Cerebral, splanchnic, and renal transit time measurement and blood volume estimation using contrast-enhanced ultrasonography

Abstract

We aimed to measure cerebral, splanchnic, and renal transit times and the associated blood volumes using contrast ultrasound. In healthy individuals, regional transit times were calculated from time-intensity curves generated as ultrasound contrast passed through the associated inflow and outflow vessels. These included the internal carotid artery and internal jugular vein (brain), the superior mesenteric artery and portal vein (intestines), and the renal artery and renal vein (kidney). An organ’s blood volume relative to the stroke volume delivered to that organ with each cardiac cycle was calculated from the product of heart rate and transit-time of contrast passage through the associated vascular bed. The fraction of systemic stroke volume received by each organ was calculated from the respective velocity time-integral and inflow vessel cross-sectional area and used to estimate absolute organ blood volume. The cohort consisted of 16 participants (age: 42 ± 13 years; 5 female) without known cerebrovascular, gastrointestinal, or renal disease. Cerebral, splanchnic, and renal transit times were obtained for 15, 14, and 8 individuals, respectively. Anatomic variability of the renal vessels confounded the acquisition of renal transit times. For all organs, transit times were reproducible and the associated blood volumes were generally comparable to reference values. Cerebral, gastrointestinal, and renal transit times/blood volumes can be reasonably acquired from contrast ultrasound, though the latter is less reliably available. Assessment of the impact on regional blood volumes of pharmacologic or other interventions is a next step towards clinical application of this technique.

INTRODUCTION

Pulmonary transit time (PTT) and pulmonary blood volume (PBV) calculated from contrast-enhanced cardiac MRI provide independent prognostic information in non-fatal major cardiac adverse events [1] and both cardiac and all-cause mortality [2]. PTT derived from contrast-echocardiography is comparable to that obtained via cardiac MRI [3] and its acquisition can be integrated into the workflow of a clinical Echocardiography Laboratory [4]. Contrast ultrasound has also been used to measure transit-times outside the pulmonary circulation, most notably to identify and quantify the degree of portal-caval shunt in advanced cirrhosis [5].

It follows that transit-time analysis via contrast-enhanced ultrasound could potentially assist in characterizing pathophysiologic processes involving additional organs and vascular distributions. The cerebral, splanchnic, and renal circulations are frequently imaged with ultrasound in the clinical setting, but whether it is plausible to measure the corresponding transit-times accurately or reproducibly remains unknown.

Therefore, as a precursor to dedicated investigations in specific disease states, we conducted a pilot study to assess the technical feasibility of obtaining cerebral, splanchnic, and renal transit times using contrast-ultrasound.

THEORY

The present analysis generalizes estimation of PBV to the blood volume contained within any organ or vascular compartment. In the absence of intra-cardiac shunt, the pulmonary circulation receives the entire cardiac output. Therefore, PBV relative to the systemic stroke volume (SV), termed relative PBV (rPBV), is equivalent to the number of cardiac cycles (‘beats’) required for the PBV to traverse the pulmonary circulation, which is equal to the PTT ‘normalized’ by the heart rate (HR) (nPTT) [6-7]:

$$\begin{gathered} \text{PBV}=(\text{cardiac output})\ast \text{PTT} \\ =(\text{SV}\ast \text{HR})\ast \text{PTT} \end{gathered}$$ (1)

$$\text{PBV}∕\text{SV}=\text{nPTT}=\text{rPBV}$$ (2)

Generalizing (1) to derive an organ blood volume (OBV) from a specific organ’s transit time (OTT) requires accounting for the fact that only a fraction of the systemic SV, the fractional organ stroke volume (fOSV), is delivered to a given organ with each cardiac cycle.

$$\begin{gathered} \text{OBV}=[(\text{fOSV}\ast \text{cardiac output}]\ast \text{OTT} \\ =[(\text{OSV}∕\text{systemic SV})\ast (\text{systemic SV}\ast \text{HR})]\ast \text{OTT} \\ \text{OBV}∕\text{OSV}=\text{nOTT}=\text{rOBV} \end{gathered}$$ (3)

Analogous to (2) for PBV, estimation of a given OBV relative to the volume delivered to it with each cardiac cycle (rOBV) is given by (3). This metric provides the number of organ-specific stroke volumes contained within the vascular bed of that organ. No direct measurement of organ-specific stroke volume is necessary to obtain this ratio – only the organ-specific transit time and contemporaneous HR are required.

