Guidelines for the Safe Use of Diagnostic Ultrasound - Rationale (Continued)

4. Rationale (Continued)

4.2 Mechanical Effects

In the absence of heating, biological effects at diagnostic exposure levels have been observed in mammalian tissues with stable gas bodies, such as lung (WFUMB 1998, AIUM 2000), and intestine. Such effects have also been observed in other tissues after injection of ultrasound gas contrast agents (microbubbles). Therefore, guidelines have been developed in terms of an exposure parameter directly related to mechanical (non-thermal) effects. At or below the recommended output limits of Section 2.4 (MI = 1.9), mechanical effects are far less likely to be important in obstetrical ultrasound because of the absence of gas bodies.

4.2.1 Exposure Parameter

Implementation of the recommendations in this document requires a basic knowledge of the meaning of the new primary exposure parameter for mechanical effects, the Mechanical Index (MI) (AIUM/NEMA 1998a, Lopez 1998). The development of the Mechanical Index has been described in detail elsewhere (AIUM/NEMA 1998a, AIUM 1993, Apfel and Holland 1991). It is approximately the largest rarefaction pressure (in MPa) in a soft-tissue attenuated ultrasound beam, divided by the square root of the centre frequency (in MHz) of the ultrasound pulse.

The MI is related to the potential for hemorrhaging of the pulmonary alveolar capillaries due to ultrasonic exposure of the lung during a diagnostic ultrasound examination. The threshold for lung hemorrhage depends on the ultrasonic pressure at the surface of the patient's lung divided by the square root of the centre frequency of the ultrasonic pulse (AIUM 1993). Although this quantity and the MI may not always be in direct proportion, in many cases, the MI provides a relative indication of the potential for lung hemorrhage due to ultrasound exposure. It can also provide an approximate relative indication of the potential for biological effects in the presence of contrast agents (AIUM 2000).

4.2.2 Biological Effects

Hemorrhaging of lung capillaries is the first and most thoroughly studied mechanical biological effect which has been observed in mammals at diagnostically relevant exposures. Beginning with the study by Child, et al., (1990), this effect has also been observed in several laboratories and in several mammalian species when the lung was directly exposed by pulsed ultrasound (Dalecki, et al., 1997, Baggs, et al, . 1996, Holland, et al., 1996, Zachary and O'Brien 1995, Tarantal and Canfield 1994, Frizzell, et al., 1994). The acoustic pressures, centre frequencies, pulse durations and pulse repetition rates were at diagnostically relevant values in the studies. The species included swine, rat, rabbit, monkey and mouse. The species with lungs most similar to humans were monkey and swine. Studies prior to 1993 have been summarized elsewhere (AIUM 1993). Some of the studies published after 1993 are summarized below in chronological order. More detailed reviews have also either been recently published or are in preparation (WFUMB 1998, AIUM 2000, NCRP in preparation).

Tarantal and Canfield (1994) reported findings of multiple, circular hemorrhagic foci of 1 - 10 mm diameter in the lungs of monkeys directly exposed by ultrasound. The lesions were usually near the pleural surface and appeared to originate from alveolar capillaries. A clinical scanner operating in "triple mode"(B-mode imaging + colour Doppler + pulsed Doppler)at maximum output was used to provide the exposure. The monkeys ranged in age from 3 months to 5 years (infant to young adult). The frequency was 4 MHz, with a pulse duration of 0.65 microseconds and a PRF of 1515 Hz. The rarefaction pressure was maximal at 1.2 cm depth where the MI was 1.8. The exposure duration was 5 minutes. The chest wall thickness in the monkeys was 0.3 to 1.2 cm and the transducer was held directly to the chest.

O'Brien and Zachary (1997) found evidence for lung hemorrhaging in adult rabbits and mice after exposure to pulsed wave ultrasound for 5 minutes with a commercial diagnostic ultrasound imaging system operating at 3 and 6 MHz, with MI values between 0.8 and 2.2. All exposures were above threshold. However, in adult pigs(10-12 weeks old, weighing about 30 kg), no hemorrhaging was observed.

Dalecki and co-workers (Dalecki, et al., 1997, Baggs, et al., 1996) reported lung hemorrhaging in neonatal and 10 day old swine after exposure by a stationary 2.3 MHz ultrasound beam with a pulse length of 10 microseconds and a 100 Hz pulse repetition frequency. The exposure duration at a single location was between 10 s and 2 minutes. In water, at the surface of the animal, the maximum negative pressure at threshold was about 1.1-1.4 Mpa. The threshold pressure at the surface of the lung was reported as 0.7-1.0 MPa. This quantity divided by the square root of the centre frequency has a value between 0.5 and 0.7. Exposures at a pressure 1.5x the threshold value showed 1 - 1.5 mm focal hemorrhages. At a factor of 2 - 3 above threshold pressure, clearly defined hemorrhagic areas were observed with linear dimensions up to 6 mm. Damage was restricted to single lobules and all hemorrhages were subcapsular, with no rupture of the parietal pleura.

Dalecki, et al., (1997) summarized the results of studies about the way in which the pressure threshold for lung hemorrhage depended on other parameters of the ultrasound exposure. These parameters included centre frequency, pulse duration and exposure duration (Holland, et al., 1996, Child, et al., 1990). The studies indicated that the pressure threshold increases as the square root of the centre frequency and that reducing the pulse duration by a factor of 10 increases the pressure threshold by a factor of 2. The threshold of lung hemorrhaging was found to be a weak function of exposure duration.

