Importance of Pulse Repetition Frequency Adjustment for 3- and 4-Dimensional Power Doppler Quantification

Transcription

1 ORIGINAL RESEARCH Importance of Pulse Repetition Frequency Adjustment for 3- and 4-Dimensional Power Doppler Quantification Andre H. Miyague, PhD, Theo Z. Pavan, PhD, Carlos A. Soares, MD, Luc De Catte, PhD, Carolina O. Nastri, PhD, Alec W. Welsh, PhD, Wellington P. Martins, PhD Received January 13, 2015, from the Department of Obstetrics and Gynecology, Medical School of Ribeirão Preto (A.H.M., C.A.S., C.O.N., W.P.M.), and Department of Physics, School of Philosophy, Sciences, and Letters of Ribeirão Preto (T.Z.P.), University of São Paulo, Ribeirao Preto, Brazil; Department of Obstetrics and Gynecology, University Hospital Evangelico, Curitiba, Brazil (A.H.M.); Woman and Fetal Medicine Institute, Curitiba, Brazil (A.H.M.); Ultrasonography and Retraining Medical School of Ribeirão Preto, Ribeirão Preto, Brazil (C.A.S., C.O.N., W.P.M.); Department of Obstetrics and Gynecology, University Hospitals Leuven, Leuven, Belgium (L.D.C.); School of Women s and Children s Health, University of New South Wales, Randwick, New South Wales, Australia (A.W.W.); and Department of Maternal-Fetal Medicine, Royal Hospital for Women, Randwick, New South Wales, Australia (A.W.W.). Revision requested February 13, Revised manuscript accepted for publication March 31, This work was supported by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior, Brazil; Conselho Nacional de Desenvolvimento Científico e Tecnológico, Brazil; Hospital das Clínicas de Ribeirão Preto, Brazil; Universidade de São Paulo, Brazil; an Escola de Ultrasonografia e Reciclagem Médica Ribeirão Preto, Brazil. Address correspondence to Wellington P. Martins, PhD, Department of Obstetrics and Gynecology, Medical School of Ribeirão Preto, University of São Paulo, Avenida Bandeirantes Andar, Hospital das Clínicas Ribeirão Preto, Campus Universitário, Ribeirao Preto SP , Brazil. wpmartins@gmail.com Abbreviations FI, flow index; PI, pulsatility index; PRF, pulse repetition frequency; RF, radiofrequency; STIC, spatiotemporal image correlation; 3D, 3-dimensional; VFI, vascularization-flow index; VI, vascularization index doi: /ultra Objectives To determine the influence of the pulse repetition frequency (PRF) and wall motion filter on the 3-dimensional (3D) power Doppler vascularization-flow index (VFI) and volumetric pulsatility index (PI) obtained from spatiotemporal image correlation (STIC) data sets acquired from a common carotid artery of a healthy participant. Methods We acquired 11 STIC data sets, 1 for each PRF value ranging from 0.6 to 9.0 khz. Vascularization-flow index and volumetric PI values were determined from the 440 static 3D data sets contained in these STIC data sets. Additionally, 3 sets of radio - frequency data were acquired for offline processing of different wall motion filter values for PRF values of 0.6, 3.3, and 10 khz. Results We constructed VFI curves and observed 2 patterns: a flattened pattern with a low PRF and a triphasic pattern with a high PRF, correlating with the known pulsed wave Doppler profile of this vessel. Volumetric PI values were around 0 for low PRF settings and increased with increasing PRF. Analysis of the radiofrequency data showed that increasing wall motion filter values gradually filtered out the low-velocity power Doppler signals while retaining the higher-velocity ones, allowing the distinction of integrated power Doppler signal velocity throughout the cardiac cycle. Conclusions We conclude that the PRF and wall motion filter dramatically influence 3D power Doppler indices and the volumetric PI, and the use of PRF values in which minimum VFI values are measured during the diastolic phase in the spectral Doppler wave may validate the use of the volumetric PI. Key Words Doppler sonography; imaging; power Doppler sonography; 3-dimensional sonography; ultrasound education The potential for 3-dimensional (3D) power Doppler sonography as a diagnostic imaging tool has generated a great deal of research interest for more than a decade. 1,2 A substantial focus on quantification of the vascularity and flow of a whole organ, such as the placenta, has come through the analysis of 3 indices, known as the vascularization index (VI), flow index (FI), and vascularization-flow index (VFI). Correlations between these indices and perfusion have been shown in ex vivo, in vivo, and in vitro models. 3 5 However, comparison within and between patients still seems impractical, as 3D power Doppler indices are highly dependent on machine settings, 6,7 attenuation, 3,5 and the sampling volume, 8 and there is a lack of standardization. Additionally, flow pulsatility during 3D power Doppler data acquisition could represent an important source of variability. 9, by the American Institute of Ultrasound in Medicine J Ultrasound Med 2015; 34:

2 We have recently addressed the influence of the cardiac cycle on the 3D power Doppler indices 10 ; in addition, spatiotemporal image correlation (STIC) 11,12 has been used to generate a volumetric pulsatility index (PI) with a research focus on the influence of gain adjustment 13 and attenuation. 14 However, there are further features and settings that are capable of influencing 3D power Doppler imaging, and these will need to be standardized before clinical application can be considered. Although power Doppler quantification is based on the amplitude of ultrasound scattering, which is related to the power Doppler signals integrated over the entire frequency range, 15 recent studies have demonstrated that the color voxel intensity and brightness vary according to the phase of the cardiac cycle, with higher 3D power Doppler indices values related to systole and lower ones to diastole. 16,17 Therefore, it seems reasonable to infer that flow velocity may play a role in 3D power Doppler quantification. Furthermore, we supposed that flow velocity thresholds defined by the pulse repetition frequency (PRF) and wall motion filter are capable of altering the quantification of 3D power Doppler signals throughout the cardiac cycle. Thus, in this study, we aimed to investigate the influence of the PRF on 3D power Doppler indices and the volumetric PI obtained from several 3D power Doppler STIC data sets acquired from a common carotid artery of a healthy participant, which has been shown to be a model of regular frequency and similar stability when compared to an in vitro flow phantom model. 10 In addition, we sought to analyze the influence of wall motion filter cutoff frequencies on power Doppler signal quantification for a single PRF, since automatic changes in the wall motion filter were previously described when changing the PRF. 6 Materials and Methods Study Model This study was approved by the local Institutional Review Board, and signed informed consent was obtained before acquisitions. The left carotid artery of a healthy, nonsmoking male participant who consented to participate was evaluated. The participant had no personal or family history of vascular disease and no record of high blood pressure. The participant was asked to fast for 8 hours and avoid consumption of caffeine, painkillers, and antiinflammatory drugs for 24 hours before the examination. performed all ultrasound scans using a Voluson E8 Expert ultrasound system (GE Healthcare, Zipf, Austria) equipped with an RSP6-16-D transducer, respecting the ALARA (as low as reasonably achievable) principle (maximum thermal index, <0.5). Briefly, STIC is an automated volume acquisition method in which several 2- dimensional frames are acquired, covering a certain angle in a fixed period containing several heartbeats (eg, 15 seconds). After the acquisition, the software identifies the heart rate by analyzing either the B-mode or the power Doppler pulsatility. Using the detected heart rate, the software rearranges every frame according to its spatial and temporal domain to obtain several 3D data sets, each representing distant moments of a single reconstructed cardiac cycle. 