3D Ultrasound Evaluation of the Fetal Ear – Comparison of an xMatrix Probe with a Conventional Mechanical Probe

Abstract

Purpose New 3 D technologies like xMatrix probes promise superiority over conventional mechanical probes and may allow a more detailed and time-saving prenatal diagnosis. In a comparison study we evaluate fetal ears. The aim of our study was to compare the following aspects of both techniques: (1) ultrasound detail resolution, (2) raw data acquisition time (AT) and (3) influence of covariates.

Materials and Methods 3 D raw data volumes of the fetal ear were collected with the V6 – 2 (V6) and with the xMatrix (X6) probe and were stored after offline customization to a single picture. Two observers scored these images independently. Furthermore, the 3 D raw data acquisition time (AT) was recorded. Concordance between observers, maternal age, body mass index (BMI), weeks of gestation and location of the placenta were evaluated.

Results Data volumes of 103 patients were analyzed. The X6 detected anatomic structures like the scapha (p = 0.0146), fossa triangularis (p = 0.0075) and cymba conchae (p = 0.0025) more often. The mean AT of the X6 was shorter compared to the V6 (p < 0.0001). A placenta location in the scanning field increased the AT only for the V6 (p < 0.01). Concordance between observers was higher for the X6 in most cases. Detailed structures were less visible at the end of pregnancy for both devices.

Conclusion The comparison study demonstrated clear advantages of the new xMatrix technology concerning an advanced and fast examination of detailed structures like the fetal ear. The importance of 3 D assessment in cases of fetal ear anomaly should be proven in further studies.

Zusammenfassung

Ziel Neue 3D-Technologien wie xMatrix-Sonden gelten gegenüber herkömmlichen mechanischen Sonden als überlegen, da sie eine detailliertere und zeitsparendere Untersuchung ermöglichen sollen. Für eine Vergleichsstudie wählten wir die fetalen Ohren aus. Zur Evaluation der Technologieunterschiede wurden folgende Aspekte geprüft: (1) die Darstellbarkeit von Detailstrukturen, (2) die benötigte Zeit zur Datenakquise und (3) der Einfluss von Kofaktoren.

Material und Methoden Es wurden 3D-Datensätze fetaler Ohren mit der V6 – 2 (V6) und mit der xMatrix (X6) Sonde aufgenommen und nach offline Bearbeitung als Einzelbild gespeichert. Diese Bilder wurden von zwei unabhängigen Untersuchern bewertet. Zusätzlich wurde die Dauer der Datenakquise (AT) erfasst. Die Übereinstimmung zwischen den Untersuchern und der Einfluss von maternalem Alter, Body-Mass-Index (BMI), Schwangerschaftsalter und Plazentalage wurden untersucht.

Ergebnisse Es konnten Datensätze von 103 Patientinnen ausgewertet werden. Die X6 stellte die Scapha (p = 0,0146), die Fossa triangularis (p = 0,0075) und die Cymba conchae (p = 0,0025) häufiger dar. Die mittlere AT der X6 ist kürzer als die der V6 (p < 0,0001). Die Übereinstimmung zwischen den Untersuchern war in fast allen Fällen für die X6 größer. Die AT verlängerte sich nur bei der V6, falls die Plazenta im Ultraschallfeld lag (p < 0,01). Mit fortgeschrittenem Schwangerschaftsalter verminderte sich die Detailauflösung beider Sonden.

Schlussfolgerung Die Vergleichsstudie zeigt klare Vorteile für die xMatrix-Technologie in Bezug auf die detaillierte und zeitsparende Darstellung der fetalen Ohren. Zukünftige Studien sollten den Stellenwert der 3D-Untersuchung bei Feten mit Ohrfehlbildungen untersuchen.

Key words

ear - 3 D ultrasound - fetus - obstetrics

Introduction

In some cases 3 D ultrasound can provide additional information to conventional 2 D technology for the evaluation of the fetal face and other surface structures [1]. New technologies such as xMatrix still promise superiority to conventional 3 D techniques like the mechanical probe. Whether this is true has not been investigated so far in a clinical study. We chose evaluation of the fetal ears as a suitable example for a comparison study because of their complex anatomy and their use as a marker in prenatal syndrome diagnosis.

In prenatal diagnosis the evaluation of the fetal ear with two-dimensional (2 D) ultrasound is part of advanced ultrasound screening [2] [3] [4] [5]. The fetal ear has a complex anatomic structure with different planes, angles and reliefs ([Fig. 1]).