Meltzer, et al.,(1998) found no evidence of hemorrhaging after intraoperative transesophageal echocardiography in adults where the MI was 1.3.

There was also evidence of hemorrhaging in murine intestine (Dalecki, et al., 1995) after 5 minutes of exposure to ultrasound exposure at diagnostically relevant frequencies, pulse durations (10 µs) and pulse repetition frequencies (100 Hz). The effect occurred in the absence of significant heating. At 2.4 and 3.6 MHz, the threshold pressure divided by the square root of the frequency was approximately 1.9.

A preliminary report has also been published by Skyba, et al, . (1998) of microbubble destruction of rat muscle capillaries in vivo using 2.3 MHz ultrasound at reported MI values ranging from 0.4 to 1. In another study, Miller and Gies (1998) injected mice with gas contrast agents and demonstrated significant enhancements in the generation of petechiae in the mouse intestine due to pulsed ultrasound exposure at 1 MHz with a 10 microsecond pulse duration and exposure levels as low as 1 MPa.

Recently, a study has been published (Dalecki, et al., 1999) that describes hemorrhaging in murine fetuses exposed to pulsed ultrasound with 10 microsecond pulses delivered with a pulse repetition frequency of 100 Hz. In this case, the hemorrhaging appeared near developing bone. Gas bodies did not appear to be a relevant factor. At 1.2 MHz, the negative pressure threshold for hemorrhage to the fetal head was about 2.5 MPa. No statistically significant hemorrhage was found at frequencies of 2.4 and 3.6 MHz at the highest negative pressure of 5 MPa.

4.2.3. Human Exposure and Clinical Significance

To determine whether an exposure is justified, the equipment operator must assess whether reliable diagnostic information can be achieved as well as the severity and likelihood of an adverse health effect.

Carstensen and co-workers (Carstensen, et al., 1992, Baggs, et al, . 1996) discussed the likelihood of finding clinical ultrasound exposures above the thresholds for capillary lung hemorrhage found in the experimental studies noted in Section 4.2.2. Their conclusion was that the largest outputs available from equipment in 1992 were very near the thresholds of macroscopic hemorrhage of lung tissue, if exposure was directly over the lung in an echocardiographic exam. This was considered the most common way in which the lung would be exposed.

It is plausible that more recent devices have higher output levels and may exceed the thresholds for lung hemorrhaging for the type of exposure described above. The output levels for the devices considered by Carstensen and co-workers are expected to have been similar to those in the survey of Patton, et al., (1994). In that study, it was found that 14 of 266 samples yielded MI values greater than 1, the largest being 1.3. However, current equipment has MI values as high as 1.9. This is the current limit for diagnostic ultrasound devices in the U.S. FDA 510(k) guidance document. Therefore, commercial devices with these values of MI appear to be able to generate suprathreshold levels.

The potential for lung hemorrhaging was found to be greater for less typical diagnostic procedures, where the focus of the sound beam could strike the surface of lung tissue. Baggs, et al., (1996) stated that this could occur if a standoff is used and the lung tissue is near the surface of the chest. They noted that the focus of the beam could also strike the surface of lung tissue at the far side of the heart, particularly in pediatric or transesophageal applications. In these circumstances, it is estimated that devices with MI values of 1.9 could lead to exposures above the thresholds for lung hemorrhaging; by about factors of 1.5 and 3 for 1 microsecond and 10 microsecond pulse lengths, respectively. The 1 microsecond pulse length is typical for Doppler and B-mode imaging exams, although pulse durations up to 10 microseconds are available in some pulsed Doppler modes. This indicates that, for such exposures, there is a risk of some pulmonary alveolar hemorrhaging of the capillaries for neonatal or infant examinations.

The evidence concerning the biological effects studies and the acoustic outputs of diagnostic ultrasound devices, suggests that a cautious but reasonable approach is to assume that, if MI exceeds 1, there is some risk of capillary hemorrhaging on the lung surface in diagnostic ultrasound examinations of neonates and infants in which the lung is exposed. However, the long term implications of the injury may not be serious (Baggs, et al., 1996). For example, Tarantal and Canfield (1994) described the hemorrhage observed in their study as anatomically mild. In both monkey and swine studies, the pathology did not indicate any disruption of the alveolar architecture. Therefore, clinically, recovery of the lesions would be expected upon the resorption of blood and reconstitution of existing architecture. No symptoms suggestive of any respiratory distress would be expected because of the lung's ability to compensate and the focal nature of the hemorrhages. Although there are no published results on whether the outcome would differ in the presence of other disease, hemorrhaging could be more extensive in the presence of coagulation disorders.

It is unlikely that there would be any significant intestinal hemorrhaging, even at the highest MI values available. However, pathologic conditions inhibiting intestinal peristalsis or promoting submucosal gas collections may increase the likelihood of such an effect.

Although studies are preliminary and the clinical significance has not been demonstrated, it is important that equipment operators be aware of the increased potential for capillary damage during a diagnostic ultrasound examination when GCAs are used, particularly in applications where malignant tumours may be exposed.

A maximum attainable value of 1.9 for the MI greatly reduces the potential for clinically significant damage from mechanical effects during diagnostic ultrasound examinations. If thresholds for biological effects must be exceeded to obtain useful diagnostic information, an examination technique that uses the real-time display of the Mechanical Index should make it relatively easy to ensure that the risk to the patient from mechanical effects is clinically justified. Above biological effects thresholds, the ALARA principle can be implemented by lowering the MI and/or reducing the dwell time to help minimize the severity of any potential injury, if required.