Validation of Placental Vascular Sonobiopsy for Obtaining Representative Placental Vascular Indices by Three-Dimensional Power Doppler Ultrasonography

Article Outline

• Abstract
• 1. Introduction
• 2. Materials and methods
  • 2.1. Study design
  • 2.2. Image acquisition
  • 2.3. Measurement of placental vascular indices by whole placenta evaluation
  • 2.4. Measurement of placental vascular indices by sonobiopsy
  • 2.5. Statistical analysis
• 3. Results
• 4. Discussion
• References
• Copyright

Abstract 

Objective

Placental vascular sonobiopsy has been proposed for obtaining a representative sample of the placental vascular tree when evaluation of the whole placenta is not feasible. We tested the hypothesis that placental vascular indices from sonobiopsy correlate well with those from the entire placenta.

Methods

Three-dimensional power Doppler ultrasound examinations were performed in 120 singleton pregnancies at 11–14 weeks' gestation. The VOCAL™ program was used to calculate placental vascularization index (VI), flow index (FI) and vascularization flow index (VFI) from stored images of each placenta by whole placenta evaluation and placenta vascular sonobiopsy. The mean of each index from four spherical sonobiopsies were compared to those from evaluation of the entire placenta for their degree of correlation and agreement.

Results

The mean VI and VFI from the two techniques were similar (13.9 [95% CI 12.3–15.8] versus 14.3 [95% CI 12.1–17.0], p = 0.62 and 6.1 [95% CI 5.2–7.1] versus 6.1 [95% CI 5.0–7.4], p = 0.93, respectively) and significantly correlated (Pearson's r = 0.70 [95% CI 0.60–0.78, p < 0.001] and r = 0.69 [95% CI 0.58–0.77, p < 0.001], respectively). The mean FI from the two techniques were significantly different (44.5 [95% CI 42.9–46.1] versus 41.3 [95% CI 39.6–43.0], p = 0.001), but correlated (r = 0.59 [95% CI 0.46–0.70, p < 0.001]).

Conclusion

Our findings suggest that placenta vascular indices from sonobiopsy have a good correlation with those from evaluation of the entire placenta. Sonobiopsy may be a valid alternative for evaluation of the placental vascular tree when visualization of the entire placenta is not feasible. Measurements of VI and VFI appear to be more reliable than FI in sonobiopsy specimen.

Keywords: Placenta, Three-dimensional power Doppler ultrasonography, Sonobiospy, Placental vascular indices

Back to Article Outline

1. Introduction 

Abnormal placental development is associated with pregnancy complications such as fetal growth restriction, preterm labor, preeclampsia and stillbirth [1]. Recent advances in ultrasound technology permit quantitative non-invasive assessment of placental vascularization using three-dimensional (3D) imaging combined with power Doppler. Vascular indices within the placenta are calculated from 3D data formed by voxels, the basic information units of volume. These indices measure the vascularization (vascularization index-VI), the intensity of blood flow at the time of the 3D sweep (Flow Index-FI), and both vascularization and blood flow (vascularization flow index-VFI) [2]. Using these indices, 3D power Doppler attempts to identify the different branches of the villous vessels, as well as the quantitative assessment of the number of vessels [3]. Preliminary studies suggest that reduction in the indices obtained by this technology may be associated with alterations in fetal growth, amniotic fluid volume and Doppler flowmetric parameters of the umbilical artery [3], [4]. This has sparked interest in the use of these indices for predicting pregnancy complications such as preeclampsia and intrauterine growth restriction [4]. Before large scale studies are performed to explore this potential, technical and methodological issues must first be addressed in order to produce valid and reproducible results [5].