During the 13^th and 14^th week of gestation, mostly the facial structures are sufficiently developed and prenatal diagnosis is achievable [6] [7].

The overall incidence of ear malformations is about 1:3800 newborns [8]. Concerning the outer ear, the incidence has been reported to be between 1:6000 [9] to 1:6830 newborns [10]. In cases of fetuses with trisomy 21 or other chromosomal abnormalities, the fetal ear size, position, orientation and location are often abnormal [2] [11] [12]. Furthermore, malformations of the outer ear are specifically indicative for several other syndromes [8].

In the past, prenatal diagnosis of fetal ears was mostly performed by 2 D ultrasound with detection rates of fetal ears of about 88 % [3]. Parameters like ear size and width were measured [2] [5] [13]. Measurement of ear length might be a useful predictor of aneuploidy [5] [11] [13]. Birnholz et al. found a specificity of 100 % and a sensitivity of 83 % for aneuploidy [2]. Similarly, Lettieri et al. published a specificity of 92 % and sensitivity of 71 % [3]. Short ears were frequently described in all cases of trisomy 13 or 18, but only in about half of those with trisomy 21 [2]. Gorlin et al. detected fetuses with ear anomalies due to trisomy 13 or 18 in just 80 – 90 % of all cases and fetuses with trisomy 21 in about 60 % [14]. Short ears with a length smaller than 1.5 standard deviations [2] or below the 10^th percentile are frequently associated with chromosomal syndromes [3].

2 D ultrasound allows assessment of the fetal ear in only one plane [2]. In contrast, recent three-dimensional (3 D) ultrasound technology enables the evaluation of detailed anatomy and the position of the fetal ear in multiple planes like the triplanar plane, the orthogonal display or the surface display. This allows additional visualization of certain planes that can be difficult to obtain by conventional imaging and has the advantage of visualizing complex structures in one image. In general, 3 D has been shown to be useful in the diagnosis of facial anomalies, facial clefts, neural tube defects, skeletal malformations and also in detecting defects of the brain, spine, heart, the skeleton and other areas [15].

As far as we know, only three studies, published in the last 15 years, have focused on the evaluation of fetal ears with 3 D ultrasound technology [16] [17] [18]. These studies, however, only used conventional techniques. In contrast, modern 3 D ultrasound systems offer higher image resolution and faster data acquisition time. One of the most recent developments is the xMatrix technology. Whether this xMatrix system is superior to conventional mechanical 3 D ultrasound technology has never been evaluated in a clinical trial.

The prenatal evaluation of the fetal ear was chosen as a suitable example to compare the xMatrix and the conventional technology in a clinical setting. Therefore, the aim of our study was to compare the following aspects of both technologies: (1) ultrasound detail resolution, (2) raw data acquisition time (AT) and (3) influence of covariates.

Methods

Study population

This cross-sectional controlled comparison study was performed in the Department of Obstetrics and Gynecology at the University Münster. Between July 2013 and July 2014, we collected 3 D raw data volumes of the fetal ear in unselected pregnant women between 18 and 40 weeks of gestation. We excluded women with an age under 18, pregnancies with oligohydramnios or fetal malformations.

Informed consent was obtained from all patients at the time of enrollment. The study was designed according to the Declaration of Helsinki and was approved by the institutional ethics board.

3 D raw data volume acquisition

All examinations were performed with an iU22 ultrasound system equipped with a 6 MHz 3 D broadband volume curved array transducer V6 – 2 (V6) and with the 6 MHz 3 D xMatrix transducer X6 – 1 (X6) (Philips Medical Systems, Bothell, WA, USA). All observers (R.S., M.K.F., M.M.) who performed the ultrasound examinations are experienced prenatal diagnosis specialists.

The examination started with a routine prenatal ultrasound screening of the fetus. After the prenatal screening, the observer started the 3 D raw data volume acquisition. Therefore, the fetal face should be in a suitable position to visualize one fetal ear. The 3 D render box was placed around the fetal ear and was customized in order to obtain a volume of the whole fetal head with the ear in the center of the box. For all 3 D raw data volume acquisition, we chose the same factory preset setting “general obstetrics”. For each 3 D probe, we saved two different 3 D raw data volumes.

Because acquisition time was not stored automatically, we recorded two blank ultrasound pictures separately for each probe, one at the beginning and one at the end of 3 D raw data acquisition. The local time was registered automatically in those pictures. The time interval between the two recorded pictures was defined as the data acquisition time (AT) for each probe.

