In Vivo ARFI Imaging of Ablation with the AcuNav™ ICE Catheter
We acquired in vivo images of a sheep heart undergoing cardiac ablation using a 64 channel AcuNav catheter. The study is described in detail in the paper in Appendix 1. We created 3 lesions in the atria two were close enough in proximity so that the first could be seen while we were creating the second. Images taken during the second ablation are shown in figure 6.
In this study, it was so difficult to align the ICE catheter to view the ablation catheter tip that we actually stopped in the middle of the procedure and glued an ablation catheter against the ICE catheter to within 4 cm of the tip. We aligned them so that when the ablation tip was deflected away, it was in the ICE imaging plane. This worked moderately well but it meant we could only image the lesion from one fixed angle.
The most difficult aspect of these studies was trying to align the ICE catheter to image the lesion location without having the metal tip in the ICE image. The metal tip causes a bright noisy echos and creates a shadow in the image. Without the catheter tip for reference, it was difficult to know if the lesion tissue was being imaged. The system we propose here will solve both of the problems just described. It will do this by showing the ICE catheter and its imaging plane on the CARTO display with the ablation catheter. This will be huge step forward and will enable ARFI technology to evaluate lesion growth in real time.
Model of ARFI Imaging with 128 Channel ICE Catheter
One of the significant challenges to be overcome to bring ARFI guided ablation to the clinic is to extend the range at which ARFI images can be made. The preliminary data in Sec C.1 demonstrates ARFI images of lesions made at a distance of about 2.5 c m. Our phantom, in vitro, and in vivo studies indicate that the 64 channel ICE catheter has a maximum ARFI imaging range of 3 c m. We present simulation results that show that the 128 channel array being built by Siemens Medic al Systems will be capable of ARFI imaging at 5-6 c m.
In 2007, we will receive 128-channel intra-cardiac arrays (see letter of support in Appendix). This catheter will be the same diameter as the commercially available AcuNav (10 fr.) catheter; but, the increased number of transmitting elements will double the imaging aperture width.
The impact of a larger array and the increased aperture will have multiple benefits for ARFI imaging. By doubling the number of channels, the effective pressure of the “pushing” pulse around the focus will be doubled. Also, the central limit theorem states that doubling the number of receive channels will reduce the noise within the beam formed receive lines by a factor of 2. Therefore, the amplification of signal and reduction of noise will cooperatively increase the signal to noise ratio (SNR) of the received reference and tracking lines by a factor of sqrt(2).
The radiation force magnitude is linearly proportional to intensity , which is also linearly proportional to displacement. Thus, the doubling of aperture size will result in a quadrupling of tissue displacements around the focus.
An ultrasound simulation program, Field II was used to examine the effective increase to the ARFI imaging pushing power. Transmit intensities of radiation force pulses focused at various depths along the center axis were calculated for both the 64 and 128-element arrays. The resulting normalized plots are shown in Figure 7. These plots reflect a near quadrupling of the maximum intensity. Currently, we have only demonstrated ICE ARFI imaging to be effective at depths less than 2.5 c m. However, the simulations suggest that the 128-element array can produce an equivalent amount of radiation force at a depth of 5.1 c m.
Doubling the lateral transducer width of the 128 channels array will also improve the lateral resolution of this array by a factor of two over the 64 channel array, for both B-mode and ARFI images. We predict the lateral resolution of the new array will be approximately 1 mm at a 4 cm depth, for both B-mode and ARFI imaging. We hypothesize that the improved transmit energy, receive sensitivity, and resolution of the 128 channel array will markedly improve its ability to image ablation lesions.
Safety of Intracardiac ARFI Imaging
The FDA (ref) describes regulatory guidelines for Mechanical Index (MI), Thermal Index (TI) and transducer heating for ultrasound imaging. Here we show our preliminary estimates and simulations of these parameters as they relate to ARFI imaging for this application.
ARFI imaging utilizes transmit pulses similar in pressure to those in B-mode imaging, but 100-200 times longer. The long “pushing pulses” utilized by ARFI have a low PRF (typically 10-60 pulses/sec), much lower than the 5-10 kHz PRF of B-mode and Doppler systems. The MI, which reflects the potential for cavitation, is very similar for ARFI and B-mode given their similar transmit frequencies and pulse pressures. The TI is greater for ARFI pulses by an amount proportional to their increased pulse length but also reduced by their decreased PRF.
We have developed and experimentally validated finite element models that evaluate soft tissue heating due to thermal absorption of the radiation force pulses39,40. LSD-DYNA3D (Livermore Software Technology Corporation, Livermore, CA) is used to solve for the dynamic temperature fields and thermal expansion displacements with a time-domain, explicit iterative solver. Modeling a homogenous region of tissue with material properties based on values for myocardial tissue found in literature, the three dimensional heating curves from an entire ARFI image acquisition was det ermined. The resulting elevational slice that included the maximum temperature rise of 0.027oC is shown in Figure 8. This simulation modeled a static region of tissue and did not account for cardiac motion, blood perfusion, and flow which would reduce heating estimates.
For soft tissue, temperature increases less than 1.0oC are considered acceptable during diagnostic ultrasound scanning and increases of up to 6oC do not require FDA justification41. Results from these simulations indicate that the long-term tissue heat accumulation due to ARFI imaging at frame rates relevant to ECG-gated acquisitions (1-2 Hz) will be negligible (i.e. less than 0.5oC). For high frame rates that approach continuous realtime imaging, additional simulations that more accurately model the environment of a beating heart will be necessary.
Another safety concern is transducer face heating. A thermocouple was placed on the center of the transducer face and catheter submerged in a water tank. ARFI images were acquired at a sampling frequency of 0.5 Hz. The resulting temperature plot is shown in Figure 9a. The plot shows the greatest temperature rise on the transducer face was approximately 2.5oC. We used a mechanical piston to move the water surrounding the transducer and simulate blood movement in the heart chamber. In the presence of minimal flow (f = 11 bpm, vmax=0.25m/s), the maximum temperature increase was approximately 1.5oC, with negligible heat accumulation on the transducer face as it was quickly dissipated within the surrounding streaming fluid.
For this test we used a conventional ARFI imaging sequence that did not utilize parallel tracking methods. Parallel tracking allows a fourfold decrease in the number of pushing pulses employed. Therefore, the temperature plots associated with this setup would be equivalent to parallel-tracking ARFI image acquisitions formed at four times the sampling rate (2.0 Hz). With these simulations and experimental temperature measurements, we have demonstrated that intra-cardiac ARFI imaging with a 64-element array has little risk of causing thermal damage to the tissue. These simulations and experiments will be repeated to determine the thermal safety of the new 128-element array.