| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||
Guest editorial |
Medical Physics Department, Royal United Hospital NHS Trust, Bath BA1 3NG, UK
The measurement of radiation is central to the safe and effective management of radiological imaging. Diagnostic ultrasound now constitutes a substantial percentage of imaging practice. At present, however, there is very little equivalent investment in the measurement of exposure to ultrasound. This editorial explores how this has come about and what may be done to respond to the present situation.
The data for use of ultrasound in diagnostic imaging speak for themselves. During the year 2003/2004, 5 937 383 ultrasound scans were reported to the Department of Health as being carried out in NHS Trusts in England [1]. This is 19.4% of all reported imaging tests. The total was about 300 000 scans more than the previous year, slightly greater than the combined growth reported for CT and MRI. In addition, much ultrasound activity probably remains unreported, being carried out in hospital departments other than radiology or within the expanding sector of primary care. Some years ago it was estimated that 25% of all imaging world-wide was due to ultrasound, and one may speculate that this is still an underestimate of the true activity.
Exact estimates for the numbers of scanners installed are less easy to determine. In my own Trust and local health community there are approximately 40 ultrasound scanners in current use. 27 of these have been installed since January 2000. This may be compared with a current complement of approximately 50 X-ray rooms, one CT scanner and one MR scanner. So, broadly, there are similar numbers of X-ray installations as ultrasound scanners. However, the comparison alters considerably when comparing ultrasound transducers with X-ray tubes, where clearly the former considerably outnumber the latter. This Trust serves a population of about 400 000 and these figures may well mirror the experience of similar acute trusts over the rest of the country.
One cause of the extensive use and continuing growth in diagnostic ultrasound has been its enviable reputation as a safe modality for imaging. About 35% of reported scans are still carried out in obstetrics and gynaecology, where safety of the unborn child takes absolute priority. Although now only one of many uses, obstetrics established ultrasound as a preferred safe modality, substantially replacing the use of X-ray imaging with its known hazard. Nevertheless, ultrasound examinations do expose the patient to external radiation, albeit associated with different hazards than those from ionizing radiation. As a result, the science of ultrasound exposure measurement has grown in partnership with its clinical use, creating knowledge of ultrasound fields and exposure on which to base estimates of hazard and to limit risk.
Rules can be great generators of activity. The Ionising Radiation (Medical Exposure) Regulations (IR(ME)R) have resulted in many more measurements being performed to quantify patient exposure to diagnostic radiological procedures. The Regulations and the resultant dose-related measurements have helped to maintain and reinforce the balance between clinically effective use of ionizing radiation, and the risks associated with its use. For ultrasound imaging, the regulation which has had the greatest impact on exposure differs from IR(ME)R in several respects. It operates, not in the UK but in the USA, through its Food and Drug Administration (FDA) [2]. It sets a control on the design of ultrasound scanners, placing limits on the allowed ultrasound output for medical use. And its emphasis has resulted in a substantial investment in acoustic measurement by manufacturers, so that they can declare to the FDA the values of specific acoustic quantities and place their scanners "on the market" in the USA. The most important overall limit, established over 20 years ago, is still applied: estimated in situ spatial peak temporal-average intensity must not exceed 720 mW cm–2. These FDA regulations strictly only have force in the USA; however, several other countries have also adopted these FDA regulations. Thus a US law has served to constrain ultrasound exposure in an international market. Acoustic measurement became embedded in the industrial scanner design process.
At the same time and on a much smaller scale, a few independent laboratories equipped themselves to carry out measurements of acoustic output. Calibration facilities such as those at the UK National Physical Laboratory, led by Dr Roy Preston, established robust calibration services, and themselves made measurements on commercial scanners. Ultrasound laboratories were established by NHS medical physics departments such as that in Newcastle under Dr Tony Whittingham and our own in Bath. Evidence emerged of occasional errors by manufacturers: these were largely corrected. But of greater general importance was the evidence from long-term studies of a general trend towards increased acoustic output. Clinical users were asking for improved performance and manufacturers competed to gain commercial edge; Doppler techniques evolved and limits to acoustic imaging were pushed. Independent surveys demonstrated an increasing availability of equipment capable of approaching the exposure limits set by the FDA. Nevertheless, such independent acoustic output measurements remain limited to very few active departments. The cost to set up an effective ultrasound measurement system, including appropriate hydrophones and acquisition and computing systems, often proved too great a challenge to a cash-strapped service when judged against more robust arguments for equipment for ionizing radiation measurement. Furthermore, even the basic measurement of total acoustic power required sensitive radiation force balances, which at that time were available only to those laboratories capable of designing and building them in-house.