3 D raw data volume analysis

The 3 D raw data volumes were evaluated offline with QLab 9 (Quantification Software 9, Philips Medical Systems, Bothell, WA, USA). QLab 9 software allows the manipulation of 3 D raw data volumes offline and the storing of customized single pictures. We stored single images of the fetal ear with optimal resolution and the best view of each 3 D data set separately for both probes.

To evaluate these pictures of the fetal ear and to compare them between the probes, we developed visual analog scales (VAS) for different criteria. We distinguished between “basic patterns” and “advanced patterns”. The “basic patterns” consisted of the visibility of the ear, the presentation of the outer contours and the possibility of determining the correct axis of the fetal ear to the fetal eye (“0” (no) and “1” (yes)).

We further rated the anatomic detail presentation of the fetal ear with a range from “0” (no ear visible) to “3” (detailed presentation of details). [Fig. 2] presents the visual scoring scale in detail.

Fig. 2 Scoring range for anatomic details of the fetal ear. Left ear a) “score 1” with just the outlines of the ear, ear in the middle b) “score 2” with more details of the structure, right ear c) “score 3” approximate anatomic structure, with details like the scapha, fossa triangularis and cymba conchae.

We defined the “advanced patterns” scale by the visibility of three detailed anatomic structures of the ear, like the scapha, the fossa triangularis and the cymba conchae. The presentation of these structures was scored with “1” (yes) or with “0” (no). During the evaluation of the pictures and the scoring of the patterns, the first observer (K.B.) was not blinded to the patient`s data and to the information about the probe used.

In contrast to that, to prove the interobserver concordance of the data analysis, a second observer (R.S.) evaluated the pictures blinded to patient data and information about the probe used.

The assessment of covariates

At the beginning of the routine ultrasound screening, the patients were interviewed in a standardized way to collect the following covariates:

“Maternal age” was described in years at the time of enrollment.

“Maternal BMI” (body mass index) was measured and calculated in weight (kg) / height² (m²) at the time of presentation.

“Weeks of gestation” was determined as weeks since the first day of the last menstruation.

“Location of placenta” was scored as “1” if the location of the placenta was in the scanning field of the ultrasound probe and as “0” if it was not.

Statistics

Statistical analysis was performed using SAS 9.2 (copyright© 2002 – 2008 by SAS Institute INC., Cary, NC, USA). Percentages, means and standard deviations were used for description. Sign test was used for the comparison between the probes for scores of the nominal scales and paired t-test for the comparison of times. A two-sided p-value< 0.05 was considered statistically significant. Kappa coefficients were evaluated to compare the data examination rating between two observers. Following the suggestions of Grouven et al. [19], a coefficient between 0.61 and 0.80 indicated “good” agreement and a coefficient greater than 0.80 indicated “very good” agreement.

The influence of the covariates “maternal age”, “maternal BMI”, “week of gestation” and “location of placenta” on the AT, the visibility of the ear (as an example of basic patterns) and the visibility of the cymba conchae (as an example of advanced patterns) were investigated using multiple linear regression for the outcome “time” and multiple logistic regression for the other outcomes with nominal scales.

Results

103 cases were included in this study. The mean (±SD) maternal age was 31 (± 5) years. All pregnant women were scanned between 18 and 41 weeks of gestation, with a mean of 27.5 (± 5.7) weeks of gestation. The mean maternal BMI was 25.8 (± 5.7) kg/m². In 41 (39.8 %) cases, the placenta was located in the scanning field of the probe. [Fig. 3] demonstrates three sample pictures of the comparison of the fetal ear from both probes.

Fig. 3 a Example of 3 D images of the fetal ear from both probes, left site V6 probe, right site X6 probe: 36th week of gestation. b Example of 3 D images of the fetal ear from both probes, left site V6 probe, right site X6 probe: 31st week of gestation. c Example of 3 D images of the fetal ear from both probes, left site V6 probe, right site X6 probe: 34th week of gestation.

[Table 1] shows the percentages of detected patterns for both probes as determined by observer 1 (K.B.). The fetal ear could be detected in most cases (V6 82.5 %, X6 76.6 %) and nearly the same was true for the contours of the ear (V6 81.6 %, X6 76.6 %). The possibility to determine the correct position of the ear in relation to the eye axis was noticed in about half of all cases (V6 48.5 %, X6 52.3 %). No significant differences between both devices were noticed for “basic patterns”. Anatomical details were detectable significantly more often with the X6 probe than with the V6 probe. The three “advanced patterns” scapha, fossa triangularis and cymba conchae could be detected significantly more often with the X6 probe than with the V6 probe (scapha (p = 0.0146), fossa triangularis (p = 0.0075) and cymba conchae (p = 0.0025)) ([Table 1]).