MATERIALS AND METHODS

This study was approved by our institution’s Institutional Review Board and all participants gave written informed consent.

Study population and setting

Healthy individuals without known pulmonary, cerebral, gastrointestinal, or renal disease were recruited to participate in the study. No participant reported a known allergy to perflutren lipid microsphere (DEFINITY^®; Lantheus Medical Imaging Inc, Billerica Massachusetts) contrast. One participant reported a history of obstructive sleep apnea, and another participant reported a history of ventricular non-compaction without cardiomyopathy. No other significant co-morbidities were present in any member of the cohort.

All participants had fasted for at least 6 hours prior to the beginning of the protocol and were encouraged to hydrate aggressively during the preceding 24 hours to optimize abdominal imaging windows. Demographics, basic anthropomorphic data, pregnancy status (where applicable), resting blood pressure and resting HR were obtained on the day of the study visit. Blood pressure and HR were recorded again at the end of the study. All components of study data acquisition were conducted at our institution’s Clinical Research Center (CRC), a dedicated research environment within our Medical Center through which qualified nursing and sonographer support was provided. A standard peripheral intravenous catheter (20 or 22 gauge) was inserted into either antecubital vein by an experienced research nurse.

Echocardiography and Vascular Ultrasound

For all participants, imaging was performed at rest by a single experienced sonographer on an EPIQ7C Matrix ultrasound instrument with echocardiographic, vascular, and abdominal capabilities (Philips, Amsterdam, The Netherlands). A limited echocardiogram (transducer X5-1c) assessed left ventricular ejection fraction (LVEF), left ventricular outflow tract (LVOT) diameter, LVOT velocity–time integral (VTI), mitral valve inflow velocity (E); mitral valve tissue Doppler velocity (e’), tricuspid annular plane systolic excursion (TAPSE), and right atrial pressure (using the inferior vena cava size and response to deep inspiration) according to American Society of Echocardiography guidelines.

After exclusion of inter-atrial shunt with agitated saline contrast studies at rest and with Valsalva, DEFINITY^® was used for ultrasound contrast administration as previously described [3-4, 6-7]. To estimate PTT, the entry of contrast into the right heart and subsequent appearance in the left heart was recorded in the apical 4-chamber view during normal respiration. The acquisition of regional transit-times occurred in the following order: pulmonary, cerebral, intestinal, renal. Each participant then underwent repeat imaging of one site to assess reproducibility. The reproducibility of PTT has been assessed previously [4] and thus repeat PTT measurements were not part of this evaluation.

The inflow-outflow vessel pairs for the brain, intestines, and kidney were as follows: internal carotid artery (ICA) and internal jugular vein (IJV); superior mesenteric artery (SMA) and portal vein; and the renal artery and renal vein, respectively. For these vessels, the following transducers were used: ICA/IJV – L12-3; SMA/portal vein – C5-1; renal artery/renal vein – C5-1. The right ICA/IJV pair was imaged with participants supine and their heads turned to the left. The SMA/portal vein pair was imaged with the participants lying supine also. The initial attempt to image the renal vessels was made with participants supine as well. If acceptable images of either kidney’s vasculature were not obtained, the right and left lateral decubitus positions were tried and the side that yielded the most favorable images was used. Prior to contrast injection at each site, the diameter of the relevant inflow vessel and the corresponding VTI were measured consecutively at the same location. After obtaining a stable image that allowed simultaneous visualization of the relevant vessel pair, the entry of contrast into the inflow vessel and subsequent appearance in the outflow vessel was recorded during normal respiration. Suspension of respiration was not used due to unpredictable arrival times of contrast to the inflow vessel. For measurement of cerebral, intestinal, and renal transit times, contrast was administered as a 0.5 mL bolus followed immediately by a 5 mL saline flush.