11,12 Spectral pulsed wave Doppler waves were initially acquired, and velocity analysis was performed by the builtin automatic Doppler spectral wave analysis software (Figure 1): angle between flow and the ultrasonic beam, 54 ; detected heart rate, 59 beats per minute; peak systolic velocity, cm/s; end-diastolic velocity, cm/s; and PI, The vascular preset and power Doppler mode were used. The machine settings were set and maintained (except for PRF) for all acquisitions as follows: depth, 4.2 cm; power, 100%; gain, 0.0; wall motion filter, mid 1; frequency, mid; flow resolution, mid 1; balance, G > 200; smooth (rise/fall), 2/2; ensemble, 12; line density, 7; power Doppler map, 5; gently color, on; artifact suppression, on; and line filter, 2. The angle of acquisition for STIC was set to 10 and the time of the acquisition to 15 seconds. The participant was asked to hold his breath during each STIC data set acquisition. One STIC data set was acquired Figure 1. Spectral pulsed wave Doppler wave acquired from a healthy participant s common carotid artery. Data were collected from the GE Voluson E8 equipment. Data Acquisition One observer (W.P.M., who had 10 years of experience with 3D sonography and 8 years of experience with STIC) 2246 J Ultrasound Med 2015; 34:

3 for each PRF, from 0.6 khz to a maximum of 9.0 khz through the smallest serial increase in the PRF value, generating 11 PRF values. If the heart rate determined after each STIC acquisition was less than 50 or greater than 64 beats per minute, the data set would not be saved, and a new one would be acquired. We used these limits because variability of up to 7 beats per minute is expected to occur in healthy young people. 18 All carotid artery data sets were analyzed offline by a second observer (A.H.M., who had 5 years of experience with 3D sonography and 3 years of experience with STIC) using 4D View version 10.5 software (GE Healthcare). To measure the influence of the wall motion filter on power Doppler signal quantification, we acquired radio - frequency (RF) data sets from the same model, for offline processing, using a Sonix RP ultrasound scanner (Ultrasonix Medical Corporation, Richmond, British Columbia, Canada) equipped with an L14-5/38 linear transducer operating at 5 MHz. The signals were collected by using a conventional Doppler pulse sequence with an ensemble length of 10 scan lines and PRFs of 0.6, 3.3, and 10 khz. For PRFs of 3.3 and 10 khz, we applied wall motion filters of 0, 10, 100, and 500 Hz and 1.5 khz, and for the PRF of 0.6 khz, we applied wall motion filters of 0, 10, 100, and 250 Hz. Frame rates were set between 15 and 17 frames per second. A total of 3 data sets were acquired for offline application of different wall motion filters to each data set. Measurements Spherical samples of cm3 (diameter, mm) made to assess the region of interest were analyzed with Virtual Organ Computer-Aided Analysis (GE Healthcare). The central point of the sphere was placed in the center of the observed flow for consistency; in this case, the sphere would avoid turbulent flow and include only the laminar flow. The histogram facility was used to calculate the VI, FI, and VFI within the volume sample; for this study, only the VFI was recorded and analyzed to reduce the number of variables, making the results easier to read and interpret, as it has been thought to most closely approximate to vascularity using 3D power Doppler sonography. 5 Each STIC data set contains several static data sets ( frames ): each frame was analyzed sequentially, providing several VFI values (40) for each STIC data set. The volumetric PI was determined as follows: [(maximum VFI value) (minimum VFI value)]/(mean VFI value). 17 The RF data were processed offline in MATLAB (The MathWorks, Inc, Natick, MA), and the power Doppler amplitude (PD) was calculated by the following formula: where I is the in-phase; Q is the quadrature component of the signals obtained after the demodulation of the RF data; and N is the length of the ensemble. Before calculating the integrated power, the I and Q signals were filtered (wall motion filter) by applying a third-order high-pass Butterworth filter. The mean integrated power was estimated from a region of interest of 0.05 cm 3 placed in the center of the observed flow. To identify the most common PRF value used in studies that assessed 3D power Doppler indices, we performed a mini systematic review of the literature, searching for the PRF values used in every included article. The review was performed in the PubMed database from January 1990 until June 2013 using the following search strategy: VFI or FI or VI and Doppler and 3-dimensional or 3D. The eligibility criteria for inclusion were whether they assessed, among other variables, the 3D power Doppler indices to quantify vascularity and flow. Statistical Analysis Statistical analysis was performed by one of the authors (W.P.M.) using Graph Pad Prism version 5.0 