Table 1 Comparison of the quality scores and the acquisition time between the V6 probe and the X6 probe for observer 1 (K.B.).

n = 103
criteria	probe V6	probe X6	p-value (sign-test)
ear visible	82.5 %	76.6 %	0.4545
contour visible	81.6 %	78.6 %	0.6291
location of the ear	48.5 %	52.3 %	0.5847
discrimination of anatomy scored “3”	8.7 %	27.1 %	0.0052[1]
scapha	21.4 %	34.0 %	0.0146
fossa triangularis	7.8 %	19.4 %	0.0075
cymba conchae	12.6 %	28.2 %	0.0025
	probe V6	probe X6	p-value (paired t-test)
acquisition time (sec)	72.5 (42.3)	51.9 (30.3)	< 0.001

Data are shown as mean (SD) or as percentage (%).

¹ all scores considered.

Observer 2 scored with similar results ([Table 2]). For the “basic pattern” and for anatomical details he likewise depicted no advantage for one probe. For the “advanced patterns” fossa triangulares (p = 0.0004) and cymba conchae (p = 0.0042), observer 2 also described significant differences in detectability between the two probes. The only exception was the detection of the scapha as a criteria of the “advanced patterns” where he scored no difference (p = 0.6900).

Table 2 Comparison of the quality scores between the V6 probe and the X6 probe for observer 2 (R.S.).

n = 87
criteria	probe V6	probe X6	p-value (sign-test)
ear visible	77.0 %	80.5 %	0.6476
contours visible	66.7 %	75.9 %	0.1338
location of the ear	36.7 %	41.4 %	0.5235
discrimination of anatomy scored “3”	9.2 %	24.1 %	0.0018[1]
scapha	26.4 %	29.9 %	0.6900
fossa triangularis	6.9 %	25.3 %	0.0004
cymba conchae	4.6 %	18.4 %	0.0042

Data are shown as percentage (%).

¹ all scores considered.

The concordance between both observers was much better for the X6 probe than for the V6 probe, with the exception of the detection of the fossa triangularis. The results of the kappa coefficient for the X6 probe were good to very good in each case, whereas it was only moderate for the V6 probe ([Table 3]).

Table 3 Interrater agreement (kappa coefficients).

criteria	probe V6 (n = 89)[1]	probe X6 (n = 93)1
ear visible	0.7634	0.8241
contours visible	0.5499	0.8385
location of the ear	0.5890	0.6189
discrimination of anatomy	0.5197	0.6124
scapha	0.5649	0.7098
fossa triangularis	0.6904	0.6448
cymba conchales	0.4315	0.7395

¹ 87 of these tests were done with both devices.

The data acquisition time of the X6 probe was significantly shorter in comparison to the V6 probe (51.9 ± 30.4 vs. 72.4 ± 42.3 sec; p < 0.001) ([Table 1]).

The covariates maternal age and BMI had no influence on the AT for both transducers. The placenta position had an effect on the AT only for the V6 probe (p < 0.01). The mean time was 22.7 seconds longer if the placenta was within the scanning field. The detectability of the “advanced patterns” decreased with higher gestational age for both devices, but slightly more for the X6 probe ([Table 4]).

Discussion

In our study, we demonstrated that the xMatrix probe has higher detail resolution in 3 D pictures of the fetal ear in comparison to a conventional mechanical 3 D probe. Especially the “advanced patterns” of the fetal outer ear like the scapha, the fossa triangularis and the cymba conchae were visible more often with the new xMatrix technology. In addition, the acquisition time (AT) of the xMatrix probe was much shorter and the concordance between the observers was higher for the X6 probe. Concerning the covariates, the location of the placenta only had an influence on the AT for the V6 probe. As the week of gestation increased, the detailed structures were less visible with both devices. The other covariates had no influences on the AT.

When 3 D ultrasound was first introduced, the acceptance was low, because of the long image processing time and the insufficient image quality. However, with the development of computer technology, acceptance has increased. Today most ultrasound equipment manufacturers have incorporated 3 D/4 D technology into their scanning systems and 3 D ultrasound has become generally accepted [15].