Each participant received approximately 0.5 mL of DEFINITY^® per injection and no participant received more than 3 mL of contrast over the course of the study. To facilitate clearance of contrast from the next vessel-pair to be imaged, participants walked for 15-20 minutes around the CRC between injections.

Organ transit times, rOBV, and OBV estimation

The recorded images of contrast traveling from the right heart to left heart and from inflow to outflow vessel were analyzed offline using QLab 10.7 software (Philips, Amsterdam, The Netherlands). The process for calculating PTT, rPBV, and PBV from these images has been described in detail previously [3, 6-7].

The regions of interest (ROI) for the non-cardiac sites were rectangular and non-uniformly sized, primarily due to anatomic variation between participants. For the inflow vessels, the ROIs were placed as close as possible to the site of diameter/VTI measurement. The outflow vessel ROIs were placed as ‘proximal’ as was feasible to minimize any non-organ related volume captured, particularly in longitudinally oriented vessels such as the IJV and renal vein.

After placement of ROIs, time-intensity curves of contrast passage through the inflow and outflow vessels were generated by the QLab software. These curves served as the inputs to our previously described inflection point algorithm, from which transit-times are calculated [8]. As described in the Theory section, rOBV was estimated as the product of OTT and average HR during the passage of contrast from inflow to outflow vessel. All participants were in sinus rhythm during the contrast injections. Absolute OBV was estimated as the product of rOBV and OSV, the latter calculated as follows: organ-specific VTI * organ-inflow vessel cross-sectional area. The fOSV was taken as the ratio of OSV to systemic SV (calculated as LVOT VTI * LVOT cross-sectional area; [9])

Data analysis

Demographic, anthropomorphic, and echocardiographic data are presented as mean ± standard deviation as are site specific stroke volumes and fOSVs. Transit times, relative blood volumes, and absolute blood volumes are given as ranges for each site. Reproducibility was assessed graphically with an X-Y plot of the initial and repeat measurements and more formally within the Bland-Altman framework.

RESULTS

Cohort characteristics and echocardiographic data

Table 1 shows general demographic, anthropomorphic, and echocardiographic characteristics. The cohort included 16 individuals (5 women, 11 men) that ranged in age from 23 to 68 years old (mean 42 ± 13 years) with a low prevalence of obesity. Only one participant had a BMI > 30 kg/m². All participants had normal biventricular systolic function and normal right and left-sided filling pressures by standard echocardiographic assessment.

Table 1 –

Cohort characteristics

Parameter	Multi-organ transit-time cohort (n = 16)
Demographic/anthropomorphic
Age (years)	42 ± 13
Gender (F/M)	5/11
Height (cm)	172 ± 11
Weight (kg)	75 ± 14
BMI (kg/m2)	25 ± 4
BSA (m2)	1.9 ± 0.2
Echocardiographic
Biplane Simpson’s LVEF (%)	64 ± 3
E/e’	8 ± 3
TAPSE (cm)	2.2 ± 0.5
Estimated RA pressure (mmHg)	4 ± 2

Data are presented as mean ± standard deviation

BMI body mass index

BSA body surface area

E mitral valve inflow velocity

e’ medial mitral valve annulus tissue Doppler velocity

RA right atrial

TAPSE tricuspid annular plane systolic excursion

Transit times and blood volume estimates

For all sites, transit times and relative blood volumes are displayed in Figure 1A and the corresponding absolute blood volume estimates are shown in Figure 1B. PTT was available for all individuals whereas inability to visualize the renal inflow and outflow vessels simultaneously precluded obtaining renal transit time in half of the participants. Technical challenges with the ultrasound machine led to the inability to acquire cerebral transit time in one subject and splanchnic transit time could not be obtained in two participants due to inadequate imaging windows.

Figure 1 – Transit-times and blood volumes for lung, brain, intestines, and kidney.

