Advertisement

Development of a Flow Phantom for Transcranial Doppler Ultrasound Quality Assurance

  • Author Footnotes
    1 These authors contributed equally.
    Fatmah Alablani
    Footnotes
    1 These authors contributed equally.
    Affiliations
    Radiology and Medical Imaging Department, College of Applied Medical Sciences, Prince Sattam bin Abdulaziz University, Kharj, Saudi Arabia

    Cerebral Haemodynamics in Aging and Stroke Medicine (CHIASM) Group, Department of Cardiovascular Sciences, University of Leicester, Leicester, UK
    Search for articles by this author
  • Author Footnotes
    1 These authors contributed equally.
    Justyna Janus
    Footnotes
    1 These authors contributed equally.
    Affiliations
    Cerebral Haemodynamics in Aging and Stroke Medicine (CHIASM) Group, Department of Cardiovascular Sciences, University of Leicester, Leicester, UK

    Medical Physics Department, University Hospitals of Leicester NHS Trust, Leicester, UK
    Search for articles by this author
  • Edward Pallett
    Affiliations
    Cerebral Haemodynamics in Aging and Stroke Medicine (CHIASM) Group, Department of Cardiovascular Sciences, University of Leicester, Leicester, UK

    Medical Physics Department, University Hospitals of Leicester NHS Trust, Leicester, UK
    Search for articles by this author
  • Toni M. Mullins
    Affiliations
    Medical University of South Carolina, Charleston, South Carolina, USA
    Search for articles by this author
  • Alanoud Almudayni
    Affiliations
    Radiology and Medical Imaging Department, College of Applied Medical Sciences, Prince Sattam bin Abdulaziz University, Kharj, Saudi Arabia

    Cerebral Haemodynamics in Aging and Stroke Medicine (CHIASM) Group, Department of Cardiovascular Sciences, University of Leicester, Leicester, UK
    Search for articles by this author
  • Emma M.L. Chung
    Correspondence
    Address correspondence to: Emma Chung, Faculty of Medicine and Life Sciences, King's College London, Shepherd's House, Guy's Campus, London, SE1 1UL, UK.
    Affiliations
    Cerebral Haemodynamics in Aging and Stroke Medicine (CHIASM) Group, Department of Cardiovascular Sciences, University of Leicester, Leicester, UK

    Medical Physics Department, University Hospitals of Leicester NHS Trust, Leicester, UK

    National Institute for Health Research Leicester Biomedical Research Centre, Leicester, UK

    Department of Women and Children's Health, Guy's Campus, King's College London, London, UK
    Search for articles by this author
  • Author Footnotes
    1 These authors contributed equally.
Open AccessPublished:August 26, 2022DOI:https://doi.org/10.1016/j.ultrasmedbio.2022.07.002

      Abstract

      Anecdotal evidence was recently brought to our attention suggesting a potential difference in velocity estimates between transcranial Doppler (TCD) systems when measuring high velocities (∼200 cm/s) close to the threshold for sickle cell disease stroke prevention. As we were unable to identify a suitable commercial TCD phantom, a middle cerebral artery (MCA) flow phantom was developed to evaluate velocity estimates from different devices under controlled conditions. Time-averaged velocity estimates were obtained using two TCD devices: a Spencer Technologies ST3 Doppler system (ST3 PMD150, Spencer Technologies, Seattle, WA, USA) and a DWL Dopplerbox (DWL Compumedics, SN-300947, Singen, Germany). These were compared with velocity estimates obtained using a Zonare duplex scanner (Zonare Medical Systems, Mountain View, CA, USA), with timed collection of fluid as the gold standard. Bland–Altman analysis was performed to compare measurements between devices. Our tests confirmed that velocities measured with the DWL TCD system were +4.1 cm/s (+3.7%; limits of agreement [LoA]: 2%, 5%; p = 0.03) higher than the Spencer system when measuring a velocity 110 cm/s and +12 cm/s higher (+5.7 %; LoA: 4.8%, 6.6%; p = 0.03) when measuring velocities of 210 cm/s, close to the diagnostic threshold for stroke intervention. We found our MCA phantom to be a valuable tool for systematically quantifying differences in TCD velocity estimates between devices, confirming that the DWL system gave consistently higher readings than the Spencer ST3 system. Differences become more pronounced at high velocities, which explains why they were not identified earlier. Our findings have clinical implications for centers using TCD to monitor patients with sickle cell disease, as extra care may be needed to adjust for bias between manufacturers when making treatment decisions about children with sickle cell with velocities close to the diagnostic threshold.

