Original Article

Definition

Microcephaly is a syndrome associated with neurologic abnormalities and mental retardation. The diagnosis has been based on measurement of the head circumference at the level of the base of the skull. With the 3D technology, different thresholds have now been proposed.

Objective

The objectives of this study were to establish reference values for the area, length and volume of fetal skull measurements by traditional multiplanar view and the VOCAL technique and to assess the intra-observer variability of measurements using both methods.

Methods

The fetal brain volumes were acquired between October 2005 and March 2006. All ultrasound examinations were performed transabdominally using a Voluson GE (GE Medical Systems, Milwaukee, WI, USA) by one fetal specialist physician. After routine 2D ultrasound evaluation, the acquisition of 3D datasets was performed as follows. The 3D volume box was adjusted for the size of the skull and the sweep angle was set between 30 and 70 degrees depending on the gestational age. The initial plane of acquisition was the axial plane at the level of the biparietal diameter. The criteria for storage of the 3D volumes were the absence of fetal movements or others artifacts and the presence of the entire skull acquired in the volume. The volume acquisition time was adjusted to obtain the best achievable resolution.

The dataset containing the brain was initially displayed on the screen in the three orthogonal planes. The sagittal view of the brain was displayed in Plane A. Two demarcating arrows were positioned at the outer border of the both parietal bones. The brain volume was measured in 3D using two different techniques: multiplanar mode and VOCAL mode. In the multiplanar mode, three perpendicular views of the fetal brain were displayed simultaneously on the screen. The length of the brain was first measured using the exact midsagittal plane (video 1). A contour of the brain was drawn using the length point just close to the inner border of the skull. The starting point was the junction of the palate and the frontal bone and the end point was the end of the occipital bone.

Video 1: how to measure the brain length.

The area length was obtained using the same multiplanar planes used for calculate the brain length and the same contour except for a line drawn just above the palate and joining the end of the palate and the inferior border of the occipital bone (video 2).

Video 2: how to measure the brain area

VOCAL technique

For rotational-based volume measurements, we used VOCAL software mode (Virtual Organ Computer-aided Analysis, 3D View, GE, KretzTechnik, Zipf, Austria). The dataset containing the fetal skull was initially displayed on the screen in the three orthogonal planes as the same used for the multiplanar method describe above. Using a rotational technique the volume of the brain was obtained by contouring its surface. The plane was rotated around the transverse axis. The contour was defined first automatically and adjusts manually in order to obtain the exact limits of the brain. The rotation step for each contours plane was 30⁰. The contour tracing was carefully performed excluding the skull and until a 180⁰ was completed. A 3D volume model of the brain was obtained and reviewed until all contours have been adequately traced. Brain volume was finally automatically calculated (video 3). All measurements were performed two different times by the same fetal specialist physician.

Video 3: How to measure the brain volume using the VOCAL technique.

The eligibility criteria consisted of: singleton pregnancies; well-defined gestational age according to the know date of the las period and confirmed by a first-trimester scan; normal first trimester anomaly scans, normal fetal anatomy and normal routine obstetrics biometry parameters.

Statistical analysis To asses repeatability (intra-observer variation), Bland-Altman analysis, as well as intra-class correlation coefficients (ICCs) were used. For this analysis, the mean difference between the two methods with 95% CIS and the limit of agreement were determined. For repeatability analyses, a good reliability was defined when the value of the ICC was greater than 0.80.

Results A total of 85 women at 15-33 weeks of gestation fulfilled the entry criteria and evaluated twice for the same operator. The cases above 25 weeks of gestation had to be excluded because there were not enough cases for modeling, altering the fitting process. Reference values Reference values for the area, length and volume of fetal skull measurement by three-dimensional (3D) ultrasound were established. Original data and predicted values for each case obtained from regression analysis using gestational age as the independent variable with lower and upper limits for 95% confidence interval is shown in the table 1 to 3, represents in the respective curves. The 95 confidence interval represents the dispersion of the mean value of the predicted curve. The percentiles 5 and 95 were obtained from a regression model, called quantile regression, chosen due to the sample size.

Table 1 - Area

Table 2 - Length

Table 3 - Volume

Comparison between area, length and volume The coefficients of variation between volumes, area and length were 2,47 for the volume, 4,68 for the area and 2,86 length respectively.

Intra-observer variability The intra-observer correlation coefficients in all cases were higher than 0.9, with significant p-values, which indicates that there is almost perfect agreement between both measurements. The range of ICS was -0.5989 to 1.3783 for brain volume using VOCAL; -3.0726 to -0.9075 for brain length and 5.8245 to -2.9222 for brain area for the multiplanar technique.

In the Bland-Altman plot, the mean difference between paired measurements by the same sonographer was 0.3897 for the brain volume, -1.9900 brain length and -4.3734 for the brain area. When visually examining these plots, we observed that intra-observer differences were evenly distributes above and below zero, suggesting no systematic error within observer.

Discussion The data from this study demonstrate good agreement between the multiplanar mode and the VOCAL for the assessment of the fetal brain. A good repeatability of the measurements was obtained but the best repeatability was for the brain volume. In addition, this study showed that the measurements of the brain calculated using the VOCAL technique has a better accuracy for predict the brain volume.

The objective of the study was to look for a different way to analyze and measure the fetal brain. We are used to measure the fetal brain using only the bi-parietal diameter and the head circumference. However, in cases of microcephaly the fetuses may have a low height of the brain (small endocranium) and these traditional parameters could not be affected leading us to misdiagnoses of microcephaly. And how we could evaluate the totality of the encephalic mass brain if we don’t have available plots to use?

That is the reason we tried to describe different methods to measure the fetal brain and to construct references values for it. And we start the study, all gestational ages were included but we failed to measure the fetal brain in the third trimester of the pregnancy because the advance mineralization of the skull and consequently the shadowing inside the brain structures. So our method was limited to the second trimester of the pregnancy. Until this moment, we still have not tested these measurements in the end of the first trimester. Another issue that still needs to be studied is the inter-observer agreement. And the next step is to study the brain volumes of the fetuses with microcephaly and test the accuracy of our methods.

Conclusions In the present study we have construct the reference ranges of the fetal brain and the volume had the best accuracy in the determination of the gestational age. Further studies using a larger sample data (more patients) and a spanning a wider gestational age are suggested to tighten up the confidence interval. A collaborative study is suggested.