software for Windows (GraphPad Software, Inc, San Diego, CA) and Microsoft Office Excel (Microsoft Corporation, Redmond, WA). The frames contained in each STIC data set were ordered consecutively, representing one entire reconstituted cardiac cycle; however, the first frame from each STIC data set might have represented any moment of the cardiac cycle. To be able to visually compare the curves of VFI values between STIC data sets, we set the frame with the lowest minimum VFI value (corresponding to the enddiastolic moment) as the first frame when constructing the graphs of the curves of VFI values along a single cardiac cycle; the other frames were ordered consecutively from this end-diastolic frame. Results N PD (db) = 10log ( 1 I2 (n) + Q 2 (n) ), N n=1 Spatiotemporal image correlation automatically assessed that heart rates were between 54 and 60 beats per minute in all cases; therefore, all acquired STIC data sets were stored without the need to reacquire any new data sets. Each STIC data set contained 40 static 3D power Doppler frames. The curves obtained by plotting the VFI values over time within each cardiac cycle clearly demonstrate that the differentiation of the signal velocity gradually occurred as PRF values increased. These curves have been J Ultrasound Med 2015; 34:

4 divided into low and high PRF values according to the observed curve patterns to facilitate their graphical comprehension. We observed 2 different curve patterns, as low PRF values (Figure 2A) displayed a flattened curve, and high PRF values (Figure 2B) displayed a curve that resembled the triphasic 2-dimensional Doppler spectral waves seen in this vessel. The volumetric PI values (Figure 3) were around 0 in the low-prf group, indicating a lack of Figure 2. A, Vascularization-flow index values during a cardiac cycle using low PRF values. B, Vascularization-flow index values during a cardiac cycle using high PRF values. Data were collected from the GE Voluson E8 equipment. discrimination between systolic and diastolic amplitude, and a gradual increase in volumetric PI values was observed for each increment in PRF values in the high-prf group, with relative stability within the PRF values of 3.2 and 4.0 khz. Analysis of the RF data demonstrated that an increase in the wall motion filter settings reduced the mean velocity of the power Doppler signals over time, allowing differentiation between maximum and minimum velocities of power Doppler signals within the cardiac cycle, which corresponded, respectively, to the systolic and diastolic phases. This finding was clearly perceivable when using PRF values of 3.3 and 10.0 khz and less so with a PRF of 0.6 khz (Figure 4). Our mini systematic review of the literature included 77 articles, and we observed that all evaluated studies (using STIC or not) used low PRF settings (0.6 or 0.9 khz; see Discussion ).The use of low PRF settings was justified by most of the authors as a desire for higher power Doppler signal sensitivity. Discussion In this study, we demonstrated that the PRF and wall motion filter dramatically influence VFI quantification. Lower PRF values, and consequently lower wall motion filter values, tend to display flattened VFI curves across the cardiac cycle; on the other hand, high PRF values, and consequently higher wall motion filter values, tend to display curves that resemble the triphasic 2-dimensional spectral waves. Volumetric PI values were around 0 in the Figure 3. Volumetric PI values according to PRF values. Data were collected from the GE Voluson E8 equipment J Ultrasound Med 2015; 34:

5 low-prf group, and they increased with each increment in the PRF value in the high-prf group. It is important to acknowledge that these results were obtained from a single STIC measurement for each PRF; nonetheless, the stability of the method has already been demonstrated in a Figure 4. Power Doppler (PD) signal velocities over time according to the PRF and wall motion filter settings. Data were collected from the Ultrasonix Sonix RP equipment. previous study. 10 We observed that increases in wall motion filter values are linearly correlated with the power Doppler signal filtering rates (Figure 4). Additionally, we observed that all previously published studies used low PRF values when examining 3D power Doppler indices. 