Introduction
Over the last 20years, Doppler ultrasound has been increasingly used for the monitoring of fetal well being in a wide variety of clinical scenarios, most commonly in the investigation of the small for gestational age (SGA) fetus and intrauterine growth restriction (IUGR). With modern ultrasound machines, obtaining Doppler signals from fetal and maternal vessels has become quick and easy. We can confidently target umbilical arteries, vessels inside the fetal skull (middle cerebral artery, MCA) and tiny vessels with diameters in the single millimetre range (ductus venosus).
Doppler tests are generally only used in high‐risk pregnancies and obstetricians heavily rely on the results in determining the management pathway for the mother and baby. Often the decision to deliver or continue the pregnancy is made on the basis of the Doppler results. It is therefore vital that Doppler assessments are performed correctly. And therein lies the catch. As with many ultrasound applications, things may not always be as easy as they seem. Despite a large body of literature, good training schemes and widely publicised international best‐practice guidelines to help the sonographer or sonologist, there are a number of ways that errors regularly creep into our assessments. Most of these errors fall into one of the following categories:
Technical acquisition errors.
Errors related to physiologic factors.
Incorrect application of measurements.
Are existing guidelines adequate?
The guidelines that are most commonly used by sonographers and sonologists in Australia and New Zealand are the International Society for Ultrasound in Obstetrics and Gynecology (ISUOG) guideline1 and the New Zealand Maternal Fetal Medicine Network (NZMFMN) guideline.2 Both guidelines provide the practitioner with comprehensive technical information. While the ISUOG guideline focuses on technique only, the NZMFMN guideline additionally provides recommended reference charts and tables.
Unfortunately, these guidelines carry certain limitations in terms of their general applicability. The first problem is that the guidelines are relatively complex. As such they are better suited to experienced practitioners wishing to fine‐tune their skill rather than the average or novice operator trying to make sense of performing complex obstetric Doppler examinations. For instance, the ISUOG recommendation: ‘the frequency for the colour Doppler mode should be adjusted to optimise the signals’1 is vague and does not provide specific guidance for the operator on what to do. An operator who does not know what the appropriate adjustment is will not learn it. An operator who knows what the appropriate adjustment is does not need this guidance. In any case, adjustment to colour Doppler frequency is almost never required because ultrasound systems automatically utilise a lower colour Doppler transmit frequency to improve the signal‐to‐noise ratio. There are other technical inaccuracies, for example, the suggestion that: ‘colour Doppler resolution increases when the colour box is reduced in size’ is simplistic. Only temporal resolution (frame rate) increases as colour box size is reduced. Axial resolution does not change and neither does the lateral or elevational resolution on most systems. To increase the spatial resolution of colour Doppler, if ever required, the operator should: (i) increase the colour Doppler frequency, (ii) increase the colour line density control and (iii) and utilise high definition zoom. Even a simple reduction of gain can improve the spatial resolution of colour Doppler, albeit at the expense of sensitivity.
While the existing guidelines are important and generally cover the subject well for the experienced operator, the guidelines may not be the best learning tool for the novice. It is best that novice operators train under the guidance of an expert and frequently refer to published guidelines to refine their technique.
Technical acquisition errors
The need for technically well acquired and optimised images has been highlighted by many authors1, 2, 3, 4 and cannot be overemphasised. The general principles of good Doppler acquisition are listed in Table1 and these apply to almost all obstetric Doppler applications. The table does not include all the conceivable optimisation steps one could take, but rather a practical set of the most relevant steps. From the pragmatic standpoint, successful spectral Doppler assessment involves a well optimised B‐mode, optimised and enlarged colour Doppler image and targeted sampling of flow within the main stream of the vessel at favourable Doppler angle, preferably 0°. The result should be a large, clear, crisp and consistent run of waveforms where automated or manual measurement can be made without any ambiguity (Figure1).
Table 1.
Steps for the acquisition of technically optimal Doppler recordings
| B‐mode optimisation |
|
| Colour Doppler optimisation |
|
| Spectral Doppler optimisation and sampling |
|
Open in a new tab
Figure 1.
Open in a new tab
Errors related to physiologic factors
Aside from the feto‐placental and utero‐placental abnormalities that drive fetal cardiovascular adaptations in IUGR and other conditions, there is a range of physiologic factors that can transiently alter the Doppler parameters during the examination. These include fetal breathing, fetal movement, maternal breathing, fetal heart rate as well as transducer pressure in the case of MCA.5, 6 For this reason, all Doppler assessments should be performed during fetal quiescence, with light transducer pressure.
Fetal breathing can lead to marked phasic change in the Doppler waveforms (Figure2). Measurements during breathing are generally discouraged. Rather than making an educated guess as to where to measure or averaging multiple waveforms, it is far better to wait until baby stops breathing.
Figure 2.
Open in a new tab
Fetal movement is problematic on two counts. First, fetal movement can displace the insonated vessel leading to poor Doppler trace or complete loss of Doppler trace. The same problem can be experienced with regard to maternal breathing occasionally requiring the mother to hold her breath during Doppler sampling. Secondly, fetal movement and increased levels of fetal activity in general can increase the fetal heart rate.
Fetal heart rate has a major effect on ratios mainly because it affects the length of diastole. The lower the heart rate, thelonger the diastole, the lower the end‐diastolic flow and the higher the measured waveform ratio. Conversely, the higher the heart rate, the shorter the diastole, the higher the end‐diastolic flow and the lower the measured waveform ratio. With regard to the umbilical artery pulsatility index (PI), a higher heart rate during fetal activity can reduce (apparently improve) the umbilical artery PI while reducing (apparently worsening) the middle cerebral artery PI. Figure3 demonstrates a SGA fetus whose MCA PI is artificially reduced during fetal activity well into the abnormal range (Figure3a) compared to baseline (Figure3b). The difference in this case is clinically significant as one measurement lies within the normal range and the other outside the normal range (Figure3c). In extreme scenarios where fetal heart rates are well beyond normal physiologic limits such as in fetuses with tachyarrhythmias or bradyarrhythmias, the UA Doppler may be completely uninterpretable. Figure4 demonstrates a very abnormal UA Doppler waveform in a fetus with Anti‐Ro antibodies and complete heart block. The fetus is otherwise well and is not growth restricted. The apparent abnormality of the umbilical artery waveform is due to the prolongation of the cardiac cycle, not due to abnormal placento‐uterine resistance.
Figure 3.
Open in a new tab
Figure 4.
Open in a new tab
Incorrect application of measurements
Background to systolic:diastolic, resistive index and PI
In the earlier days of obstetric Doppler ultrasound, measurements of waveform ratios included the systolic:diastolic (SD) ratio and the resistive index (RI). The systolic:diastolic ratio is defined as SD=PSV/EDV where PSV is the peak systolic velocity and EDV is the end‐diastolic velocity. The resistive index is defined as RI=(PSV−EDV)/PSV. Both SD and RI are easy to measure. All the operator needs to do is to position one cursor on peak systole and one on end diastole. The SD and RI, however, have some limitations. The primary limitation is that neither ratio accounts for what happens in the long period of time between systole and diastole. It is therefore possible for two different waveforms that represent two different flow scenarios to share the same values of SD and RI (Figure5).
Figure 5.
Open in a new tab
The SD ratio additionally suffers from another major annoyance. In growth restricted fetuses with reducing end‐diastolic velocity, the SD ratio rapidly becomes very large, so much so that the measurement shoots off the reference chart and sometimes well beyond the margins of the page. If this was not bad enough, the situation becomes even worse when the UA end‐diastolic velocity hits rock bottom (zero). At this point, the SD ratio instantly becomes invalidated, undefined and incalculable because division by zero is mathematical nonsense. The RI overcomes some of these limitations, but it still does not address the large information void between peak systole and end diastole. And this is precisely where the PI has come in and why the PI has become the preferred waveform ratio in obstetrics and other ultrasound subspecialties. The PI is defined as a PI=(PSV−EDV)/MeanV. Because of the mean velocity parameter in the denominator, the PI accounts for the entire cardiac cycle and is sensitive to small changes in waveform shape. While the SD ratio and RI are the same for the waveforms in Figure5, the PI is different (4.5 on the left and 2.3 on the right). The PI is therefore able to successfully differentiate between these two waveforms.
The source of potential problems during PI measurements
The measurement of the PI requires the estimation of mean flow velocity over one cardiac cycle or over several complete cardiac cycles. There are generally three methods to achieve this: (i) automatic trace done by the machine (henceforth referred to as ‘autotrace’), (ii) manual ‘trace by points’ method or (iii) manual continuous trace method. As a matter of simplicity and efficiency, most operators choose the first option, letting the machine determine not only the waveform envelope, but also the peak systolic velocity, end‐diastolic velocity and the boundaries of the cardiac cycle. It's a big ask for a piece of computer software. Unsurprisingly, the software sometimes fails. To make matters worse, the software usually performs well in normal fetuses, but becomes unstable when waveforms become abnormal in fetuses in trouble even if the waveforms are reasonably well optimised.
Even apparently small errors in the autotrace measurement can lead to relatively large errors in the PI estimation as shown in Figure6. Further examples of inaccurate automatic measurements including incorrect (jagged) tracings, failure to detect flow reversal and incorrect detection of the cardiac cycle boundaries are shown in Figure7. On some machines, the autotrace is also relatively faint and the misplacement of the measurement trace may be difficult to appreciate (Figure8).
Figure 6.
Open in a new tab
Figure 7.
Open in a new tab
Figure 8.
Open in a new tab
In summary, when using autotrace, it is essential that the operator carefully checks that the waveform and the cardiac cycle were detected correctly.
How about Doppler safety?
Ultrasound imaging constitutes a transfer of energy into the fetus and can increase the temperature of the insonated tissues. This is especially true in spectral Doppler mode where higher intensities are used. Regardless of the good safety track‐record of ultrasound, all operators should abide by the principles of prudent use and ALARA.7, 8 The ISUOG Doppler guideline1 recommends that ‘the displayed thermal index (TI) should be ≤1.0’. This is a sound recommendation, however, the problem is that multiple international studies9, 10, 11 have demonstrated that operators of ultrasound have little functional understanding of what the thermal indices are, what they indicate, which ones should be used when and how to keep the thermal index under control should it suddenly spike during spectral Doppler sampling. The NZMFMN guideline2 is a little more specific and prompts the operator to take specific action: ‘always keep TIb<0.5 if possible or at least <1 by reducing the acoustic output power’ and for umbilical artery sampling: ‘avoid sampling in such a way that the Doppler beam is directed towards fetal eyes’ (Figure9).
Figure 9.
Open in a new tab
As a matter of interest, in November 2015, I ran a Google image search for ‘MCA PI fetus’ and reviewed images that have come up on the first page. There were 9 MCA Doppler images that contained all the technical parameters including the output display standard (ODS) indicators (TI and MI). Of the nine images, four images displayed the wrong ODS index (TIs instead of TIb) and in the remaining five images, the correct index was displayed but the exposure was too high (>1). Not a single image was obtained correctly from the biological safety perspective and not a single image complied with ISUOG recommendations. This is a worrying finding. It shows that despite international best‐practice guidelines in place, operators of ultrasound generally pay little attention to biological safety of the fetus. The complacency of the experts also affects the next generation of sonography trainees.9 Having reviewed over 50 case‐based assessments from ASUM DMU trainees this year, I was disappointed to find how many senior ultrasound trainees neglected the basic safety considerations in obstetric Doppler imaging. Some trainees were scanning on machines not even set up to display the appropriate thermal index, for instance displaying the TIs where a TIb was required. Virtually all of the trainees dutifully parroted back some trivial aspects of ALARA and prudent use in their writing, but then completely disregarded these fundamental principles in their clinical practice.
Sonographers and sonologists should not automatically assume that the manufacturer has set up the ultrasound machine defaults and presents correctly.9 I was shocked to come across a promotional image of a certain ultrasound machine manufacturer (I will refrain from vilifying this manufacturer by naming them here) containing a fetal Doppler image obtained with a TIb of 4.4. Translating what this parameter means, in this case the acoustic power output of the transducer was 4.4 times greater than the power output that could increase the temperature in fetal bone near the focal point by 1°C. Four point four times greater! There are absolutely no circumstances in obstetric imaging when such high levels of acoustic output power would be necessary or even remotely justified.
Operators of ultrasound need to keep in mind that it is our responsibility to control the biological safety of the fetus. This includes performing Doppler ultrasound only when clinically justified (not as a matter of routine),12 keeping exposure times reasonable and reducing the acoustic output if the thermal index is high. Finally, it is our duty to instil these safety principles firmly into the minds of the next generation of sonographers and sonologists.
Conclusion
We find ourselves truly in the age of obstetric Doppler. By definition, 10% of the fetal population will be diagnosed as SGA and these fetuses will receive multiple Doppler examinations. There are many other fetuses that will also require Doppler tests because of maternal hypertensive disorders, decreased fetal movements, fetal abnormality and other maternal or fetal factors. Doppler is an indispensable and reliable assessment tool. Doppler saves lives. It is relatively easy to perform and interpret, but it is not without its nuances and potential problems. Sonographers and sonologists need to be aware of these pitfalls and use their technical expertise to overcome them. By adhering to the best‐practice guidelines, we can ensure that the Doppler tests we perform are accurate, clinically relevant and also safe.
References
- 1.ISUOG Clinical Standards Committee. ISUOG Practice Guidelines: use of Doppler ultrasonography in obstetrics. Ultrasound Obstet Gynecol2013; 41: 233–9. [DOI] [PubMed] [Google Scholar]
- 2.NZMFMN. New Zealand Obstetric Doppler Guideline. New Zealand Maternal Fetal Medicine Network 2015. Available at: http://www.healthpoint.co.nz/public/new-zealand-maternal-fetal-medicine-network/?solo=otherList&index=4 [accessed 12 November 2015].
- 3.Mari G. Middle cerebral artery peak systolic velocity for the diagnosis of fetal anemia: the untold story. Ultrasound Obstet Gynecol 2005; 4: 323–30. [DOI] [PubMed] [Google Scholar]
- 4.Nicolaides K, Rizzo G, Hecher K, Ximenes R. Doppler in obstetrics. The Fetal Medicine Foundation 2002. Available at: https://fetalmedicine.org/var/uploads/Doppler-in-Obstetrics.pdf [accessed 12 November 2015].
- 5.Vyas S, Campbell S, Bower S, Nicolaides K. Maternal abdominal pressure alters fetal cerebral blood flow. Br J Obstet Gynaecol 1990; 97: 740–2. [DOI] [PubMed] [Google Scholar]
- 6.Su YM, Lv GR, Chen XK, Li SH, Lin HT. Ultrasound probe pressure but not maternal Valsalva maneuver alters Doppler parameters during fetal middle cerebral artery Doppler ultrasonography. Prenat Diagn 2010; 30(12–13): 1192–7. [DOI] [PubMed] [Google Scholar]
- 7.Bioeffects Committee of the American Institute of Ultrasound in Medicine. American Institute of Ultrasound in Medicine Consensus Report on potential bioeffects of diagnostic ultrasound, executive summary. J Ultrasound Med2008; 27: 503–15. [DOI] [PubMed] [Google Scholar]
- 8.Nelson TR, Fawlkes JB, Abramowicz JS, Church C. Ultrasound biosafety considerations for the practicing sonographer and sonologist. J Ultrasound Med 2009; 28: 139–50. [DOI] [PubMed] [Google Scholar]
- 9.Necas M. New Zealand sonographers do not outperform their European or American colleagues in the knowledge of ultrasound safety. AJUM 2010; 13(4): 28–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sheiner E, Shoham‐Vardi I, Abramowicz JS. What do clinical usersknow regarding safety of ultrasound during pregnancy? JUltrasound Med 2007; 26: 319–25. [DOI] [PubMed] [Google Scholar]
- 11.Sheiner E, Abramowicz JS. Clinical end users worldwide show poor knowledge regarding safety issues of ultrasound during pregnancy. J Ultrasound Med 2008; 27: 499–501. [DOI] [PubMed] [Google Scholar]
- 12.Alfirevic Z, Stampalija T, Medley N.Fetal and umbilical Doppler ultrasound in normal pregnancy. Cochrane Database Syst Rev 2015; 4: CD001450. doi: 10.1002/14651858.CD001450.pub4. [DOI] [PMC free article] [PubMed] [Google Scholar]