Placental vascular indices have generally been calculated from evaluation of the entire placenta [6]. Unfortunately, except during the first trimester, the present 3D power Doppler technology does not permit evaluation of the entire placental vascular tree in a single sweep [7], [8], [9], [10]. To circumvent this limitation, placental vascular sonobiopsy has been proposed as a means to obtain a representative sample of the placental vascular tree [11], [12]. This involves acquisition of a 3D image of the placenta and defining the region of interest using the sphere mode in the Virtual Organ Computer-aided Analysis (VOCAL™) program, which automatically calculates the vascular indices. This technique has been shown to be feasible, clinically reproducible, and the placental vascular indices so obtained have been correlated with fetal biometry [4], [11], [12]. To our knowledge, the indices obtained by this technique have not been validated by comparing them to the indices from the whole placenta. This validation is necessary to ensure that the indices obtained from sonobiopsy are truly representative of the entire placenta.

The objective of this study was to test the hypothesis that placental vascular indices obtained from placental sonobiopsy correlate with and are representative of those from evaluation of the entire placenta.

Back to Article Outline

2. Materials and methods 

2.1. Study design 

This was part of an ongoing prospective cohort study of patients undergoing first trimester ultrasound examinations at our perinatal center from January to September 2009. Subjects included were consecutively selected women with singleton pregnancies between 11 and 14 weeks' in whom the entire placenta was visualized in a single image by two-dimensional (2D) ultrasound. Approval for the study was obtained from our institutional review board.

2.2. Image acquisition 

All images were acquired using Voluson 730 Expert ultrasound machines (GE Medical Systems, Milwaukee, WI, USA) equipped with a 4–8 MHz transducer. The same pre-established instrument power settings (angio mode: cent; smooth: 4/5; FRQ: low; quality: 16; density: 6; enhance: 16; balance: GO150; filter: 2; actual power: 2 dB; pulse repetition frequency: 0.9) were used in all cases. The entire view of the placenta was identified by 2D-ultrasound and the volume box was adjusted in order to scan the entire placenta. The angle of volume acquisition varied from 45° to 90° and duration was 10–15 s. For posterior and laterally located placentas a slight lateral inclination of the transducer was used to acquire the images. After scanning the entire volume the ultrasound images were stored on a removable hard disk for subsequent analysis.

2.3. Measurement of placental vascular indices by whole placenta evaluation 

Each image was recovered in succession for processing. Evaluation of the entire placenta was performed using the rotational technique in the VOCAL™ program included in the 4DView computer software (GE Medical Systems, Milwaukee, WI, USA). This involved rotating the image at 30° intervals and outlining the contour of the placenta six times in the A plane of the three orthogonal multiplanar system. When completed, the VOCAL™ program automatically calculated the vascular indices from computer algorithms (Fig. 1).

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 1 

  Assessment of placental vascular indices by whole placenta evaluation using the rotational technique (VOCAL™).

2.4. Measurement of placental vascular indices by sonobiopsy 

Placental vascular indices were then obtained from the placental images by taking four spherical sonobiopsies using the sphere mode in the VOCAL™ program. Two biopsies were taken on each side of the midline and approximately equidistant from each other along the length of the placental image. This involved positioning the limits of a virtual reference axis between the basal and chorionic plates (excluding both of them) and measuring the volume of a sphere by automatically rotating around that fixed axis [11], [12]. The sizes of the spheres varied, depending on the size of the placenta. The vascular indices for each sphere were automatically calculated by the VOCAL™ program (Fig. 2). The mean of each index from the four sonobiopsies was calculated and used for analysis.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 2 

  Assessment of placental vascular indices by sonobiopsy (one of four) using the sphere mode (VOCAL™).

All measurements in the VOCAL™ program were performed by the same examiner.

2.5. Statistical analysis 

Normal distribution of each index was verified using the Kolmogorov–Smirnov test. Indices that were not normally distributed (VI and VFI) were logarithmically transformed to achieve normal distribution. The mean of the placental indices from the two techniques were compared using the paired-sample t test. Correlation between each of the two pairs of indices was assessed using Pearson's correlation coefficient (r) and linear regression analysis. All regression graphs were plotted with logarithmic transformed indices to make them visually comparable.