9,12,19 22 This experiment was conducted with a single-vessel model, whereas the main purpose of 3D power Doppler sonography is to evaluate the vascularization and flow of a whole organ or tissue. Thus, this research will be extrapolated to larger vascular fields in the future. Another limitation was that we used 2 different ultrasound systems for this experiment because one cannot avoid the inherited drawbacks such as the automatic adjustment in the wall motion filter for each PRF when usingthe GE Voluson system; in that case, we used a second ultrasound machine (the Ultrasonix Sonix RP) to collect the RF data for offline processing, therefore avoiding such inconveniences. For techniques such as 3D power Doppler sonography to be established, it is important that there is an established conventions for machine settings, and more research in this direction is indicated. Although STIC provides a novel perspective into 3D power Doppler assessment, a great deal of time is required to analyze a single STIC data set, as each acquisition may contain as many as 40 static 3D data sets. Potentially, standardization of machine settings could be an important step toward automated analysis of STIC data sets. In this study, we chose to use a small sphere in the center of the vessel to quantify 3D power Doppler indices based on the assumption that the PRF would mainly affect quantification of the FI. We observed that the velocity of the power Doppler signal throughout the cardiac cycle was better distinguished when using higher PRF Figure 5. Histograms obtained from the same vessel during the systolic phase using different PRF settings. A, Pulse repetition frequency of 9.0 khz. B, Pulse repetition frequency of 0.6 khz. Note in the graphic representation that most of the color voxels were translated as 100%. Data were collected from the GE Voluson E8 equipment. J Ultrasound Med 2015; 34:

6 values, as these settings clearly increased the maximum power Doppler velocity threshold displayed in the histogram (Figure 5A). When the power Doppler signal velocity within the sample volume is higher than the established maximum level, the VFI will be quantified as 100% independent of its velocity (Figure 5B). We observed that all published studies used low PRF values when quantifying 3D power Doppler signals. However, our study showed a flattened pattern when using low PRF values, reflecting a reduced ability to distinguish between the different integrated amplitudes in the center of the vessel during the cardiac cycle. When increasing the PRF, the ability of power Doppler sonography to identify differences in integrated Doppler amplitudes was improved; however, higher PRF values (ie, 7.5 and 9 khz) reduced the ability to display power Doppler signals originating from low blood flow velocity. In a study evaluating the effect of machine settings on 3D static data sets, Raine- Fenning et al 6 demonstrated a complex relationship between PRF and 3D power Doppler indices, as automatic changes in the power and wall motion filter occurred by increasing the PRF. It is well known that the wall motion filter is essential for estimating tissue vascularization, 23,24 and usually, it cannot be adjusted in a graduated manner as other machine settings. The wall motion filter is usually adjusted by changing the PRF, which justifies the importance of adopting objective criteria for an appropriate choice of the PRF. By analyzing the RF data, we could demonstrate that the wall motion filter plays a key role in distinguishing high and low power Doppler velocities throughout the cardiac cycle. A proportion of the power Doppler signals, which are probably those with low-velocity flows, are increasingly filtered as the wall motion filter value is raised. In comparison, as PRF values are raised, the wall motion filter gradually filters out the power Doppler signals while retaining the higher-velocity ones. The automatic changes in the wall motion filter that occur when increasing PRF values generate a band of power Doppler signals that are defined by upper and lower velocity thresholds. The upper threshold is defined by the PRF: the higher the PRF value, the higher the power Doppler signal velocity to be decoded; the lower threshold is defined by the wall motion filter: the higher the wall motion filter value, the higher the rate of filtered power Doppler signals arising from lowvelocity blood flow. Therefore, when increasing PRF values, the concomitant influence of the PRF and wall motion filter allowed, respectively, higher power Doppler velocity thresholds and higher filtering rates of power Doppler signals from low-velocity blood flow. Although power Doppler signal generation is based on the amplitude of ultrasound scattering represented by the blood cells, it is important to acknowledge that the PRF and wall motion filter influence quantification of maximum, minimum, and average VFI values. These findings are in accordance with those reported in a previous study. 