratio; and NLR, negative likelihood ratio.
differentiating between the various shades of gray [19], which justifies the need for objective evaluation. The histogram is the most commonly used type of graph, and this shows the distribution of frequencies for continuous or discrete data. In the US field, it has always been used to calibrate gray scales. However, it could aid in determining the echotexture of tissues. Because of the easy application and reproducibility of GSH, its use in other tissues such as caesarian scars [20] and the uterine cervix during pregnancy [21] has been described. The technique to obtain the GSH is easy and reproducible [20,21], and in our experience is necessary performing few US exams to obtain the ideal expertise. Since there is no consensus in the literature regarding the best method for studying lung maturity through US [22–24], two types of easily performed evaluation were compared in this study: subjective analysis and analysis through GSH. In our sample, no cases of RDS were found. We believe that this was due to the use of antenatal steroids in pregnancies less than 34 weeks, which is routine in our service. This practice has been shown to be effective in reducing the risk of RDS. Respiratory distress was found in 20 newborns, mostly from the group of mothers with gestations
5
of less than 36 weeks, which were already expected to present some degree of lung immaturity. In a pioneering study on the evaluation of lung and hepatic echogenicity through GSH, Maeda et al. [16] evaluated 52 normal fetuses between 24 and 38 weeks and observed that lung echogenicity was greater between 30 and 38 weeks than between 24 and 29 weeks, but with no difference in hepatic echogenicity. The ratio of the means for lung/hepatic echogenicity was less than one between 24 and 29 weeks and greater than one between 30 and 35 weeks. However, these authors did not correlate lung echogenicity with development of respiratory distress among newborns. In a more recent study, Serizawa and Maeda [25] evaluated predictions of lung maturity through GSH among 22 newborns with respiratory distress and 25 newborns with mature lungs. They observed values for lung echogenicity, hepatic echogenicity and the lung/hepatic echogenicity ratio in the group with respiratory distress of 32.21%, 37.03% and 0.865, respectively. In the visual comparison of this study, the means were lower. However, the different gestational period evaluated in this study (28–40 weeks) and the fact that we did not exclude pregnant women who received corticoid at a gestational age younger than 34 weeks need to be emphasized. In this study, the evaluation through GSH obtained better results than the subjective evaluation, especially in the group in which maturity was already expected ( 36 weeks), due to the aforementioned difficulty in distinguishing between subtle differences in shades of gray [19]. GSH was effective for predicting lung immaturity, especially if the relationship between the means for lung and liver echogenicity was used, without requiring any other propaedeutic resource. In the group in which immaturity was expected, GSH produced specificity of 89.1% and sensitivity of 61.9%. One point that should be noted is that this related to a specific test, thus demonstrating its importance concerning decision-making in weighing up the risk of prematurity in situations of elective resolution of pregnancies. In summary, evaluation of fetal lung maturity through GSH was more effective than the subjective method for predicting respiratory distress in newborns.
Declaration of interest The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.
References 1. Burri PH. Structural aspects of postnatal lung development: alveolar formation and growth. Biol Neonate 2006;89:313–22. 2. Hillman NH, Kallapur SG, Jobe AH. Physiology of transition from intrauterine to extrauterine life. Clin Perinatol 2012;39:769–83. 3. Gleason CA, Devaskar SU. Avery’s diseases of the newborn. 9th ed. Philadelphia (PA): Elservier Saunders; 2012. 4. Jobe AH. What is RDS in 2012? Early Hum Dev 2012;88:S42–4. 5. Rubaltelli FF, Dani C, Reali MF, et al. Acute neonatal respiratory distress in Italy: a one-year prospective study. Acta Paediatr 1998; 87:1261–8. 6. Hansen AK, Wisborg K, Uldbjerg N, Henriksen TB. Elective caesarean section and respiratory morbidity in the term and nearterm neonate. Acta Obstet Gynecol Scand 2007;86:389–94.
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Ryerson University on 12/03/14 For personal use only.
6
A. P. A. Beck et al.
7. Hansen AK, Wisborg K, Uldbjerg N, Henriksen TB. Risk of respiratory morbidity in term infants delivered by elective caesarean section: cohort study. Br Med J 2008;336:85–7. 8. Ashwood ER, Palmer SE, Taylor BA, Pangree S. Lamellar body counts for rapid maturity testing. Obstet Gynecol 1993;81: 619–26. 9. Gluck MD, Kulovich MV, Borer RC, et al. Diagnosis of the respiratory distress syndrome by amniocentesis. Am J Obstet Gynecol 1971;109:440–5. 10. Hallman M, Gluck L. Phosphatidilglicerol in lung surfactant II, subcelular distribution and mechanism of biosynthesis in vitro. Biochim Biophys Acta 1975;409:172–91. 11. Luo G, Norwitz E. Revisiting amniocentesis for fetal lung maturity after 36 weeks gestation. Obstet Gynecol 2008;1:61–8. 12. Cayea PD, Grant DC, Doubilet PM, Johns TB. Prediction of fetal lung maturity: inaccuracy of study using conventional ultrasound instruments. Radiology 1985;155:473–5. 13. Feingold M, James SR, Cetrulo CL, Koza D. Fetal lung to liver reflective ratio and lung maturity. J Clin Ultrasound 1987;15: 384–7. 14. Liu J. Lung ultrasonography for the diagnosis of neonatal lung disease. J Matern Fetal Neonatal Med 2014;27:856–61. 15. Allison JW, Barr LL, Massoth RJ, et al. Understanding the process of quantitative ultrasonic tissue characterization. Radiographics 1994;14:1009–108. 16. Maeda K, Utsu M, Yamamoto N, Serizawa M. Echogenicity of fetal lung and liver quantified by the grey-level histogram width. Ultrasound Med Biol 1999;25:201–8.
J Matern Fetal Neonatal Med, Early Online: 1–6
17. Hadlock FP, Harrist RB, Martinez-Poyer J. In utero analysis of fetal growth: a sonographic weight standard. Radiology 1991;181: 129–33. 18. de Klerk AM, de Klerk RK. Use of continuous positive airway pressure in preterm infants: comments and experience from New Zealand. Pediatrics 2001;108:761–3. 19. Julesz B, Gilbert EM, Frish HL. Inability of humans to discriminate between visual features that agree in second-order statistics – revisited. Perception 1973;2:391–405. 20. Santos Filho Ode O, Nardozza LM, Arau´jo Junior E, et al. [Cesarean uterine scar evaluation by the grey-level histogram]. Rev Assoc Med Bras 2010;56:99–102. 21. Furtado MR, Pires CR, Araujo Ju´nior E, et al. Transvaginal grey scale histogram of the cervix at 20–25 weeks of pregnancy. Aust N Z J Obstet Gynaecol 2010;50:444–9. 22. La Torre R, Cosmi E, Anceschi MH, et al. Preliminary report on a new and noninvasive method for the assessment of fetal lung maturity. J Perinat Med 2003;31:431–4. 23. Piazze JJ, Anceschi MM, Picone G, et al. Association between maternal-fetal Doppler velocimetry and fetal lung maturity. J Perinat Med 2003;31:484–8. 24. Cobo T, Bonet-Carne E, Martı´nez-Terro´n M, et al. Feasibility and reproducibility of fetal lung texture analysis by Automatic Quantitative Ultrasound Analysis and correlation with gestational age. Fetal Diagn Ther 2012;31:230–6. 25. Serizawa M, Maeda K. Noninvasive fetal lung maturity prediction based on ultrasonic gray level histogram width. Ultrasound Med Biol 2010;36:1998–2003.