Bland–Altman analysis was performed comparing each of the indices obtained by sonobiopsy to those from the whole placenta [13]. Because initial plots of the indices indicated that spread of the differences in each index as measured by the two techniques increased with increase in magnitude of the index, we plotted graphs with logarithmic transformed values (Log-Index_whole − Log-Index_sonobiopsy) against their average ([Log-Index_whole + Log-Index_sonobiopsy]/2) [13]. Bias, precision and agreement were evaluated using the back-transformed mean differences, standard deviations and 95% limits of agreement respectively.

All statistical analyses were performed using STATA version 10.0 (Stata Corp., College Station, TX) and MedCalc version 11.0.0 (MedCalc Software, Mariakerke, Belgium). Tests with p values < 0.05 were considered significant.

Back to Article Outline

3. Results 

From an ongoing cohort study, 120 consecutively selected subjects with singleton pregnancies were included in this study. The mean maternal age was 31.9 ± 5.8 years and the median gestational age was 12.0 weeks (range 11.1–13.8 weeks). A slight majority of the patients (73/120, 60.8%) was multiparous. Most of the study subjects were Caucasian (60%) or Hispanic (25%). Placenta location was anterior in 54.2% (65/120), posterior in 44.2% (53/120) and lateral or fundal in 1.6% (2/120) of cases (Table 1).

Table 1. Study Subject characteristics (N = 120).

Parameter	Value
Maternal age (mean ± SD)	31.9 ± 5.8
Gestational age (median [range])	12.0 (11.1–13.8)
Maternal weight (Lbs) (mean ± SD)	163.2 ± 37.7
Race, n (%)
Caucasian	72 (60)
Hispanic	30 (25)
African–American	10 (8.3)
Other	8 (6.7)
Parity (%)
Nulliparous	47 (39.2)
Multiparous	73 (60.8)
Placenta Location (%)
Anterior	65 (54.2)
Posterior	53 (44.2)
Lateral or Fundal	2 (1.6)

SD-standard deviation, Lbs-pounds.

A comparison of placental vascular indices from whole placenta evaluation and sonobiopsy is shown in Table 2. Measures of FI were normally distributed, while VI and VFI required logarithmic transformation to achieve normal distribution. The mean VI and VFI obtained from sonobiopsy were similar to those from whole placenta evaluation (13.9 [95% CI 12.3–15.8], vs. 14.3 [95% CI 12.1–17.0], p = 0.62 and 6.1 [95% CI 5.2–7.1], vs. 6.1 [95% CI 5.0–7.4], p = 0.93, respectively). On the other hand, the mean FI from the two techniques were significantly different (44.5 [95% CI 42.9–46.1], vs. 41.3 [95% CI 39.6–43.0], p = 0.001).

Table 2. Comparison of means and correlation of placental vascular indices from sonobiopsy and whole placenta evaluation (N = 120).

VI-vascularization index, FI-Flow Index, VFI-vascularization flow index.

Table 2 and Fig. 3 show results of Pearson's coefficient of correlation tests and linear regression analysis respectively. Significant correlations were noted between all the indices obtained by the two techniques (Pearson's r = 0.70 [95% CI 0.60–0.78, p < 0.001] for VI, r = 0.59 [95% CI 0.46–0.78, p < 0.001] for FI, and r = 0.69 [95% CI 0.58–0.77, p < 0.001] for VFI). The slopes of the regression lines for all indices were significantly different from zero (all p < 0.001). The linear models for VI and VFI were a better fit for the relationship between the measurements by the two techniques than for FI (coefficients of determination, R² = 0.49 for VI and 0.47 for VFI vs. 0.35 for FI).

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 3 

  Placental Vascular indices from sonobiopsy in relation to whole placenta evaluation with 95% CI of predicted (N = 120).

