Stanley B Barnett

International recommendations and guidelines for the safe use of diagnostic ultrasound in medicine

Modern sophisticated ultrasonographic equipment is capable of delivering substantial levels of acoustic energy into the body when used at maximum outputs. The risk of producing bioeffects has been studied by international expert groups during symposia supported by the World Federation for Ultrasound in Medicine and Biology (WFUMB). These have resulted in the publication of internationally accepted conclusions and recommendations. National ultrasound safety committees have published guidelines as well. These recommendations and safety guidelines offer valuable information to help users apply diagnostic ultrasound in a safe and effective manner. Acoustic output from ultrasound medical devices is directly regulated only in the USA and this is done by the Food and Drug Administration (FDA). However, there is also a modern trend towards self-regulation which has implications for the worldwide use of diagnostic ultrasound. It has resulted in a move away from the relatively simple scheme of FDA-enforced, application-specific limits on acoustic output to a scheme whereby risk of adverse effects of ultrasound exposure is assessed from information provided by the equipment in the form of a real-time display of safety indices. Under this option, the FDA allows a relaxation of some intensity limits, specifically approving the use of medical ultrasound devices that can expose the fetus or embryo to nearly eight times the intensity that was previously allowed. The shift of responsibility for risk assessment from a regulatory authority to the user creates an urgent need for awareness of risk and the development of knowledgeable and responsible attitudes to safety issues. To encourage this approach, it is incumbent on authorities, ultrasound societies and expert groups to provide relevant information on biological effects that might result from ultrasonographic procedures. It is obvious from the continued stream of enquiries received by ultrasound societies that effective dissemination of such knowledge requires sustained strenuous effort on the part of ultrasound safety committees. There is a strong need for continuing education to ensure that appropriate risk/benefit assessments are made by users based on an appropriate knowledge of the probability of biological effects occurring with each type of ultrasound procedure. The primary purpose of this paper is to draw attention to current safety guidelines and show the similarities and areas of general agreement with those issued by the parent ultrasound organisation, the WFUMB. It is equally important to identify gaps in our knowledge, where applicable.

The sensitivity of biological tissue to ultrasound

Mammalian tissues have differing sensitivities to damage by physical agents such as ultrasound. This article evaluates the scientific data in terms of known physical mechanisms of interaction and the impact on pre- and postnatal tissues. Actively dividing cells of the embryonic and fetal central nervous system are most readily disturbed. As a diagnostic ultrasound beam envelopes a small volume of tissue, it is possible that the effects of mild disturbance may not be detected unless major neural pathways are involved. There is evidence that ultrasound can be detected by the central nervous system; however, this does not necessarily imply that the bioeffect is hazardous to the fetus. Biologically significant temperature increases can occur at or near to bone in the fetus from the second trimester, if the beam is held stationary for more than 30 s in some pulsed Doppler applications. In this way, sensory organs that are encased in bone may be susceptible to heating by conduction. Reports in animals and humans of retarded growth and development following frequent exposures to diagnostic ultrasound, in the absence of significant heating, are difficult to explain from the current knowledge of ultrasound mechanisms. There is no evidence of cavitation effects occurring in the soft tissues of the fetus when exposed to diagnostic ultrasound; however, the possibility exists that such effects may be enhanced by the introduction of echo-contrast agents.

Heating of guinea-pig fetal brain during exposure to pulsed ultrasound

Ultrasound-induced temperature elevations in fresh and formalin-fixed fetal guinea-pig brains were measured during in vitro insonation, with a stationary beam in a tank containing water at 38 degrees C. The pulsing regimen used 6.25 microseconds pulses, repeated at a frequency of 4 kHz emitted from a focussed transducer operating with a centre frequency of 3.2 MHz. The greatest temperature rise in brain tissue occurred close to bone and correlated with both gestational age and progression in bone development. After a 2 min insonation with a spatial peak temporal average intensity (ISPTA) of 2.9 W/cm2, a mean temperature elevation of 5.2 degrees C was recorded in fetuses of 60 days gestation (dg). The same exposure produced an increase of 2.6 degrees C in the centre of whole brains of 60 dg fetuses when the bony cranium was removed. As most of the heating occurs within 40 s, these findings have implications for the safety of pulsed Doppler examinations where dwell-time may be an important factor.

INTRACRANIAL TEMPERATURE ELEVATION FROM DIAGNOSTIC ULTRASOUND

Tissues of the central nervous system are sensitive to damage by physical agents, such as heat and ultrasound. Exposure to pulsed spectral Doppler ultrasound can significantly heat biologic tissue because of the relatively high intensities used and the need to hold the beam stationary during examinations. This has significant implications for sensitive neural tissue such as that exposed during spectral Doppler flow studies of fetal cerebral vessels. Recent changes in the FDA regulation allow delivery of almost eight times higher intensity into the fetal brain by ultrasound devices that incorporate an approved real-time output display in their design. In this situation, ultrasound users are expected to assess the risk/benefit ratio based on their interpretation of equipment output displays (including the thermal index, TI) and an understanding of the significance of biologic effects. To assist in the assessment of potential thermally mediated bioeffects, a number of conclusions can be drawn from the published scientific literature: the amount of ultrasound-induced intracranial heating increases with gestational age and the development of fetal bone; pulsed spectral Doppler ultrasound can produce biologically significant heating in the fetal brain; the rate of heating near bone is rapid, with approximately 75% of the maximum heating occurring within 30 s; blood flow has minimal cooling effect on ultrasound-induced heating of the brain when insonated with narrow focused clinical beams; the threshold for irreversible damage in the developing embryo and fetal brain is exceeded when a temperature increase of 4 degrees C is maintained for 5 min; an ultrasound exposure that produces a temperature increase of up to 1.5 degrees C in 120 s does not elicit measurable electrophysiologic responses in fetal brain; for some exposure conditions, the thermal index (TI), as used in the FDA-approved output display standard, underestimates the extent of ultrasound-induced intracranial temperature increase.