6 The PRF also interferes with volumetric PI determination: low PRF values limit the differentiation of high power Doppler signal velocities (above the threshold established by the PRF), making them close to the low power Doppler signals in all frames, therefore reducing the volumetric PI values; on the other hand, high PRF values (ie, >5.0 khz) generate high wall motion filter values, which disable the ability to detect power Doppler signals during diastole, increasing the volumetric PI values. When evaluating the intraobserver reliability of volumetric impedance indices (ie, volumetric PI, volumetric resistive index, and volumetric systolic-to-diastolic ratio), Welsh et al 25 observed that the maximum VI and VFI values were very close to the minimum values, and, as a reflection of these findings, the volumetric impedance indices were around 0 for the volumetric PI and volumetric resistive index and around 1 for the volumetric systolic-to-diastolic ratio. These findings may be explained by the low PRF value (0.6 khz) used by the authors, which reduced the ability of Virtual Organ Computer-Aided Analysis to distinguish between different signal velocities that occurred during a cardiac cycle. Pulse repetition frequency values of 3.2 and 4.0 khz were considered more suitable for the assessed flow, as these values generated curve patterns similar to those observed in Doppler spectral waves. Thus, in cases in which pulsed wave Doppler imaging shows diastolic flow, we suggest that the observer gradually decrease the PRF value until the consequent reduction in the wall motion filter allows display of a minimal power Doppler signal on the screen during the diastolic phase, therefore allowing quantification of different signal velocities within the cardiac cycle and, finally, construction of the volumetric PI. Further in vivo studies are indicated to determine the optimum balance of the wall motion filter and PRF for generation of the volumetric PI. Finally, the combination of STIC imaging with 3D power Doppler imaging allows a measurement that is thought to approximate tissue impedance that possesses internal standardization across the cardiac cycle; in other words, because of the volumetric PI formula (where the denominator is the mean VFI value), if one increases gain or reduces the attenuation effect, it will proportionally increase the maximum, minimum, and mean VFI values, therefore 2250 J Ultrasound Med 2015; 34:

7 creating an internal standardization. Nevertheless, our findings do not yet correlate with clinical practice, as 3D power Doppler sonography still lacks standardization. As research advances in this area, we believe that demonstration of the influence of the PRF and wall motion filter on 3D power Doppler quantification and volumetric PI estimation will represent an important step in optimizing settings for this new tool. References 1. Martins WP. Three-dimensional power Doppler: validity and reliability. Ultrasound Obstet Gynecol 2010; 36: Pairleitner H, Steiner H, Hasenoehrl G, Staudach A. Three-dimensional power Doppler sonography: imaging and quantifying blood flow and vascularization. Ultrasound Obstet Gynecol 1999; 14: Jones NW, Hutchinson ES, Brownbill P, et al. In vitro dual perfusion of human placental lobules as a flow phantom to investigate the relationship between fetoplacental flow and quantitative 3D power Doppler angiography. Placenta 2009; 30: Morel O, Pachy F, Chavatte-Palmer P, et al. Correlation between uteroplacental three-dimensional power Doppler indices and the uterine real blood flow: evaluation in a pregnant sheep experimental model. Ultrasound Obstet Gynecol 2010; 36: Raine-Fenning NJ, Nordin NM, Ramnarine KV, et al. Determining the