Results of Bland–Altman analysis to assess bias, precision and agreement between the two techniques for measuring placental vascular indices are shown in Fig. 4 and Table 3. Fig. 4 is a plot of the differences between each log-transformed index obtained by the two techniques against their average. The mean differences (measures of bias), standard deviations (measures of precision) and 95% limits of agreement of the back-transformed indices are shown in Table 3. This analysis showed that the mean measures of VI and VFI from the two techniques did not significantly differ (back-transformed difference 0.97 [95% CI – 0.86 to 1.09], p = 0.62 and – 0.99 [95% CI – 0.86 to 1.14], p = 0.89, respectively), indicating absence of systematic bias. In contrast, the mean measures significantly differed for FI (back-transformed difference 1.08 [95% CI 1.04–1.13], p < 0.001), suggesting significant bias towards underestimation. The standard deviations were 1.95, 1.25 and 02.18 for VI, FI and VFI, respectively. The 95% limits of agreement were 0.26–3.60 for VI, 0.71–1.66 for FI and 2.18–4.62 for VFI.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 4 

  Bland–Altman plots of placental vascular indices obtained from sonobiopsy and whole placenta evaluation showing 95% limits of agreement (N = 120).

Table 3. Comparison of bias (mean difference), precision (standard deviation) and agreement (limits of agreement) between placental indices from whole placenta evaluation and sonobiopsy (N = 120).

VI-vascularization index, FI-Flow Index-FI, VFI-vascularization flow index.

Back to Article Outline

4. Discussion 

This is the first study that formally tests the validity of placental vascular indices obtained from placenta vascular sonobiopsy. We sought to determine if the placental vascular indices obtained from sonobiopsy correlate with and are representative of those from evaluation of the entire placenta. To facilitate assessment of the indices from both the entire placenta and from sonobiopsy, we limited our study to only first trimester pregnancies where the entire placenta could be captured in a single ultrasound sweep. Our results confirm that all the indices obtained from sonobiopsy correlate well with those from the entire placenta. While the measurements of VI and VFI from sonobiopsy were similar to those from the whole placenta, measures of FI were significantly different. All the indices obtained by sonobiopsy were associated with relatively poor precision and wide limits of agreement. Of all the indices estimated by sonobiopsy, VI and VFI were more reliable, showing no significant systematic bias, although with poor precision and agreement. FI was the least reliable, showing significant bias towards underestimation.

The good correlation of all the placental vascular indices obtained by sonobiopsy with those from evaluation of the entire placenta is reassuring with regards to the prospect of using sonobiopsy as an alternative when whole placenta evaluation is not feasible. The observation that the mean VI and VFI from the two techniques did not significantly differ suggests that they may be the preferred indices when sonobiopsy is used. It is unclear, why measures of FI from sonobiopsy correlate with, but significantly differ from those obtained by whole placenta evaluation. We posit that it may be a reflection of the pattern of blood flow in the placenta. FI is a measure of placental blood flow at the time of the 3D ultrasound sweep and blood flow in the placenta is highest at the cord insertion and lower towards the basal plate. Because FI from whole placenta evaluation represents average placental blood flow while FI from sonobiopsy represent flow in selected regions of the placenta, it is plausible for the two sets of measurements to be correlated, but significantly differ in magnitude.

A direct comparison of our results to those of prior studies is not possible, since they used different methodologies and sonobiopsy techniques. All studies to date have included a wide range of gestational ages, evaluated correlation of the indices with gestational age and determined intra- and inter-observer agreements [4], [11], [12], [14]. None of the studies compared the vascular indices from sonobiopsy to those from whole placenta evaluation. The original technique described by Merce et al. involved obtaining a single spherical biopsy between the basal and chorionic plates [12]. This was applied to 30 normal pregnancies between 14 and 40 weeks' and shown to be reproducible (intra-observer correlation coefficient 0.88). Measures of FI and VFI showed better intra-observer agreement than VI. The same technique was used by the same investigators to evaluate 99 normal pregnancies between 14 and 40 weeks' gestation [11]. All indices were shown to correlate significantly with gestational age, fetal biometry (except VI and fetal weight) and umbilical artery Doppler resistance index. FI showed the most, while VI showed the least correlation with fetal parameters. We did not evaluate the relationship between the placental vascular indices and gestational age in our study since we limited enrolment to a narrow gestational age range.