In vivo heating of the guinea-pig fetal brain by pulsed ultrasound and estimates of thermal index.

Temperature was measured in the brain in live near-term fetal guinea pigs (62-66 d gestational age), during in utero exposure to a fixed beam of pulsed ultrasound at intensity ISPTA 2.82 W/cm2. Mean temperature increases of 4.3 degrees C close to parietal bone and 1.1 degrees C in the mid-brain were recorded after 2-min exposures. These values were lower (12%) than those obtained for ultrasound-induced heating near the bone in dead fetuses insonated in utero. A significant cooling effect of vascular perfusion was observed only when guinea pig fetuses reached late gestation, near term, when the cerebral vessels were well developed. The estimated value for the thermal index (TIB), as used in AIUM/NEMA output display standard, underestimated the measured temperature increase at the bone-brain interface. The ratio of measured temperature to the TIB is 1.3. A modification of the cranial thermal index provided a more reasonable, conservative, estimate of the temperature increase at a biologically significant point of interest at the brain-bone interface.

Ultrasound-induced temperature increase in guinea-pig fetal brain in utero: third-trimester gestation.

Temperature increase was measured at various depths in the brain of living fetal guinea pigs during in utero exposure to unscanned pulsed ultrasound at ISPTA 2.8 W/cm2. Mean temperature increases of 4.9 degrees C close to parietal bone and 1.2 degrees C in the midbrain were recorded after 2-min exposures. When exposures were repeated on the same sites in each fetus after death, the corresponding mean temperature increases were 4.9 degrees C and 1.3 degrees C, respectively. Cerebral blood perfusion had little cooling effect on ultrasound-induced heating in the guinea pig fetus of 57-61 days gestational age.

Ultrasonic heating of the brain of the fetal sheep in utero

The fetal sheep was used as a model to determine the extent of ultrasound-induced heating of brain tissue in procedures involving pulsed Doppler examination of fetal intracranial arteries. Temperature measurements were recorded in late-gestation fetuses insonated in utero. The centre frequency was 3.5 MHz and a pulse repetition rate of 6 to 10 kHz produced a power output of 0.6 or 2 W. The brain was insonated in the near field of a focussed beam where the -6-dB beam width was 1.7 cm for the 0.6-W transducers and 1.2 cm for the 2-W transducers. Mean (standard error) maximal temperature increases of 3.0 degrees C (0.3) and 12.5 degrees C (1.3), respectively, were recorded in dead fetuses. The mean values obtained in normally perfused living fetuses were lower by 43% and 30%, respectively, showing that vascular perfusion substantially limited ultrasonic heating in sheep fetal brain tissue. There were no changes in blood flow to the heated brain tissue as measured using radiolabelled microspheres.

Routine ultrasound scanning in first trimester: What are the risks?

Acoustic exposure from modern ultrasonographic devices is capable of disturbing biological tissue to varying extent depending on the type of ultrasound examination and the particular tissue under investigation. There is no strong evidence that these biological effects present a serious health hazard, however, knowledge is incomplete, particularly from human studies. Although ultrasound induced heating can be significant in later pregnancy, it is unlikely that diagnostic ultrasound poses a significant thermal risk to the developing embryo when used according to published safety guidelines. Nevertheless, uncertainties remain, particularly for nonthermal effects in early pregnancy where shear stresses from radiation pressure may become an important factor. The likelihood of producing some biological effects can be enhanced by new procedures such as the use of gas encapsulated echo-contrast agents. The particular sensitivity of the embryo to physical damage together with uncertainties of both risk and benefit suggest that caution should be applied to the scanning of early first trimester uncomplicated pregnancy.

Ultrasound-induced heating in a foetal skull bone phantom and its dependence on beam width and perfusion

The cooling effect of single and multiple perfusing channels has been measured in a model of human foetal skull bone heated by wide and narrow beams of simulated pulsed spectral Doppler ultrasound (US). A focussed transducer operating with a centre frequency of 3.5 MHz, that emitted pulses of 5.7 micros duration with a repetition frequency of 8 kHz, was used. This produced a beam of power 100 +/- 2 mW with -6 dB diameters of 3.1 mm and 7.8 mm at 9 cm and 6 cm, respectively, from the transducer face. Arterial perfusion was simulated by allowing distilled water to flow in a large single channel or a grid of fine channels near the heated bone target. This study has established that: 1. perfusion-induced cooling is significantly enhanced when the bone phantom is heated by a wide rather than a narrow beam; 2. irrespective of the US beam width, a grid of small channels is more effective in cooling a heated bone target than a single larger diameter channel with the same volume flow rate; 3. the measured temperature rise and rate of temperature rise support the prediction of inverse proportionality to the US beam width; and 4. the perfusion time constants determined in our phantom model are 2 to 30 times larger than that assumed for the thermal index (TIB) algorithm.

Ultrasound-induced temperature increase in the guinea-pig fetal brain in vitro

The temperature of the brain of fetal guinea pigs was measured in vitro during exposure to an unscanned beam of pulsed ultrasound at intensity ISPTA 2.8 W/cm2. A mean temperature increase of 5.1 degrees C recorded after 2 min of insonation confirms results of an earlier similar study. The water-bath exposure system provided enhanced cooling of superficial tissue by acoustic streaming. When the scalp was removed, the ultrasound-induced temperature increase was substantially reduced (by 35%) due to cooling through radiation force-induced bulk fluid streaming along the direction of propagation in the water bath. The measured temperature increase in guinea pig fetal brain correlated with a modified cranial thermal index.