In contrast to 2 D ultrasound, 3 D ultrasound technologies offer many benefits. With a variety of display modes, like the gray render modes (multiplanar, surface, transparent or inversion mode) and color render modes (pure color mode or glass body render mode), 3 D ultrasound can produce images that can never be achieved with 2 D technology [15]. Using these different modes of imaging, pictures of malformations can be presented in a more comprehensible manner to colleagues or future parents. Another advantage of 3 D technology is the possibility of digital storage on a personal computer and virtual examinations by reloading these volumes with the same quality. This allows navigating through them at any time at any place using the appropriate software regardless of an ultrasound system.

Studies with 3 D ultrasound and the fetal ear are still rare. These first studies evaluated ear length measurements and ear shapes and developed reference charts for 3 D ultrasound [16], in relation to gestational age [20]. Additionally, ear rotation in fetuses with autosomal trisomy [12] and quantitative measurements of ear length, ear width and ear area [18] in fetuses with chromosomal aberrations were investigated. One study in 1998 described ear morphology (lying axis, orientation and cranial location) and stated its usefulness for prenatal diagnosis and genetic counseling [17]. All of these authors concluded consistently that 3 D ultrasound displayed fetal ear abnormalities with greater accuracy and clarity than 2 D ultrasound.

In addition to these studies, which only evaluated the shape and orientation of the outer ear, we examined the fine anatomic structures of the fetal ear. To the best of our knowledge, there are no other studies published which investigate these structures with 3 D ultrasound. New developments like the xMatrix probe promise superiority to conventional 3 D techniques like the mechanical probe with respect to the detectability of fine structures. Our study is the first to confirm these promises. If these detail structures are affected, cases of syndromes or genetics failures must be investigated in further studies. Up to now, there is no valid parameter for the quantification of detailed structures of the fetal ear in ultrasound.

Our results demonstrate a distinction in quality between 3 D probes due to their different technologies. The xMatrix probe is less affected by the placenta location compared to the V6 probe. Additionally, the observer agreement of the X6 probe is high and is considerably higher compared to the V6 probe. In contrast to a conventional probe, the xMatrix probe generates volumetric imaging with electronic focus in both the azimuthal and elevation directions and steering through the entire 3 D field of view. This could lead to a smaller 3 D slice thickness and enhance the spatial and contrast resolution, including an assessed C-plane like in our study. In comparison to conventional arrays which may have up to a few hundred elements, the xMatrix array operates with over 9000 active elements.

Overall, the greatest benefit of the X6 probe is the enhanced detectability of anatomical details and the faster 3 D volume data acquisition time (AT). Furthermore, in the future, 4 D live scanning could be possible with the xMatrix technology and this could allow fast acquisition and rendering of volume data at high frame rates with acceptable resolution.

The authors think that recent real-time 3 D (4 D) transducer assessment could be more efficient and precise compared to 3 D static volume analysis, because of faster allowing for rapid correction of the gain setting and insinuation angle with the advantage of images with optimal resolution. This should be proven in further studies.

Additionally the xMatrix probe provides better ergonomics with respect to weight and size compared to conventional motorized 4 D transducers.

We only examined “normal” fetal ears, without any malformation during the prenatal examination. A detailed postnatal follow-up was not obtained. This might be a limitation of our study. Further studies could explore whether variations in these detail structures could be a sign for genetic malformation syndromes. Additional follow-up studies could compare pre- and postnatal findings in order to prove the reliability of 3 D examination of fetal ears. In contrast to our study, both ears have to be investigated in the daily routine because in about 70 – 90 % cases malformations are only unilateral [8]. Therefore, information of one unsuspicious ear would not exclude a malformation or syndrome. In 58 – 61 % of cases, the right side is concerned.

Further limitations of our study were that the scoring of the fetal ears was just blinded for observer 2 (R.S.) and that an evaluation of the interobserver variability of 3 D data acquisition was not performed. However, the scoring results were quite similar between the two observers, thereby suggesting independence of the results from blinding. Additionally, the concordance in data examination between the observers was quite high.

Conclusion

The new 3 D xMatrix technology is superior to the conventional 3 D mechanical probe. Advanced and fast examination of detailed structures like the fetal ear is feasible with recent 3 D technology. Assessment of fetal ears could generate important additional information in prenatal diagnosis und should be further evaluated in cases with an ear anomaly.

Fig. 1 Sonographic image of the anatomic structures of the outer ear.