Guiot et al. employed a modified spherical sonobiopsy technique which involved taking five samples per placenta; one central and four lateral, two on either side, equidistant from the center [14]. They evaluated 15 normal and 30 IUGR pregnancies between 23 and 37 weeks' gestation and demonstrated that placental indices were independent of gestational age and placenta location. VI and VFI were lower in the presence of elevated umbilical artery Dopplers. FI showed the least variability between samples and was lower than controls only in the most severe cases of placental impairment (absent end-diastolic umbilical artery blood flow). Most recently, Noguchi et al. obtained four spherical biopsies in each plane of the placenta in 208 normal and 13 IUGR pregnancies between 12 and 40 weeks' [4]. They demonstrated good intra- and inter-observer agreement (correlation coefficient >0.85) as well as a weak correlation between the indices and gestational age. After 32 weeks', all indices in the IUGR pregnancies were significantly lower compared to the normal pregnancies.

The distributions of the placental vascular indices in our study deserve a special comment. We observed that only FI was normally distributed while VI and VFI required logarithmic transformation to achieve normal distribution. In contrast, the only study dedicated to the distribution of placental vascular indices throughout gestation, demonstrated normal distribution of all indices [15]. This difference may be attributable to the larger sample size (N = 314) or the wider gestational age range of the subjects in that study (12–40 weeks).

A major strength of our study is the rigorous methodology employed to directly compare sonobiopsy to whole placenta evaluation. In addition to the traditional correlation coefficients and linear regression analyses, we used the approach proposed by Bland and Altman for assessing agreement between two methods of clinical measurements [13]. Because correlation coefficients may be misleading when used to assess agreement between two methods, this analysis augmented our ability to evaluate the two measurement techniques. Secondly, we used the same pre-established instrument power settings for all examinations. This controlled for the known effect of machine settings on measured placental vascular indices [16], [17]. Finally, by employing the two measurement techniques on the same images, we eliminated differences that could be attributed to different pregnancy and subject characteristics.

Despite these strengths there are limitations which should be taken into account when interpreting our results. First, while our sample size is larger than those in many of the previous studies, it is still relatively small. This may account for the relatively wide confidence intervals as well as the relatively poor measures of precision and agreement. Second, we limited our study to only first trimester pregnancies in which the entire placenta could be captured in a single ultrasound sweep. While this was necessary to make the study feasible, it resulted in the exclusion of the higher gestational ages where sonobiopsy is often necessary. Because placental characteristics may change with advancing gestational age, care must be taken when extrapolating our findings to higher gestational ages. However, the largest study evaluating the distribution of placental vascular indices over a wide range of gestational ages showed that all indices were constant throughout pregnancy [15]. If this is confirmed, our findings may be safely extrapolated to higher gestational ages.

In conclusion, we have demonstrated that placental vascular indices obtained from sonobiopsy correlate well with those from evaluation of the entire placenta. Measures of VI and VFI appear to be the most reliable, while FI shows significant bias towards underestimation. As research continues to evaluate the role of placental vascular indices in predicting placenta-related pregnancy complications, our results suggest that sonobiopsy may be a valid alternative for evaluation of the placental vascular tree when visualization of the entire placenta is not feasible. In such cases, VI and VFI would be the preferred parameters of evaluation.