      Key Words

      Introduction

      Transcranial Doppler (TCD) ultrasound is used clinically for monitoring blood flow, detecting intracranial stenosis and diagnosing brain death (
      • Blanco P
      • Abdo-Cuza A.
      Transcranial Doppler ultrasound in neurocritical care.
      ;
      • Ali MF.
      Transcranial Doppler ultrasonography (uses, limitations and potentials): A review article.
      ). For example, the Optimizing Primary Stroke Prevention in Children with Sickle Cell Anemia (STOP II) trial reported that an abnormal time-averaged mean maximum velocity (TAMMV) detected using TCD, prompting regular blood transfusion therapy, results in a 92% reduction in stroke incidence in patients with sickle cell disease (
      • Adams RJ
      • McKie VC
      • Hsu L
      • Files B
      • Vichinsky E
      • Pegelow C
      • Abboud M
      • Gallagher D
      • Kutlar A
      • Nichols FT.
      Prevention of a first stroke by transfusions in children with sickle cell anemia and abnormal results on transcranial Doppler ultrasonography.
      ). The American Society of Hematology 2020 guidelines for sickle cell disease suggest a threshold for regular blood transfusion therapy of two non-imaging TAMMV TCD measurements ≥200 cm/s or a single measurement >220 cm/s in the proximal portion of the middle cerebral artery (MCA). As imaging transcranial color Doppler (TCCD) displays the time-averaged mean velocity (TAMV), which is lower than the TAMMV measured using non-imaging TCD, a threshold of two measurements with TAMV ≥185 cm/s or a single TAMV measurement >205 cm/s has also been proposed (
      • Chou ST
      • Alsawas M
      • Fasano RM
      • Field JJ
      • Hendrickson JE
      • Howard J
      • Kameka M
      • Kwiatkowski JL
      • Pirenne F
      • Shi PA.
      American Society of Hematology 2020 guidelines for sickle cell disease: Transfusion support.
      ).
      The aim of this study was to investigate anecdotal reports that systems from some manufacturers tend to provide slightly higher readings than other TCD systems across a range of velocities up to the diagnostic threshold of 200 cm/s for sickle cell disease stroke intervention. As this has potential implications for patient care, our Ultrasound Quality Assurance team at University Hospitals of Leicester National Health Service (UHL NHS) Trust agreed to develop a bespoke TCD phantom to test whether there were any differences in velocity estimates between manufacturers, as well as to explore the relationship between imaging TCD TAMV and non-imaging TCD TAMMV estimates.
      Non-imaging TCD systems are equipped with a relatively low 2-MHz-frequency ultrasound (US) probe to facilitate transmission of ultrasound through the skull. Backscattered echoes from moving red blood cells return to the transducer with a phase shift that can be analyzed to estimate the component of motion in the direction of the beam (
      • Aaslid R
      • Markwalder TM
      • Nornes H.
      Noninvasive transcranial Doppler ultrasound recording of flow velocity in basal cerebral arteries.
      ;
      • Purkayastha S
      • Sorond F.
      Transcranial Doppler ultrasound: Technique and application.
      ).
      The Doppler equation allows estimation of velocity
      Δf=2f0vcosθc
      (1)


      where Δf is the Doppler frequency shift measured using pulsed ultrasound with a transmitted frequency of fo, moving through tissue at the speed of sound, c =1540 m/s. The Doppler angle, θ, is the angle between the US beam and the direction of motion of red blood cells moving with velocity (v).
      The MCA carries up to 60% of carotid artery blood flow (
      • Rutgers D
      • Blankensteijn J
      • Van Der Grond J
      Preoperative MRA flow quantification in CEA patients: flow differences between patients who develop cerebral ischemia and patients who do not develop cerebral ischemia during cross-clamping of the carotid artery.
      ;
      • Moppett I
      • Mahajan R.
      Transcranial Doppler ultrasonography in anaesthesia and intensive care.
      ) and is commonly used for TCD monitoring because of the ease of access through the temporal bone window and favorable Doppler angle. On the basis of computed tomography (CT) studies, the angle between the beam and MCA is approximately 30° (
      • Ringelstein E
      • Kahlscheuer B
      • Niggemeyer E
      • Otis S.
      Transcranial Doppler sonography: Anatomical landmarks and normal velocity values.
      ). As non-imaging TCD systems assume a Doppler angle of 0°, MCA velocities displayed by TCD correspond to 77%–93% of the true (angle-corrected) MCA flow velocity (
      • Ainslie PN
      • Hoiland RL.
      Transcranial Doppler ultrasound: Valid, invalid or both?.
      ). It is also worth emphasizing that the “mean velocity” displayed by non-imaging TCD systems does not correspond to the mean flow velocity within the vessel, but to the much higher “time-averaged mean of the maximum velocity” envelope (or TAMMV). For laminar flow conditions, the TAMMV is approximately twice the TAMV (Fig. 1).
      Fig 1
      Fig. 1Diagram clarifying the difference between the time-averaged mean maximum velocity (TAMMV), often displayed as the “mean” on non-imaging transcranial Doppler (TCD), and the time-averaged mean velocity (TAMV) estimated by TCCD (or non-imaging TCD), which corresponds to the mean flow velocity. For laminar flow, the TAMMV would be twice the TAMV.
      Transcranial color doppler (TCCD) duplex imaging provides an attractive alternative to TCD, facilitating angle correction based on the B-mode image and inclusion of color-flow information. Underestimation of velocities caused by the absence of angle correction, coupled with overestimation resulting from display of the “mean of the maximum envelope” means that TCD and TCCD (or imaging TCD) are measuring different quantities (TAMMV and TAMV), summarized in Figure 1. Diagnostic thresholds for the two techniques will therefore not be directly interchangeable, as systems are calculating and displaying different velocity measures. This is reflected in the current sickle cell intervention guidelines which state a lower threshold for measurements made using imaging systems compared with non-imaging systems (abnormal 200 cm/s TAMMV corresponds to a TAMV threshold of 185 cm/s on imaging TCD machines) (
      • Jones A
      • Granger S
      • Brambilla D
      • Gallagher D
      • Vichinsky E
      • Woods G
      • Berman B
      • Roach S
      • Nichols F
      • Adams RJ.
      Can peak systolic velocities be used for prediction of stroke in sickle cell anemia?.
      ;
      • Jordan LC
      • Casella JF
      • Debaun MR.
      Prospects for primary stroke prevention in children with sickle cell anaemia.
      ). Despite several studies, the appropriate thresholds for treatment remain a point of some uncertainty. Assuming laminar flow conditions, it becomes possible to estimate TCCD TAMV based on halving the TCD TAMMV and then applying a 30° angle correction. This is equivalent to applying a scaling factor of 0.57. However, this conversion will be valid only at low flow velocities consistent with the assumption of laminar parabolic flow.
      TAMVmeasuredusingTCCD=0.57TAMMVonTCD
      (2)