Table 4 Influence of the covariates on acquisition time, visibility of the ear and visibility of cymba conchae for V6 and X6. Results of multiple regression analysis.

	acquisition time		visibility of the ear		visibility of cymba
V6	sec/unit	p-value	odds ratio/unit	p-value	odds ratio/unit	p-value
age unit:1 year	1.47	0.11	1.02	0.79	1.07	0.36
BMI unit: 1 point	0.13	0.86	1.04	0.45	0.08	0.71
gestation unit: 1 week	0.13	0.86	0.98	0.66	0.85	0.012
location of placenta in scanning field	22.71	0.01	1.87	0.26	1.21	0.78
X6
age unit:1 year	0.39	0.59	1.00	0.96	0.90	0.12
BMI unit: 1 point	–0.94	0.58	1.02	0.73	1.04	0.45
gestation unit: 1 week	0.04	0.94	0.99	0.83	0.78	< 0.001
location of placenta in scanning field	– 2.85	0.67	1.61	0.35	1.74	0.34

No conflict of interest has been declared by the author(s).

• References

• 1 Merz E, Abramowicz JS, Baba K. et al. 3D imaging of the fetal face – Recommendations from the International 3D Focus Group. Ultraschall in Med 2012; 33: 175-182
• 2 Birnholz JC, Farrell EE. Fetal ear length. Pediatrics 1988; 81: 555-558
• 3 Lettieri L, Rodis JF, Vintzileos AM. et al. Ear length in second-trimester aneuploid fetuses. Obstet Gynecol 1993; 81: 57-60
• 4 Shimizu T, Salvador L, Hughes-Benzie R. et al. The role of reduced ear size in the prenatal detection of chromosomal abnormalities. Prenat Diagn 1997; 17: 545-549
• 5 Awwad JT, Azar GB, Karam KS. et al. Ear length: A potenzial sonografic marker for Down syndrome. Int J Gynaecol Obstet 1994; 44: 233-238
• 6 Hill MA. Early human development. Clin Obstet Gynecol 2007; 50: 2-9
• 7 Andresen C, Matias A, Merz E. Fetal Face: The whole picture. Ultraschall in Med 2012; 33: 431-440
• 8 Bartel-Friedrich S, Wulke C. Classification and diagnosis of ear malformations. Laryngorhinootologie 2007; 86: 77-95
• 9 Brent B. The pediatrician's role in caring for patients with congenital microtia and atresia. Pediatr Ann 1999; 28: 374-383
• 10 Conway H, Wagner KJ. Congenital anomalies of the head and neck. Plast Reconstr Surg 1965; 36: 71-79
• 11 Sforza C, Dellavia C, Zanotti G. et al. Soft tissue facial areas and volumes in subjects with Down syndrome. Am J Med Genet A 2004; 130A: 234-239
• 12 Ginsberg NA, Cohen L, Dungan JS. et al. 3-D ultrasound of the fetal ear and fetal autosomal trisomies: a pilot study of a new screening protocol. Prenat Diagn 2011; 31: 311-314
• 13 Chitkara U, Lee L, Oehlert JW. et al. Fetal ear length measurement: a useful predictor of aneuploidy?. Ultrasound Obstet Gynecol 2002; 19: 131-135
• 14 Gorlin RJ, Cohen MM, Levin LS. Chromosomal syndromes: common and/or well-known syndromes. In: Gorlin RJ, Cohen MM, Levin LS. (eds.) Syndromes of the Head and Neck. 3rd edn. Oxford: Oxford Univ Press; 1990: 35-45
• 15 Merz E, Abramowicz JS. 3D/4D ultrasound in prenatal diagnosis: Is it time for routine use?. Clin Obstet Gynecol 2012; 55: 336-351
• 16 Hatanaka A, Rolo L, Mattar R. et al. Reference intervals for fetal ear length between 19 and 24 weeks of pregnancy on 3-dimensional sonography. J Ultrasound Med 2011; 30: 1185-1190
• 17 Shih J, Shyu M, Lee C. et al. Antenatal depiction of the fetal ear with three-dimensional ultrasonography. Obstet Gynecol 1998; 91: 500-505
• 18 Chang C, Chang F, Yu C. et al. Fetal ear assessment and prenatal detection of aneuploidy by the quantitative three-dimensional ultrasonography. Ultrasound Med Biol 2000; 26: 743-749
• 19 Grouven U, Bender R, Ziegler A. et al. Der Kappa- Koeffizient. Dtsch Med Wochenschr 2007; 132: 65-68
• 20 Hatanaka AR, Rolo LC, Mattar R. et al. P28.15: Analysis of the fetal ear position by three dimensional rendering mode using a novel method: the ear index-preliminary results. Ultrasound Obstet Gynecol 2012; 40: 171-310