Back to Article Outline

References 

1. Pijnenborg R, Vercruysse L, Hanssens M. The uterine spiral arteries in human pregnancy: facts and controversies. Placenta. 2006 Sep–Oct;27(9–10):939–958
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (970 KB)
   • CrossRef
2. Pairleitner H, Steiner H, Hasenoehrl G, Staudach A. Three-dimensional power Doppler sonography: imaging and quantifying blood flow and vascularization. Ultrasound Obstet Gynecol. 1999 Aug;14(2):139–143
   • View In Article
   • MEDLINE
3. Guimaraes Filho HA, da Costa LL, Araujo Junior E, Nardozza LM, Nowak PM, Moron AF, et al. Placenta: angiogenesis and vascular assessment through three-dimensional power Doppler ultrasonography. Arch Gynecol Obstet. 2008 Mar;277(3):195–200
   • View In Article
   • CrossRef
4. Noguchi J, Hata K, Tanaka H, Hata T. Placental vascular sonobiopsy using three-dimensional power Doppler ultrasound in normal and growth restricted fetuses. Placenta. 2009 May;30(5):391–397
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (410 KB)
   • CrossRef
5. Alcazar JL. Three-dimensional power Doppler derived vascular indices: what are we measuring and how are we doing it?. Ultrasound Obstet Gynecol. 2008 Sep;32(4):485–487
   • View In Article
   • CrossRef
6. Rizzo G, Capponi A, Cavicchioni O, Vendola M, Arduini D. Placental vascularization measured by three-dimensional power Doppler ultrasound at 11 to 13 + 6 weeks' gestation in normal and aneuploid fetuses. Ultrasound Obstet Gynecol. 2007 Sep;30(3):259–262
   • View In Article
   • CrossRef
7. Hafner T, Kurjak A, Funduk-Kurjak B, Bekavac I. Assessment of early chorionic circulation by three-dimensional power Doppler. J Perinat Med. 2002;30(1):33–39
   • View In Article
   • MEDLINE
   • CrossRef
8. Matijevic R, Kurjak A. The assessment of placental blood vessels by three-dimensional power Doppler ultrasound. J Perinat Med. 2002;30(1):26–32
   • View In Article
   • MEDLINE
   • CrossRef
9. Pretorius DH, Nelson TR, Baergen RN, Pai E, Cantrell C. Imaging of placental vasculature using three-dimensional ultrasound and color power Doppler: a preliminary study. Ultrasound Obstet Gynecol. 1998 Jul;12(1):45–49
   • View In Article
   • MEDLINE
10. Yu CH, Chang CH, Ko HC, Chen WC, Chang FM. Assessment of placental fractional moving blood volume using quantitative three-dimensional power Doppler ultrasound. Ultrasound Med Biol. 2003 Jan;29(1):19–23
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (119 KB)
    • CrossRef
11. Merce LT, Barco MJ, Bau S, Kupesic S, Kurjak A. Assessment of placental vascularization by three-dimensional power Doppler “vascular biopsy” in normal pregnancies. Croat Med J. 2005 Oct;46(5):765–771
    • View In Article
    • MEDLINE
12. Merce LT, Barco MJ, Bau S. Reproducibility of the study of placental vascularization by three-dimensional power Doppler. J Perinat Med. 2004;32(3):228–233
    • View In Article
    • MEDLINE
    • CrossRef
13. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986 Feb 8;1(8476):307–310
    • View In Article
    • CrossRef
14. Guiot C, Gaglioti P, Oberto M, Piccoli E, Rosato R, Todros T. Is three-dimensional power Doppler ultrasound useful in the assessment of placental perfusion in normal and growth-restricted pregnancies?. Ultrasound Obstet Gynecol. 2008 Feb;31(2):171–176
    • View In Article
    • CrossRef
15. de Paula CF, Ruano R, Campos JA, Zugaib M. Quantitative analysis of placental vasculature by three-dimensional power Doppler ultrasonography in normal pregnancies from 12 to 40 weeks of gestation. Placenta. 2009 Feb;30(2):142–148
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (769 KB)
    • CrossRef
16. Raine-Fenning NJ, Nordin NM, Ramnarine KV, Campbell BK, Clewes JS, Perkins A, et al. Evaluation of the effect of machine settings on quantitative three-dimensional power Doppler angiography: an in-vitro flow phantom experiment. Ultrasound Obstet Gynecol. 2008 Sep;32(4):551–559
    • View In Article
    • CrossRef
17. Schulten-Wijman MJ, Struijk PC, Brezinka C, De Jong N, Steegers EA. Evaluation of volume vascularization index and flow index: a phantom study. Ultrasound Obstet Gynecol. 2008 Sep;32(4):560–564
    • View In Article
    • CrossRef