Individual organ transit times and relative blood volumes are shown in panel (A) and absolute blood volumes and absolute blood volumes indexed to body surface area are shown in panel (B). Note that the scale on the vertical axis in this panel is logarithmic. The number of participants with transit-time measurements for each site is as follows: lung – n = 16; brain – n = 15; intestines – n = 14; kidney – n = 8.

nOTT – normalized organ transit time (equivalent to relative organ blood volume)

The range of transit times and relative blood volumes for the various sites were as follows: lung: 2.6-9.5 seconds, 3.3-8.1 beats; brain: 4.2-12.5 seconds, 4.1-10.6 beats; intestines: 2.6-10.9 seconds, 2.9-9.2 beats; kidney: 2.2-9 seconds, 2.4-9.5 beats. The ranges of absolute blood volumes were (mL): lung: 176-670, brain: 32-97, intestines: 21-149, kidney: 13-45. These PTT, rPBV, and PBVi values are comparable to our prior findings using the same technique [7]. The brain, intestines, and kidney data are comparable to reference values for regional blood volumes (mL): brain – 48-111; intestines – 72-371; kidney – 9-53 [10].

Reproducibility of transit time measurements

Repeat measurements were available for 15 participants. We obtained reproducibility data for each site as follows: brain: 5 participants (33%), intestines: 6 participants (43%), kidney: 4 participants (50%). Across all sites, the average time elapsed between the original and follow-up measurements was 56 ± 32 minutes (18-135 minutes) and was site-dependent owing to the fixed order of regional transit-time acquisition. Figure 2 displays the reproducibility data as Bland-Altman plots (A) and an X-Y plot (B). The bias is relatively low for all sites with only one splanchnic transit time point falling outside the 2-standard deviation range. Correspondingly, that same point is an outlier in terms of deviation from the identity line.

Figure 2 – Reproducibility of organ transit times.

Panel (A) displays the Bland-Altman plot for organ transit time measurements. Bias is relatively low for each site and most data points fall within 2 standard deviations. The initial-repeat measurement pairs are plotted against each other in panel (B). Most pairs are close to the identity line. The number of participants with reproducibility measurements for each site is as follows: brain – n = 5; intestines – n = 6; kidney – n = 4.

Fractional organ stroke volume

For each site, fOSV is presented in Table 2. The estimates of fOSV for the gastrointestinal tract and kidney are comparable to standard values [11]. Our fOSV for the brain is higher, particularly when taking into consideration that our estimate focuses on the internal carotid artery and excludes the vertebral artery. However, more recent data suggest that brain fOSV may be higher than previously reported in younger women [12], who were relatively over-represented in our cohort.

Table 2 –

Organ-specific stroke volumes

Organ	Inflow vessel	Diameter (cm)	VTI (cm)	Stroke volume (mL)	fOSV (range)
Lung (n = 16)	LVOT	2.10 ± 0.20	21 ± 3	70 ± 18	1
Brain (n = 15)	ICA	0.45 ± 0.07	55 ± 12	9 ± 3	0.12 ± 0.03 (0.05-0.20)
Gut (n = 14)	SMA	0.59 ± 0.09	51 ± 14	14 ± 6	0.20 ± 0.08 (0.09-0.36)
Kidney (n = 8)	Renal artery	0.38 ± 0.10	62 ± 11	7 ± 3	0.10 ± 0.04 (0.05-0.16)

Data are presented as mean ± standard deviation

fOSV fractional organ stroke volume

ICA internal carotid artery

LVOT left ventricular outflow tract

SMA superior mesenteric artery

VTI velocity-time integral

DISCUSSION

This pilot study demonstrates that acquiring cerebral, intestinal, and renal transit times by contrast ultrasound is technically feasible and that the related organ blood volume estimates are reasonably accurate and precise. Transit-times and corresponding blood volumes can be obtained with any contrasted imaging modality for any organ or vascular bed in which blood flow through the inflow and outflow vessels can be imaged simultaneously (or sequentially). Comparatively, ultrasound is low cost, does not require ionizing radiation like CT, is not restricted by the presence of implanted metal devices or hardware as with MRI, and is less constrained than either CT or MRI by patient size, the presence of renal dysfunction, or the ability to lie flat. Additionally, its availability and portability allow for bedside deployment in the clinic, inpatient, or ICU settings.