      Vascular flow phantoms have a role to play in the development and testing of blood flow measurement systems. Accuracy in velocity estimates across the full range of clinically relevant velocity measurements should be reported in the technical specifications of all new commercial TCD devices. According to the International Electrotechnical Commission, phantoms should be developed using stable and reproducible components with acoustical and physical properties that are physiologically realistic and comparable to those of human tissue, vessels and blood (
      • Ramnarine KV
      • Nassiri DK
      • Hoskins PR
      • Lubbers J.
      Validation of a new blood-mimicking fluid for use in Doppler flow test objects.
      ). To fulfil these requirements, blood-mimicking fluid (BMF), tissue-mimicking material (TMM) and C-flex tubing are often used in vascular phantom development (
      • Ramnarine KV
      • Nassiri DK
      • Hoskins PR
      • Lubbers J.
      Validation of a new blood-mimicking fluid for use in Doppler flow test objects.
      ,
      • Ramnarine KV
      • Hoskins PR
      • Routh HF
      • Davidson F.
      Doppler backscatter properties of a blood-mimicking fluid for Doppler performance assessment.
      ;
      • Lubbers J.
      Application of a new blood-mimicking fluid in a flow Doppler test object.
      ).
      Although several Doppler phantoms are commercially available (
      • Yonan K
      • Greene E
      • Sharrar J
      • Caprihan A
      • Qualls C
      • Roldan C.
      Middle cerebral artery blood flows by combining TCD velocities and MRA diameters: In vitro and in vivo validations.
      ;
      • Kim SK
      • Kwak HS
      • Chung GH
      • Hwang SB.
      Why is middle cerebral artery plaque augmented by contrast media? A phantom study using middle cerebral artery stenotic silicon model.
      ;
      • Bakenecker AC
      • Von Gladiss A
      • Schwenke H
      • Behrends A
      • Friedrich T
      • Lüdtke-Buzug K
      • Neumann A
      • Barkhausen J
      • Wegner F
      • Buzug TM.
      Navigation of a magnetic micro-robot through a cerebral aneurysm phantom with magnetic particle imaging.
      ), these are intended to mimic larger extracranial vessels, such as the carotid arteries, and the range of velocities represented is limited. To the best of our knowledge, this is the first experimental study to compare MCA velocity readings between TCD manufacturers, as well as comparing velocities displayed on TCD machines against TCCD estimates and reference values obtained by timed collection of fluid. This in vitro study aimed to compare phantom “MCA” velocity estimates between two commercially available TCD machines (the DWL and Spencer ST3) to determine whether there are significant differences between MCA velocity readings, especially in the ∼200 cm/s velocity range relevant to stroke prevention in patients with sickle cell anemia (
      • Chou ST
      • Alsawas M
      • Fasano RM
      • Field JJ
      • Hendrickson JE
      • Howard J
      • Kameka M
      • Kwiatkowski JL
      • Pirenne F
      • Shi PA.
      American Society of Hematology 2020 guidelines for sickle cell disease: Transfusion support.
      ). As part of this study, we also aimed to quantify differences between MCA velocity estimates based on imaging TCD and non-imaging TCD (TCCD).

      Methods

      Phantom construction and setup

      A 3-mm-internal-diameter C-flex tube (Cole-Parmer, London, UK) was placed in a container at an angle of 30° to mimic the insonation geometry of the MCA. This was then surrounded by a validated agar-based TMM with speed of sound, attenuation and backscatter properties similar to those of soft tissue (Fig. 2) (
      • Ramnarine KV
      • Anderson T
      • Hoskins PR.
      Construction and geometric stability of physiological flow rate wall-less stenosis phantoms.
      ). The tubing was connected to a programmable gear pump (UHL Medical Physics Department), which was used to circulate a validated BMF (
      • Ramnarine KV
      • Hoskins PR
      • Routh HF
      • Davidson F.
      Doppler backscatter properties of a blood-mimicking fluid for Doppler performance assessment.
      ,
      • Ramnarine KV
      • Anderson T
      • Hoskins PR.
      Construction and geometric stability of physiological flow rate wall-less stenosis phantoms.
      ). The BMF was then filtered using a 40-μm sieve and de-gassed by running the fluid through the pump for at least an hour before taking each set of measurements. Air bubbles and particles were reduced by placing filter fabric material (40-denier tights, Marks and Spencer, UK) over the entrance to a funnel bubble trap (Fig. 2). The programmable gear pump was used to generate controlled steady flow, and pulsatile (60 bpm) flow, of the BMF across a range of clinically relevant velocity values. These specifically included the lowest and highest values of steady flow typically observed during cardiopulmonary bypass of ∼20–100 cm/s (
      • Polito A
      • Ricci Z
      • Di Chiara L
      • Giorni C
      • Iacoella C
      • Sanders SP
      • Picardo S.
      Cerebral blood flow during cardiopulmonary bypass in pediatric cardiac surgery: The role of transcranial Doppler—A systematic review of the literature.
      ), as well as normal “textbook” pulsatile flow at ∼55 cm/s (
      • Aaslid R
      • Markwalder TM
      • Nornes H.
      Noninvasive transcranial Doppler ultrasound recording of flow velocity in basal cerebral arteries.
      ), and high pulsatile flow at the sickle cell diagnostic threshold of ∼200 cm/s (
      • Chou ST
      • Alsawas M
      • Fasano RM
      • Field JJ
      • Hendrickson JE
      • Howard J
      • Kameka M
      • Kwiatkowski JL
      • Pirenne F
      • Shi PA.
      American Society of Hematology 2020 guidelines for sickle cell disease: Transfusion support.
      ).
      Fig 2
      Fig. 2Phantom setup, fabrication and schematic of flow circuit. MCA = middle cerebral artery; US = ultrasound; V = velocity.

      Phantom validation

      Reference values: Timed collection method

      The true velocity of the BMF was measured by breaking the circuit and collecting the fluid into a measuring cylinder over a set period. The volume flow was measured and calculated as an average of three measurements. The velocity was then obtained using the formula
      v=Q/A
      (3)


      where v is the velocity (in cm/s), Q is a volume flow rate (in cm3) and A is the cross-sectional area (πr2) of the 3-mm tube (in cm2).

      Reference values: Angle-corrected TCCD ultrasound measurements

      A portable ultrasound imaging device (Zonare Medical Systems, Mountain View, CA, USA), equipped with an L10-5 linear array transducer, was used to estimate the angle-corrected time-averaged mean velocity (TAMV) at each setting to confirm the flow rate before acquiring each non-imaging TCD measurement. TAMV estimates were compared with timed collection as the gold standard (Table 1). Note that duplex ultrasound imaging, combining B-mode and pulsed wave (PW), was not feasible for measuring velocities at ∼200 cm/s because of aliasing. It is well known that the pulse repetition frequency (PRF) needs to be at least twice the Doppler shift being measured, which we were unable to achieve with our imaging system for the 5-cm depth of the sample.
      Table 1Mean velocity estimates and agreement for each of the four flow conditions tested
      The TAMMV values have not been corrected for Doppler angle.
      Flow conditionTAMV estimateTAMMV estimate
      Timed-collection (gold standard) (cm/s)TCCD ultrasound (cm/s)Bias (%) (LoA) p value
      Wilcoxon signed rank test with statistical significance at p ≤ 0.05. The bias for the TAMV estimates compares timed-collection with angle-corrected TCCD estimates. The bias for the TAMMV estimates compares measurements made using the two TCD devices without correction for angle.
      Spencer (cm/s)DWL (cm/s)Bias (%) (LoA) p value
      Wilcoxon signed rank test with statistical significance at p ≤ 0.05. The bias for the TAMV estimates compares timed-collection with angle-corrected TCCD estimates. The bias for the TAMMV estimates compares measurements made using the two TCD devices without correction for angle.
      Low steady flow9.05 (9–9.1)9 (9–9)0.5 (–0.6, 1.8), p = 0.518 (17–18)18 (18–18)1.9 (–4.5, 8), p > 0.9
      Normal pulsatile flow33 (32.5–33)33 (33-33)0.5 (–1, 2), p > 0.957 (57–57)57.5 (57–58)0.4 (–1.2, 2), p > 0.9
      Flow during CPB74.5 (74.3–74.5)74 (74–75)–0.13 (–1.9, 1.6), p > 0.9107.5 (107–108)111.5 (111–112)3.7 (2, 5), p = 0.03
      Very high pulsatile flow
      For example, in sickle cell anemia.
      152.4 (152–152.6)N/AN/A195 (195–207)207 (207–2018)5.7 (4.8, 6.6), p = 0.03
      CPB = cardiopulmonary bypass; LoA = limits of agreement; TAMMV = time-averaged mean maximum velocity; TAMV = time-averaged mean velocity; TCCD = transcranial color Doppler.
      low asterisk The TAMMV values have not been corrected for Doppler angle.
      Wilcoxon signed rank test with statistical significance at p ≤ 0.05. The bias for the TAMV estimates compares timed-collection with angle-corrected TCCD estimates. The bias for the TAMMV estimates compares measurements made using the two TCD devices without correction for angle.
      For example, in sickle cell anemia.

      Transcranial Doppler ultrasound assessment

      Two commercial TCD machines, the Doppler Box (DWL Compumedics, Germany) and Spencer TCD (ST3 PMD150, Spencer Technologies, Seattle, WA, USA), equipped with standard 2-MHz transducers were used to insonate the phantom vessel at a depth of 5 cm. Ultrasound coupling gel was applied between the transducer and the surface of the phantom. The power M-mode display was used to determine the depth of the strongest signal, and then the probe was moved along the surface of the phantom until the depth was 5 cm. The probe was fixed perpendicular to the phantom's surface (θ = 30°) using a clamp. Equipment settings included a sample depth of 50 mm, gate size of 10 mm and wall filter set to “low.” The velocity scale was set sufficiently high to prevent aliasing artefacts. The Doppler gain was set just below saturation. The time-averaged mean of the maximum velocity envelope (TAMMV value) displayed on the screen was recorded three times for each flow setting. Although measurements for each flow condition were carried out on separate days, measurements were taken using both test devices on the same day under identical flow conditions, and measurements were alternated between the two devices being tested. The imaging TCCD Zonare measurements were performed at the start and end of each measurement session to confirm that there had been no change in conditions over the duration of the experiment. Timed collection of fluid was conducted last.
      The phantom was initially set up based on a typical MCA waveform with a mean velocity of 55 cm/s obtained from a separate research study (University of Leicester ethics approval, Ethics Reference: 24267-aa1171-ls: cardiovascular sciences). All volunteers to this study gave written informed consent for their anonymized data to be used in future research. The phantom's programmable pump TAMMV settings were adjusted until flow in the phantom closely resembled that of the healthy volunteer (Fig. 3).
      Fig 3
      Fig. 3Comparison of (A) human and (B) phantom transcranial Doppler measurements used for phantom validation. From these images it can be clearly seen that the mean velocity of 55 cm/s displayed on the screen is the mean of the maximum velocity envelope (TAMMV) rather than the time-averaged mean velocity (TAMV), which would be lower than the diastolic value. PI = pulsatility index.
      Measurements were obtained across a range of clinically relevant settings intended to mimic (i) very low steady flow on cardiopulmonary bypass (CPB) (
      • Polito A
      • Ricci Z
      • Di Chiara L
      • Giorni C
      • Iacoella C
      • Sanders SP
      • Picardo S.
      Cerebral blood flow during cardiopulmonary bypass in pediatric cardiac surgery: The role of transcranial Doppler—A systematic review of the literature.
      ), (ii) normal textbook pulsatile flow (∼55 cm/s) (
      • Aaslid R
      • Markwalder TM
      • Nornes H.
      Noninvasive transcranial Doppler ultrasound recording of flow velocity in basal cerebral arteries.
      ), (iii) high steady flow during CPB (
      • Burrows FA.
      Transcranial Doppler technology: The noninvasive monitoring of cerebral perfusion during cardiopulmonary bypass. In: Brain injury and pediatric cardiac surgery.
      ) and (iv) very high pulsatile flow found in sickle cell disease (∼200 cm/s) (
      • Yawn BP
      • Buchanan GR
      • Afenyi-Annan AN
      • Ballas SK
      • Hassell KL
      • James AH
      • Jordan L
      • Lanzkron SM
      • Lottenberg R
      • Savage WJ.
      Management of sickle cell disease: Summary of the 2014 evidence-based report by expert panel members.
      ).

      Statistical analysis

      Measured velocity distributions were checked for normality using a Shapiro–Wilks test and found to be non-normal. Continuous data were summarized as a median and range. Agreement between measurement methods was assessed using a correlation plot with Bland–Altman analysis used to estimate the bias ± 2 SD (percentage difference between measurements) and limits of agreement (LoA). For group comparisons, paired t-tests were applied, with significance assumed at a conventional p value ≤0.05. Statistical analysis was performed using Prism 7 (GraphPad Software, San Diego, CA, USA).

      Results

      Time-averaged mean flow velocity (TAMV) confirmed by breaking the circuit and measuring the flow of fluid into a measuring cylinder revealed an average volume flow under each test condition of (i) 40 mL/min for low (steady) flow, (ii) 142 mL/min for normal (pulsatile) flow, (iii) 316 mL/min for steady flow under CPB and (iv) 647 mL/min for very high pulsatile flow conditions. These corresponded via eqn (3) to the following flow rates: (i) 9 cm/s (low, steady flow), (ii) 33 cm/s (normal pulsatile flow), (iii) 74.4 cm/s (normal steady flow) and (iv) 152 cm/s (high pulsatile flow) (Table 1). Spectrograms from all of the flow profiles tested closely resembled physiological conditions, with a similar pulsatility index (PI), backscatter intensity, and TAMMV as observed clinically. Although we did not have access to publishable comparative clinical examples, our team has extensive experience in clinical TCD monitoring and were able to reproduce clinically observed MCA waveforms with high realism.
      TCCD angle-corrected time-averaged mean velocities (TAMVs) were measured three times at each setting to confirm the flow rate before and after acquiring each TCD measurement. No statistically significant differences were observed between angle-corrected TCCD estimates of TAMV and the true mean flow velocity determined by timed collection. In all cases, the bias between TCCD and the gold standard was <0.5 %, with p > 0.05, which is suggestive of no significant difference between timed collection and TCCD (see Table 1, TAMV estimates). Note that for abnormal high flow, we were unable to perform a TCCD measurement because of aliasing. The reason we performed both TCCD and timed collection measurements was that TCCD can be performed at the start and end of each experiment to verify the measurement conditions without disturbing the flow.
      It can be seen from comparison of the TAMMV estimates on the right-hand side of Table 1 that the DWL system displays consistently higher readings than the Spencer ST3 system. Although differences between the two devices were not significant for velocities in the low to normal range, at a TAMMV of ∼110 cm/s a significant bias of 3.7% (LoA: 2%, 5%; p = 0.03) begins to emerge, which is equivalent to the DWL system reading +4.1 cm/s higher than the Spencer system (p = 0.002). This increased to 5.7% (LoA: 4.8%, 6.6%; p = 0.03) at 211 cm/s, equivalent to the DWL system reading +12 cm/s higher than the Spencer device (p = 0.0008).
      The growing percentage bias between the Spencer and DWL devices can also be seen as a deviation from the line of perfect agreement in Figure 4A. Figure 4B is the Bland–Altman plot based on percentage differences (n = 14). Across the full range of velocities tested (9–210 cm/s), DWL readings were, on average, 2.9% higher than those of the Spencer device.
      Fig 4
      Fig. 4Direct comparison of transcranial Doppler time-averaged mean maximum velocity (TAMMV) values reveals a divergence from the line of perfect agreement at high velocities (Panel A). Panel B: Bland–Altman analysis, assessing agreement between the DWL and Spencer devices (as a percentage of the TAMMV), suggests an overall bias of 2.9% and limits of agreement ranging from –2% to 8%.
      It is also of interest to compare TCD TAMMV estimates with timed collection reference values (Table 1). At low flow velocities, TAMMV values are broadly consistent with the multiplication factor of 1.74, based on assumption of a 30° Doppler angle and laminar flow. However, as velocities increase, this relationship breaks down and TAMMV and TAMV values start to become closer together (although the TAMV is always far below the TAMMV value). This is likely due to the assumption of laminar flow being invalid at high velocities, such that vmax ≠ 2vmean. Extreme care should be taken in moving between TAMMV and TAMV estimates, as the scaling factor appears to be velocity dependent. This means that diagnostic thresholds, or normative values, initially defined using non-imaging TCD cannot readily be converted to TCCD (TAMV) estimates, or vice versa.

      Discussion

      Phantoms provide an important means of ensuring medical ultrasound devices are functioning within expected parameters; however, TCD quality assurance is rarely performed in clinical practice. This is possibly due to the assumption that uncertainties in Doppler angle and other limitations of TCD velocity estimation are likely to affect measurements more than a systematic bias between devices. Our results suggest that sickle cell anemia patients with velocities close to the published threshold for intervention will be more likely to be referred for stroke prevention therapy if scanned with a DWL TCD machine than a Spencer ST3. A clinical audit would be needed to establish how many children may have had been allocated a different treatment pathway because their measurements were within the 12 cm/s average difference between manufacturers.
      This study developed an MCA flow phantom to facilitate comparison of imaging and non-imaging velocity estimates across a range of clinically relevant flow velocities and two non-imaging TCD manufacturers (Spencer and DWL). The phantom is suitable for adult or pediatric TCD evaluation as the waveforms and depth are adjustable; however, all of the measurements obtained in the current study were obtained at a depth of 5 cm. The phantom was validated by comparing the waveform shape, backscatter intensity and velocity of a healthy volunteer to our in vitro setup. This indicated that our laboratory phantom could reproduce a typical pulsatile MCA waveform with a TAMMV of 55 cm/s.
      Non-imaging DWL TCD measurements were only significantly different at high velocities; close to the sickle cell stroke intervention threshold of 200 cm/s, TAMMV, measurements obtained using the DWL system were ∼12 cm/s higher than the Spencer estimates.
      Gold standard reference measurements (from timed collection) were in good agreement with angle-corrected TCCD TAMV values, and confirmed that the “mean” displayed by non-imaging TCD systems refers to the “mean of the peak velocity envelope” (TAMMV), which is around twice the mean velocity for low flow velocities (TAMV < 50 cm/s). As flow increases, the flow is no longer laminar, and this relationship no longer holds. These results confirm and agree with previous studies (
      • Jones AM
      • Seibert JJ
      • Nichols FT
      • Kinder DL
      • Cox K
      • Luden J
      • Carl EM
      • Brambilla D
      • Saccente S
      • Adams RJ.
      Comparison of transcranial color Doppler imaging (TCDI) and transcranial Doppler (TCD) in children with sickle-cell anemia.
      ;
      • Krejza J
      • Rudzinski W
      • Pawlak M
      • Tomaszewski M
      • Ichord R
      • Kwiatkowski J
      • Gor D
      • Melhem E.
      Angle-corrected imaging transcranial Doppler sonography versus imaging and nonimaging transcranial Doppler sonography in children with sickle cell disease.
      ) and American Society for Hematology (ASH) guidelines (
      • Liem RI
      • Lanzkron SD
      • Coates T
      • Decastro L
      • Desai AA
      • Ataga KI
      • Cohen RT
      • Haynes Jr, J
      • Osunkwo I
      • Lebensburger JD.
      American Society of Hematology 2019 guidelines for sickle cell disease: Cardiopulmonary and kidney disease.
      ;
      • Chou ST
      • Alsawas M
      • Fasano RM
      • Field JJ
      • Hendrickson JE
      • Howard J
      • Kameka M
      • Kwiatkowski JL
      • Pirenne F
      • Shi PA.
      American Society of Hematology 2020 guidelines for sickle cell disease: Transfusion support.
      ) suggesting differing thresholds for sickle cell stroke intervention are needed for non-imaging TCD compared with imaging TCD. The original STOP study diagnostic thresholds were arrived at on the basis of measurements obtained using a non-imaging Nicolet TCD (
      • Nichols FT
      • Jones AM
      • Adams RJ.
      Stroke Prevention in Sickle Cell Disease (STOP) study guidelines for transcranial Doppler testing.
      ) and may need to be systematically adjusted to account for differences between non-imaging TCD manufacturers and imaging TCCD (with differing thresholds for TAMMV vs. TAMV, as suggested within the ASH guidance).
      TCD manufacturers could also do more to allow estimation of TAMV by introducing angle-correction pre-sets based on typical insonation geometry and additional signal processing of the spectrogram to quantify the TAMV. It is surprising to us that accuracy assessment and calibration across the full range of clinically measured velocities are not a requirement of medical device regulatory approvals.
      Care should be exercised when interpreting the results of previous research, as researchers frequently refer to the “mean” without specifying whether this is the TAMMV or TAMV. Extreme care should be taken in converting TAMMV to TAMV based on an assumption of laminar parabolic flow as this does not appear to be a valid assumption for TAMV >50 cm/s.

      Conclusions

      An MCA velocity ultrasound Quality Assurance phantom was developed to investigate differences in velocity estimates between TCD systems. Our in vitro results indicate a small percentage difference between the TCD systems we tested, which became significant at high velocities relevant to the threshold for sickle cell stroke intervention.

      Acknowledgments

      The PhD student conducting this study (F.A.) is funded by the Radiology and Medical Imaging at Prince Sattam bin Abdulaziz University.

      Conflict of interest disclosure

      The authors declare that they have no conflicts of interest.

      References

        • Aaslid R
        • Markwalder TM
        • Nornes H.
        Noninvasive transcranial Doppler ultrasound recording of flow velocity in basal cerebral arteries.
        J Neurosurg. 1982; 57: 769-774
        • Adams RJ
        • McKie VC
        • Hsu L
        • Files B
        • Vichinsky E
        • Pegelow C
        • Abboud M
        • Gallagher D
        • Kutlar A
        • Nichols FT.
        Prevention of a first stroke by transfusions in children with sickle cell anemia and abnormal results on transcranial Doppler ultrasonography.
        N Engl J Med. 1998; 339: 5-11
        • Ainslie PN
        • Hoiland RL.
        Transcranial Doppler ultrasound: Valid, invalid or both?.
        American Physiological Society, Bethesda, MD2014
        • Ali MF.
        Transcranial Doppler ultrasonography (uses, limitations and potentials): A review article.
        Egypt J Neurosurg. 2021; 36: 1-9
        • Bakenecker AC
        • Von Gladiss A
        • Schwenke H
        • Behrends A
        • Friedrich T
        • Lüdtke-Buzug K
        • Neumann A
        • Barkhausen J
        • Wegner F
        • Buzug TM.
        Navigation of a magnetic micro-robot through a cerebral aneurysm phantom with magnetic particle imaging.
        Sci Rep. 2021; 11: 1-12
        • Blanco P
        • Abdo-Cuza A.
        Transcranial Doppler ultrasound in neurocritical care.
        J Ultrasound. 2018; 21: 1-16
        • Burrows FA.
        Transcranial Doppler technology: The noninvasive monitoring of cerebral perfusion during cardiopulmonary bypass. In: Brain injury and pediatric cardiac surgery.
        CRC Press, Boca Raton, FL2019: 129-142
        • Chou ST
        • Alsawas M
        • Fasano RM
        • Field JJ
        • Hendrickson JE
        • Howard J
        • Kameka M
        • Kwiatkowski JL
        • Pirenne F
        • Shi PA.
        American Society of Hematology 2020 guidelines for sickle cell disease: Transfusion support.
        Blood Adv. 2020; 4: 327-355
        • Jones AM
        • Seibert JJ
        • Nichols FT
        • Kinder DL
        • Cox K
        • Luden J
        • Carl EM
        • Brambilla D
        • Saccente S
        • Adams RJ.
        Comparison of transcranial color Doppler imaging (TCDI) and transcranial Doppler (TCD) in children with sickle-cell anemia.
        Pediatr Radiol. 2001; 31: 461-469
        • Jones A
        • Granger S
        • Brambilla D
        • Gallagher D
        • Vichinsky E
        • Woods G
        • Berman B
        • Roach S
        • Nichols F
        • Adams RJ.
        Can peak systolic velocities be used for prediction of stroke in sickle cell anemia?.
        Pediatr Radiol. 2005; 35: 66-72
        • Jordan LC
        • Casella JF
        • Debaun MR.
        Prospects for primary stroke prevention in children with sickle cell anaemia.
        Br J Haematol. 2012; 157: 14-25
        • Kim SK
        • Kwak HS
        • Chung GH
        • Hwang SB.
        Why is middle cerebral artery plaque augmented by contrast media? A phantom study using middle cerebral artery stenotic silicon model.
        Neuroradiology. 2019; 61: 1173-1180
        • Krejza J
        • Rudzinski W
        • Pawlak M
        • Tomaszewski M
        • Ichord R
        • Kwiatkowski J
        • Gor D
        • Melhem E.
        Angle-corrected imaging transcranial Doppler sonography versus imaging and nonimaging transcranial Doppler sonography in children with sickle cell disease.
        AJNR Am J Neuroradiol. 2007; 28: 1613-1618
        • Liem RI
        • Lanzkron SD
        • Coates T
        • Decastro L
        • Desai AA
        • Ataga KI
        • Cohen RT
        • Haynes Jr, J
        • Osunkwo I
        • Lebensburger JD.
        American Society of Hematology 2019 guidelines for sickle cell disease: Cardiopulmonary and kidney disease.
        Blood Adv. 2019; 3: 3867-3897
        • Lubbers J.
        Application of a new blood-mimicking fluid in a flow Doppler test object.
        Eur J Ultrasound. 1999; 9: 267-276
        • Moppett I
        • Mahajan R.
        Transcranial Doppler ultrasonography in anaesthesia and intensive care.
        Br J Anaesth. 2004; 93: 710-724
        • Nichols FT
        • Jones AM
        • Adams RJ.
        Stroke Prevention in Sickle Cell Disease (STOP) study guidelines for transcranial Doppler testing.
        J Neuroimaging. 2001; 11: 354-362
        • Polito A
        • Ricci Z
        • Di Chiara L
        • Giorni C
        • Iacoella C
        • Sanders SP
        • Picardo S.
        Cerebral blood flow during cardiopulmonary bypass in pediatric cardiac surgery: The role of transcranial Doppler—A systematic review of the literature.
        Cardiovasc Ultrasound. 2006; 4: 47
        • Purkayastha S
        • Sorond F.
        Transcranial Doppler ultrasound: Technique and application.
        Semin Neurol. 2012; 32: 411-420
        • Ramnarine KV
        • Nassiri DK
        • Hoskins PR
        • Lubbers J.
        Validation of a new blood-mimicking fluid for use in Doppler flow test objects.
        Ultrasound Med Biol. 1998; 24: 451-459
        • Ramnarine KV
        • Hoskins PR
        • Routh HF
        • Davidson F.
        Doppler backscatter properties of a blood-mimicking fluid for Doppler performance assessment.
        Ultrasound Med Biol. 1999; 25: 105-110
        • Ramnarine KV
        • Anderson T
        • Hoskins PR.
        Construction and geometric stability of physiological flow rate wall-less stenosis phantoms.
        Ultrasound Med Biol. 2001; 27: 245-250
        • Ringelstein E
        • Kahlscheuer B
        • Niggemeyer E
        • Otis S.
        Transcranial Doppler sonography: Anatomical landmarks and normal velocity values.
        Ultrasound Med Biol. 1990; 16: 745-761
        • Rutgers D
        • Blankensteijn J
        • Van Der Grond J
        Preoperative MRA flow quantification in CEA patients: flow differences between patients who develop cerebral ischemia and patients who do not develop cerebral ischemia during cross-clamping of the carotid artery.
        Stroke. 2000; 31: 3021-3028
        • Yawn BP
        • Buchanan GR
        • Afenyi-Annan AN
        • Ballas SK
        • Hassell KL
        • James AH
        • Jordan L
        • Lanzkron SM
        • Lottenberg R
        • Savage WJ.
        Management of sickle cell disease: Summary of the 2014 evidence-based report by expert panel members.
        JAMA. 2014; 312: 1033-1048
        • Yonan K
        • Greene E
        • Sharrar J
        • Caprihan A
        • Qualls C
        • Roldan C.
        Middle cerebral artery blood flows by combining TCD velocities and MRA diameters: In vitro and in vivo validations.
        Ultrasound Med Biol. 2014; 40: 2692-2699