relationship between three-dimensional power Doppler data and true blood flow characteristics: an in-vitro flow phantom experiment. Ultrasound Obstet Gynecol 2008; 32: Raine-Fenning NK, Nordin NM, Ramnarine KV, 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; 32: Martins WP, Raine-Fenning NJ, Ferriani RA, Nastri CO. Quantitative three-dimensional power Doppler angiography: a flow-free phantom experiment to evaluate the relationship between color gain, depth and signal artifact. Ultrasound Obstet Gynecol 2010; 35: Kudla MJ, Alcázar JL. Does sphere volume affect the performance of three-dimensional power Doppler virtual vascular sampling for predicting malignancy in vascularized solid or cystic-solid adnexal masses? Ultrasound Obstet Gynecol 2010; 35: Martins WP, Lima JC, Welsh AW, et al. Three-dimensional Doppler evaluation of single spherical samples from the placenta: intra- and interobserver reliability. Ultrasound Obstet Gynecol 2012; 40: Miyague AH, Raine-Fenning NJ, Polanski L, et al. Assessing the repeatability of 3D Doppler indices obtained by static 3D and STIC power Doppler: a combined in-vitro/in-vivo flow phantom study. Ultrasound Obstet Gynecol 2013; 42: DeVore GR, Falkensammer P, Sklansky MS, Platt LD. Spatio-temporal image correlation (STIC): new technology for evaluation of the fetal heart. Ultrasound Obstet Gynecol 2003; 22: Martins WP, Welsh AW, Falkensammer P, Raine-Fenning NJ. Re: spatiotemporal imaging correlation (STIC) technical notes about STIC triggering and choosing between power Doppler or high-definition color flow. Ultrasound Med Biol 2013; 39: Miyague AJ, Raine-Fenning NJ, Pavan TZ, et al. Influence of gain adjustment on three-dimensional power Doppler indices and on STIC volumetric pulsatility indices using a flow phantom. J Ultrasound Med 2013; 32: Miyague AJ, Pavan TZ, Grillo FW, Teixeira DM, Nastri CO, Martins WP. Influence of attenuation on three-dimensional power Doppler indices and STIC volumetric pulsatility index: a flow phantom experiment. Ultrasound Obstet Gynecol 2014; 43: Szabo TL. Diagnostic Ultrasound Imaging: Inside Out. 1st ed. Philadelphia PA: Elsevier Science; Alcázar JL, Kudla MJ. Ovarian stromal vessels assessed by spatiotemporal image correlation high-definition flow in women with polycystic ovary syndrome: a case-control study. Ultrasound Obstet Gynecol2012; 40: Martins WP, Welsh AW, Lima JC, Nastri CO, Raine-Fenning NJ. The volumetric pulsatility index as evaluated by spatiotemporal image correlation (STIC): a preliminary description of a novel technique, its application to the endometrium and an evaluation of its reproducibility. Ultrasound Med Biol 2011; 37: O Brien IA, O Hare P, Corrall RJ. Heart rate variability in healthy subjects: effect of age and the derivation of normal ranges for tests of autonomic function. Br Heart J 1986; 55: Welsh AW, Hou M, Meriki N, Stevenson G. Use of four-dimensional analysis of power Doppler perfusion indices to demonstrate cardiac cycle pulsatility in fetoplacental flow. Ultrasound Med Biol2012; 38: Hata T, Tanaka H, Noguchi J. Transvaginal 3-D power Doppler ultrasound evaluation of the fetal brain at weeks of gestation. Ultrasound Med Biol 2012; 38: Araujo Júnior E, Nardozza LM, Nowak PM, Rolo LC, Guimarães Filho HA, Moron AF. Three-dimensional power Doppler placental vascularisation indices in early pregnancy: a pilot study. J Obstet Gynaecol 2011; 31: Collins SL, Stevenson GN, Noble JA, Impey L, Welsh AW. Influence of power Doppler gain setting on Virtual Organ Computer-Aided Analysis indices in vivo: can use of the individual sub-noise gain level optimize information? Ultrasound Obstet Gynecol 2012; 40: Elfarnawany M, Pinter SZ, Lacefield JC. Improved objective selection of power Doppler wall filter cut-off velocity for accurate vascular quantification. Ultrasound Med Biol 2012; 38: Pinter SZ, Lacefield JC. Detectability of small blood vessels with highfrequency power Doppler and selection of wall filter cut-off velocity for microvascular imaging. Ultrasound Med Biol 2009; 35: Welsh AW, Hou M, Meriki N, Martins WP. Spatiotemporal image correlation derived volumetric Doppler impedance indices from spherical samples of the placenta: intraobserver reliability and correlation with conventional umbilical artery Doppler indices. Ultrasound Obstet Gynecol 2012; 40: J Ultrasound Med 2015; 34: