Radio frequency catheter ablation (RFCA) is used for coagulation and destruction of dysfunctional tissues in the fields of oncology [1], cardiology [2], dermatology [3], and vascular diseases [4]. One common application in cardiology is the elimination of abnormal electrical pathways responsible for cardiac arrhythmias [2], particularly those shown to be resistant to drug therapy [5]. Much like other thermal ablation procedures, RFCA results in localized coagulation and desiccation of the target tissue while avoiding uncontrolled damage to neighboring structures. The ablation procedure is generally guided by electrophysiological and anatomic mapping as well as by careful radio frequency (RF) power titration [6]. The size of the induced lesion is mainly determined by the extent and duration of the heat-affected area. Hence, real-time treatment monitoring is essential to optimize the outcome of the intervention. The ablation process is usually monitored via simple temperature or impedance measurements at the ablation tip [7,8]. However, heat diffusion and the use of irrigated ablation tips can substantially affect the size and shape of the heated area, resulting in failed treatments [9]. To this end, several imaging techniques have been proposed for ablation monitoring. For example, intravascular ultrasound (IVUS) and magnetic resonance imaging (MRI) allowed for a more precise placement and navigation of the ablation catheter and visualization of the RFCA-induced morphological tissue alterations [10]. Transformations of the tissue composition in coagulated or desiccated areas result in light scattering and absorption changes detectable via optical methods, such as spectroscopy or optical coherence tomography (OCT) [11]. Infrared thermal imaging furthermore allows the quantification of tissue temperature with high resolution, but is restricted to superficial tissues [12]. Ultrasound (US), x-ray computed tomography (CT), or MRI images were shown sensitive to temperature variations in the tissue; however, real-time mapping of lesion formation is impeded with these techniques due to either limited temporal resolution or low contrast [13].

Optoacoustics (OA) has been suggested for ablation monitoring as early as 1993 [14], chiefly owing to its high sensitivity to changes in optical properties resulting from chemical transformations in ablated tissues [15] and to temperature variations [16]. OA has been used for temperature monitoring in forming lesions [17] and recently adopted for volumetric tomographic ablation monitoring in real time [18]. Volumetric OA tomography has also been shown to clearly discern vascular and organ morphology as well as extrinsically labeled structures in vivo [19], making it highly suitable for precise anatomical navigation. The high imaging speed of state-of-the-art OA tomography is efficient in capturing the dynamics of RFCA treatments with subsecond temporal resolution in two [15] and three dimensions [18]. However, in previous studies the excitation light was delivered into the ablated area through thick layers of turbid tissues, limiting applicability in realistic clinical scenarios involving monitoring of deep tissue lesions [20].

Herein, we present a conceptually different approach for simultaneous RF ablation and OA monitoring (RAOM) of the lesion formation. It combines the delivery of both electrical current and pulsed light within a single catheter (Fig. 1) while detection of the generated OA responses is performed from outside the body using a spherical matrix array for optimal volumetric OA image formation. The integrated catheter consists of a bundle of 96 copper-coated multimode fibers [Fig. 1(a)]. The excitation light and electrical current are coupled into the proximal end of the bundle [Fig. 1(b)] and are delivered to the tissue at its distal end [Fig. 1(c)]. The individual, custom-made light guides (IVG Fibers, Toronto, Canada) consist of step-index multimode optical fibers with a silica core and a fluorine-doped glass cladding, enabling efficient propagation of visible and near-infrared light with a transmission efficiency of approximately 30%. The fibers have a core diameter of 200 μm (220 μm including cladding) and a numerical aperture (NA) of 0.2. The light guides are further coated with a 25 μm thin copper film. The copper coating was removed at the proximal end of the bundle to maximize light coupling efficiency. This was achieved by closely packing the fibers within a conventional optical fiber connector (inner diameter 2.5 mm, SMA905, Thorlabs, Newton, USA) and securing them using a high-temperature epoxy (353NDPK, Thorlabs, Newton, USA). After the epoxy was cured, the proximal end was polished to optical quality. Figure 1(b) shows the facet of the polished proximal end. A low-resistance electrical connection between the separate copper-coated light guides was achieved using solder in the vicinity of the proximal end. Additionally, a copper cable was soldered to the same location, enabling the connection to the RF generator. The copper-coated light guides align to each other in a nearly hexagonal pattern ensuring an even distribution of the ablation current at the distal tip. The bundle was embedded into a steel ferule with an outer diameter of 6 mm and an inner aperture of 4 mm using high-temperature epoxy and polished to optical quality. The minimal short-term bending radius of the copper-coated fibers was experimentally found to be 2 mm, comparable to conventional 200 μm fibers, while the bending radius of the assembled bundle was less than 8 mm. Figure 1(c) shows the polished facet of the distal end with and without light transmitted through the catheter. The assembled RAOM catheter [Fig. 1(d)] was electrically insulated using PVC tubing (Tygon, Carl Roth GmbH, Karlsruhe, Germany), only exposing the ablation tip at the distal end. A schematic of the simultaneous RF ablation and OA signal detection experiment is shown in Fig. 1(e). The ultrasound array consists of 256 detection elements distributed on a spherical cap with 90° apex angle ($0.59\pi$ solid angle) and 4 cm radius. Its individual elements have a central frequency of 4 MHz and 100% detection bandwidth, resulting in nearly isotropic 3D imaging resolution of $\sim 200\mathrm{\mu m}$ around the geometrical center of the sphere. OA signal excitation was achieved via an optical-parametric-oscillator-(OPO)-based laser (Innolas Laser GmbH, Krailling, Germany) coupled into the proximal end of the RAOM catheter. The distal end of the catheter delivered short ($<10\mathrm{ns}$) laser pulses with $\sim 6\mathrm{mJ}$ energy and pulse repetition rate of 10 Hz, resulting in light fluence of $\sim 48\mathrm{mJ}/{\mathrm{cm}}^2$ at the fiber tip. The wavelength of the laser was tuned to 780 nm, corresponding to the highest lesion-specific OA contrast [18]. The 256 detection channels were simultaneously digitized at 40 megasamples per second by a custom-made data acquisition system (Falkestein Mikrosysteme GmbH, Taufkirchen, Germany) triggered by the $Q$-switch output of the laser. The same trigger signal was used to switch off the RF current during the OA signal acquisition to avoid signal cross-talk. The acquired signals were deconvolved with the impulse response of the matrix array elements and bandpass filtered between 0.1 and 2 MHz to smoothen the images. The reconstructions were performed with a graphics-processing-unit-based 3D back-projection algorithm [21,22].

 

Fig. 1. Design and application of the combined RAOM catheter. (a) The catheter ablation tip, comprising copper-coated light guides embedded in a steel ferule using high-temperature epoxy. (b) Proximal end facet of the catheter, optimized for high light coupling efficiency. (c) Distal end facet of the catheter, delivering both RF current and pulsed-light illumination. (d) Fully assembled bundle. (e) Layout of the combined RAOM monitoring experiment.

Download Full Size | PDF

Performance of the RAOM catheter was first separately characterized in the OA imaging and RF ablation modes. For OA imaging, we used a two-layer agarose phantom [Fig. 2(a)]. The first layer mimicked strong tissue scattering and was used to quantify the OA signal levels generated by the catheter tip due to back-scattered light. It consisted of a 3 mm thick layer of agarose mixed with 1.2% (by volume) of Intralipid. The second 1.5 mm thick layer of the phantom mimicked tissue optical absorption of ${\mu }_a=0.66{\mathrm{cm}}^{-1}$ at 780 nm [23] and comprised agarose mixed with ink. The distal end of the catheter was positioned in direct contact with the scattering layer of the phantom, and OA imaging was performed without RF ablation. Figure 2(b) displays the side view of the recorded volumetric OA image where the absorbing layer is clearly visible at a depth of 3 mm in the phantom (P). Part of the light emitted by the catheter is back-scattered toward the ablation tip where it is absorbed by both the copper surrounding the light guides as well as the steel ferule [see Fig. 1(a)]. The signal generated at the catheter tip (C) is, however, much weaker in comparison to that generated by the tissue-mimicking absorbing layer. The catheter tip also acts as a partial acoustic reflector of the omnidirectional OA signals generated in the phantom. This results in shadow signals detected by the transducer (R). However, these artifacts do not interfere with the signals originating from the region of interest and can easily be cropped. The top view of the volumetric OA image shown in Fig. 2(c) further illustrates the uniform illumination provided by the RAOM catheter. Ablation performance of the RAOM catheter was subsequently evaluated by generating lesions in a porcine heart tissue sample. The catheter was connected to a custom-built generator allowing precise control of the output RF power. The samples were immersed in phosphate-buffered saline (PBS, Sigma Aldrich, St. Louis, USA) and the ground electrode with $20{\mathrm{cm}}^2$ area was positioned under the tissue. The catheter delivered tone bursts of electric current at 20 kHz carrier frequency with a duty cycle of 3% (600 cycles in a burst, 10 Hz repetition frequency). On average, 9 W of electric power was delivered for 10 s, 20 s, 40 s, and 60 s into the tissue samples. Photographs of the generated lesions are shown in Fig. 2(d). The catheter formed a homogenous white coagulum having a typical pallor and a small depression due to desiccation without any visible charring. Longer ablation durations generated deeper lesions, reaching a maximal depth of $\sim 1\mathrm{cm}$ after 60 s. The tissue beyond the coagulation region appears unaffected in all four tissue samples. The uniform lesion shape indicates a homogenous current distribution due to the evenly distributed copper-coated light guides.

 

Fig. 2. Characterization of the RFOA catheter in standalone OA imaging and RF ablation modes. (a) Experimental arrangement for OA imaging with the catheter placed in direct contact with a scattering and absorbing phantom. (b) Reconstructed volumetric OA images of the phantom containing signals generated by the absorbing layer (P) and the catheter tip (C) as well as the signal reflected by the catheter tip (R). Maximal intensity projection (MIP) along the $x$ direction is shown. (c) The corresponding MIP image made along the depth (z) direction over the range indicated by the gray dashed lines in panel (b). (e) Top- and side-view photographs of lesions generated in porcine heart tissue.

Download Full Size | PDF

The real-time ablation monitoring performance was then evaluated in a 4 cm thick porcine tissue sample, which was placed between the RAOM catheter and the surface of the spherical detection array. Ablation was carried out for 30 s, and OA signals were acquired for 180 s to cover the cooling period. Light fluence decay was volumetrically corrected by dividing the reconstructed volumetric image with the solution of the light diffusion equation for a point source, i.e., a 3D exponential decay in the form of $(1/d)\exp (-{\mu }_{\text{eff}}d)$, where ${\mu }_{\text{eff}}=3{\mathrm{cm}}^{-1}$ is the effective attenuation coefficient and $d$ is the distance in centimeters between the corrected voxel and the distal end of the fiber bundle. For distances $d$ smaller than the radius of the ablation tip, the fluence was assumed to be constant. This particular correction function was applied as a purely qualitative measure aimed at achieving better contrast uniformity across the OA images.

Figure 3 displays OA images of the porcine tissue sample prior [Fig. 3(a)], during [Fig. 3(b)] and after [Fig. 3(c)] the RF ablation procedure together with an OA signal time trace of a coagulated and a noncoagulated voxel [Fig. 3(d)]. As expected, the lowest OA signal intensity appears prior to the ablation due to the lowest temperature in the sample and lack of coagulation. A strong increase in the OA signal amplitude can be observed as the lesion progresses [Fig. 3(b), first 30 s in Fig. 3(d)]. The signal increase is attributed to the enhanced lesion contrast associated to tissue coagulation [15] as well as to the strong temperature dependence of the Grüneisen parameter in tissues [16]. While previously suggested approaches were afflicted by strong light attenuation in deep tissues thereby necessitating signal averaging [15], direct delivery of the excitation light into the ablated region via the catheter-based approach allows for monitoring of lesion without signal averaging. This represents a significant advantage of the integrated RAOM method for real-time clinical application. In Fig. 3(c) the OA images are further compared to gross pathology of the specimen taken after the RF ablation experiment, confirming a uniform coagulum without charring and a good qualitative correspondence between the yellow colored volume in the OA images and the appearance of coagulated area in the sliced specimen. As expected, cooler tissue [Figs. 3(c) and 3(d), $t=180\mathrm{s}$] exhibits lower OA signal levels as compared to the end of ablation time point ($t=30\mathrm{s}$), this is attributed to the temperature dependence of the OA signals [16]. The OA signal levels at $t=180\mathrm{s}$ remain higher than in the preablated specimen, supposedly due to incomplete cooldown and residual thermal diffusion effects in the rest of the sample. Note, however, that the signal in the coagulated zone [blue square in Fig. 3(b)] does not decline significantly during the cooldown period [blue plot in Fig. 3(d)]; this is ascribed to an increase in optical absorption coefficient caused by denaturized tissue proteins in the coagulum [17].

 

Fig. 3. Real-time OA monitoring of RF ablation in ex-vivo porcine heart tissue. (a) OA images acquired from unablated tissue. (b) Noninvasive OA imaging of the lesion formation in 3D. (c) OA image acquired after the cooldown phase along with the photograph of coagulated area from a sliced specimen. Images on the top are top-view MIPs while the bottom row shows a single ($z–y$) slice through the center of the reconstructed OA volumes, as indicated by the white dashed line in (a). (d) Time evolution of the OA signal during and after RF ablation in the locations indicated in panel (b).

Download Full Size | PDF

The presented results illustrate the basic feasibility of the suggested integrated RAOM approach for simultaneous RF ablation and real-time OA monitoring of the lesion progression. Because of direct light delivery through the catheter, the ablated region is efficiently illuminated, thus enabling monitoring of deep-seated lesions. Evidently, in vivo experimentation is essential to demonstrate the applicability of the proposed monitoring configuration in a real clinical setting. For this, several outstanding technical issues need also to be addressed. The ablation tip diameter of 6 mm is to be reduced to the typical 4 mm electrodes used in RFCA; this can be achieved by reducing the size of the encapsulating steel ferrule, packing the copper-coated fibers more densely, and/or reducing the number of fibers, the last also resulting in a more flexible catheter design. This would allow for the integration of the RAOM catheter into conventional steerable catheter shafts, thus adding additional functionality, such as electrocardiographic and temperature monitoring at the tip. Both the ablation catheter and its tip could be further adapted to fit different types of ablation procedures. For instance, RF tumor ablation is regularly performed with large-area ablation catheters, achievable by using a longer steel ferule at the ablation tip of the RAOM catheter [24]. The presented results indicate the basic feasibility of identifying changes related to tissue heating and coagulation with the suggested RAOM approach, making its potential combination with existing catheter and monitoring modalities simple and cost-effective. We observed dynamic changes in the OA images of forming RF lesions that corresponded well with the gross lesion pathology. However, the observed changes in the OA signal allow for only a qualitative assessment of the forming lesion, as the method does not allow for differentiating between alterations in the optical absorption due to coagulation versus changes of the temperature-dependent Grüneisen parameter. This can be possibly achieved via a multispectral imaging approach [25], thus attaining both real-time and quantitative feedback on the temperature distribution and the size of coagulated area during the intervention. Yet, real-time temperature mapping may still be possible in uncoagulated tissue areas where no alterations of the optical tissue properties have occurred [25].

In conclusion, the suggested catheter combining RF ablation and light delivery for OA excitation in a single flexible and adaptable design represents an advantageous solution for optimizing the outcome of RFCA interventions. The high spatiotemporal resolution and deep-tissue imaging capacity of the integrated ablation monitoring approach anticipate its general applicability in a number of RF ablation procedures.