Life moved on. Earlier FDA limits prevented the use of Doppler systems in the USA, leaving pioneers in Europe to demonstrate the value of Doppler in obstetric imaging. The US reaction was to remove the specific intensity limit placed on obstetric exposure, allowing manufacturers to make available, for obstetric applications, intensities which had previously been reserved for peripheral vascular applications only. Agreement to relax the regulation occurred, however, only with an additional agreement to introduce so-called "safety indices": the Thermal Index (TI) and the Mechanical Index (MI) [3–5]. These safety indices are now displayed on most new scanners, and are intended to guide users in the safe use of the scanner. For those involved in ultrasound exposure, however, this represented a considerable escalation in the challenge of acoustic measurement. Acoustic power must be measured to calculate TI, with special arrangements to measure "power per centimetre" from imaging arrays. MI requires measurement of hydrophone position and acoustic frequency, as well as peak rarefaction pressure, where acoustic pressure alone previously had been sufficient. Much more significantly, however, the measurement strategy changed from an investigation of conditions giving rise to maximum exposure, to a measurement regimen applied to all conditions for which a safety index value might be displayed. Users were now expected to note and act upon the displayed value whereas, before, the user had almost no information on which to control output for safety reasons. This resulted, for example, in general caution against the use of Doppler in early pregnancy. The introduction of a safety index display has, at least in principle, enabled users to determine Doppler conditions with low TI and to use these conditions preferentially.
Comparison between exposure measurements for diagnostic radiology and for diagnostic ultrasound is, perhaps, illuminating. Output of X-ray equipment is evaluated using an ionization chamber to measure exposure and a tube potential divider to estimate tube potential. More recent emphasis on patient exposure and dose has resulted in the widespread installation and use of dose–area product (DAP) meters, and the introduction of surveys with thermoluminescent dosemeters to establish and use diagnostic reference levels. Risk assessment depends on well-established Monte Carlo modelling and comparison with an established body of scientific bio-effects evidence. By comparison, ultrasound measurement uses a hydrophone to measure the acoustic pressure wave and a radiation force balance for total acoustic power. In situ estimation of acoustic exposure may be modelled, and temperature rise can be estimated. Whilst detailed differences exist, there is close equivalence in principle between ultrasound and X-rays. Both have established methods for measuring key parameters of the radiation. Both have models to predict radiation exposure in situ. Both have means to convert measured exposure to dose, although this is, at present, more highly developed for ionizing radiation than for ultrasound, for which the concept of dose has still to be properly developed. Most notably, ultrasound lacks robust means to predict risk.
But the measurement challenge for ultrasound and the costs are significant. Manufacturers are now obliged to measure the acoustic conditions relating to the full range of scanner settings, and to calculate all the safety indices according to the defining formulae. They must also take account of equipment variability, most notably that resulting from the variability of transducer manufacture. For those involved with independent measurement, full evaluation of the correctness of the displayed index values is generally impossible except for very simple scanners, within any reasonable time scale. The calculation of TI in pulsed Doppler mode, for example, requires axial intensity profiles to be measured, which demands automated hydrophone scanning to be practical. Indeed, such are the financial implications for industry that feasibility studies are in place to compute rather than measure acoustic quantities, based on validated computer models.
Although a full evaluation is beyond the scope of any independent laboratory, it is possible for an experienced laboratory to carry out spot checks of equipment for selected, critical output conditions. This is happening in a small number of hospitals and laboratories and not surprisingly, evidence is now emerging of discrepancies between displayed safety index values and those measured independently. For the larger part, the differences are small, and often the measured value is lower than that displayed, so erring on the side of caution. Nevertheless, some large errors have been found and reported. Clearly, a means designed to inform users about safety must have their confidence, and there remains work to be done both by industry and independent laboratories to ensure that the indices can always be relied upon to give valid guidance. For hospitals this requires, necessarily, investment in appropriate measurement equipment and allocation of scientists' and technologists' time to the task.
The safety indices were developed to advise the user of possible thermal and mechanical hazard. For the measurement scientist, an alternative sensible approach is to avoid the limitations of calculation inherent in the TI and to measure temperature rise directly. This approach carries the added advantages for hospital-based laboratories of relative simplicity, low cost, and speedy measurement. The development of the Thermal Test Object by the National Physical Laboratory is one approach which has provided valuable results reported in two projects funded by the Department of Health [4]; this approach deserves more extensive evaluation than it has experienced so far. However, the highest temperature rises are at the transducer surface, and measurement and management of surface temperature remains a central issue for both manufacturers and users. Many clinical scanners may cause the transducer surface to heat by 20°C in air, and can cause temperatures very close to the limit allowed by the International Electrotechnical Commission (IEC), 43°C, when in contact with the skin [5]. Of particular concern is the behaviour of interstitial transducers, especially transvaginal. A recent study from the Medical Physics Department at St George's Hospital has demonstrated that some scanner conditions may cause unacceptably high surface temperatures for some transvaginal transducers [6].