Prior studies of cerebrovascular transit time have primarily used trans-cranial ultrasound [13-14], a technique that is both sensitive and specific for assessment of the intra-cranial vasculature but may have limited utility in patients with inadequate transcranial acoustic windows. In addition, trans-cranial ultrasound may be less widely available than standard ultrasound imaging. Given the choice of inflow and outflow vessels, our cerebral transit time and blood volume measurements reflect both macro- and microcirculatory properties of intra-cranial blood flow and thus may be dynamic due to cerebral autoregulation. Measurements can be further affected by contralateral vascular compromise, which can lead to relative hyperemia of the imaged vessel. In addition, the anatomic variability of the cerebral circulation suggests that the absolute volume estimates may need to be adjusted by an unknown factor of between 1 and 2, the former representing the extreme and unlikely scenario where the ICA perfuses and the ipsilateral IJV drains the entire cerebral blood volume and the latter a scenario where that vessel pair supplies and drains exclusively the ipsilateral hemisphere. This inherent heterogeneity may have also contributed to the observed overestimation of cerebral fOSV. Nonetheless, estimates of relative or absolute cerebral blood volume may have clinical utility, particularly in serial evaluations of perfusion disorders [15], such as before and after thrombolysis, mechanical intervention for acute stroke, or in the evaluation of vasospasm (i.e. reversible vasoconstriction syndrome or aneurysmal subarachnoid hemorrhage).

As the splanchnic circulation’s role in the physiology and management of heart failure continues to evolve [16], quantification of intestinal blood volume in response to interventions targeting this portion of the circulation may further elucidate the underlying mechanisms. The ability to estimate renal blood volume, in isolation or in conjunction with PBV, could contribute to further understanding of the cardio-renal syndrome, a complex disorder in which the relative roles of these volumes are difficult to assess [17-18]. Ultrasound-based assessment of the regional blood volumes measured in this study may also be useful in detecting differences in regional blood flow due to physiologic maneuvers such as meal intake or postural changes, which could be useful in diagnosing or gauging response to treatment of dysautonomias [19-22], a family of disorders that often involves abnormalities in the autonomic nervous system’s regulation of blood volume distribution.

Limitations

The small sample size as well as the young and generally healthy cohort may limit the generalizability of these results. Due to non-technical factors that cannot be systematically addressed (i.e. intrinsic anatomic variation), renal transit time could not be reliably or predictably obtained, which clearly restricts its use. As noted, there is unavoidable uncertainty in the exact vascular volume being assessed by the cerebral transit time. This uncertainty extends to the splanchnic transit time as well as the SMA does not necessarily perfuse all the small and large intestines. Ideally, to capture the entire gastrointestinal tract’s blood volume, the inflow vessel would be the portion of the abdominal aorta just proximal to the celiac artery, but visualizing this vessel in conjunction with the portal vein was not possible. Use of the portal vein as the splanchnic outflow vessel, though practical from an imaging standpoint, may be confounded by a contribution from splenic vein drainage. The absolute blood volume estimates include the fOSV, which introduces sources of error beyond those contained in the transit time and relative blood volume measurements (i.e. the LVOT and inflow vessel VTIs and cross-sectional areas). However, our absolute blood volume estimates were within the ranges reported in prior work, suggesting that the fOSV measurements are satisfactory. Furthermore, use of the relative metrics for serial measurements (i.e. before and after a change in position or a pharmacologic intervention) eliminates this error in addition to conferring the ability to quantify the change in organ volume while accounting for the change in organ stroke volume, analogous to the analysis of PBV and systemic SV in prior work involving measurement of PTT during exercise [7].

CONCLUSION

In summary, measurement of cerebral, splanchnic, and renal transit times and subsequent estimation of the respective blood volumes using contrast-ultrasound is feasible and the results are consistent with reference standards obtained using other modalities. Further exploration of this technique’s utility in characterizing disorders that involve these regional blood volumes seems reasonable.