Many challenges remain. High-fidelity acoustic measurements require appropriate measurement devices, suitable for laboratory-based and for field measurements. It is commonly inconvenient or impossible to take a scanner away from its clinical setting for measurement. There is a pressing need for a well-engineered, portable commercial acoustic power meter capable of measuring acoustic power down to 10 mW or less. This would make available a direct and simple means to confirm the accuracy of displayed TI for imaging, and for identifying critical conditions for transducer heating. Hydrophone technology is well developed, and the challenges of phase calibration and spatial and temporal resolution associated with higher-frequency beams are being met. Modern ultrasound imaging systems now operate with complex sequences of ultrasound pulses of a variety of waveforms, where current hydrophone-based measurement systems may sometimes be inadequate. Non-linear propagation effects must be managed properly, and there will be measurement challenges in carrying out all measurements under low-amplitude, quasi-linear conditions. Whilst hydrophone technology is well developed, the practical use still remains clumsy even now. Hydrophones in buckets worked well as the measurement need developed, but it is time to have available an integrated easy-to-use portable package including hydrophone, positioning, data acquisition and output. Such a system would enable on-line measurement of pressure waveform, acoustic frequency, MI and temporal-average intensity, all key quantities for exposure measurement. Partnership between hospitals, standards laboratories, academics and industry is essential to progress. The UK is particularly fortunate in having local scientific and technical skills of both the National Physical Laboratory (Teddington, UK) and Precision Acoustics (Dorchester, UK) to support this area.
The availability of measurement equipment is only part of the story. Ultrasound measurement must become part of the scientific ethos of all NHS Trusts, mediated through their medical physics and bioengineering services. Commitment to a regimen of measurement support should be integrated within the risk management policies of Trusts, so ensuring accountability and the ability to respond robustly to the question "What evidence to you have that your ultrasound scanners are safe to use on patients?" Training in ultrasound measurement is already integrated within the IPEM training scheme for Medical Physicists and Clinical Engineers, and skill development for these scientists will become part of their continued professional development, supported by the national occupational standards for Healthcare Scientists. As a result, NHS Trusts will have available local skills to support and advise on ultrasound output and safety of ultrasound scanning in all disciplines.
I have heard it suggested that ultrasound measurements should lie with a central Government-funded laboratory, and that manufacturers should submit all new equipment to such a laboratory for measurement, prior to offering it for sale in the UK. This suggestion has some merit and the establishment of this type of specialist laboratory would provide a valuable resource, making available specialist skills, measurement equipment and ultrasound beam data of importance to all users of ultrasound. It seems unlikely, however, that users would tolerate significant delays between the introduction of a new scanner, transducer or mode, and the output being confirmed as appropriate for use. The task itself is substantial, with new systems, modes, and transducers being introduced monthly, and beam complexity challenging the current measurement technology. With the advent of fully computer-managed scanners, software upgrades are often distributed direct to users. Furthermore, the precedents are not good for the longevity of such a laboratory, given the recent decision to close the Medical Devices Agency "Ultrasound Equipment Evaluation Project". Such centrally-funded schemes always ultimately close because funding is linked to political expediency. Consequently, the clear priority is to establish a desire at a local level to engage with this task, and by so doing to embed scientific understanding at the point of service delivery.
The measurement of radiation is central to the tradition of radiology. The success of diagnostic ultrasound has resulted in the introduction into healthcare of very large numbers of ultrasound scanners, now approaching in number that of X-ray installations. Current hospital practice makes little or no attempt to evaluate the ultrasound radiation generated by this equipment and there is little investment in measurement tools, personnel or commitment of time to this role. Given the capital investment in ultrasound scanners, perhaps approaching £2 million per Trust, it is difficult indeed to find an argument against the purchase of a hydrophone and power balance for a few thousand pounds, or the investment in skills to understand their use and interpret the findings. Ultrasound exposure measurement still remains a sparsely-practiced science. It would be greatly to the benefit of the health service, for the users of ultrasound scanners and for the patients who are exposed to ultrasound waves, to have a visible and recognised ultrasound measurement service as an integral part of the radiation protection service within each hospital.
Received for publication January 11, 2005. Accepted for publication February 11, 2005.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| BJR | DMFR | IMAGING | ALL BIR JOURNALS |
