An Individually Controlled Multitined Expandable Electrode Using Active Cannula-Based Shape Morphing for On-Demand Conformal Radiofrequency Ablation Lesions

of this article.


Introduction
As a leading cause of mortality worldwide, tens of millions of people suffer from various cancers each year and the number is still ever increasing. [1][2][3] Due to its high fatality rate in later stages, timely treatment is strongly suggested once malignant cancer is detected. Compared with other standard treatment modalities (e.g., chemotherapy and radiotherapy) that might cause severe side effects, [4][5][6] radiofrequency ablation (RFA) is regarded as a promising minimally invasive cancer treatment with relatively gentle side effects and has been widely adopted for years, [7,8] particularly for small-sized ones (diameters < 3 cm). [9] During RFA treatments, a key point lies in the generation of ideal RFA lesions that cover specific areas/volumes where target tissues exist. For large areal and highly efficient ablations, intensive efforts have been made for the development of new devices, such as expandable multitined electrodes and perfusion electrodes. [10][11][12][13][14][15][16][17][18] Although these devices can effectively ablate the target tissue either by increasing the surficial area of electrodes (e.g., expandable multitined electrodes) or by cooling the overheated surface of electrodes (e.g., perfusion electrode), the shapes of RFA lesions are almost the same and cannot be adjusted for target tissues that are irregularly shaped. [19][20][21][22] They usually end up with unnecessarily excessive ablation and subsequently often bring irreversible damage to the organs' functions. Therefore, in the clinical practice, conformal RFA lesions is desired as it can shorten the time that patients take to recover and improve their live quality, especially for the organs, for example, kidney, that have limited capability to regenerate after injury. [23] To satisfy the diverse morphologic demanded in RFA lesions, several groups proposed various designs. Among them, the most influential strategy is to utilize multiple electrodes. [24][25][26][27] The multiple electrode strategy, which distinguishes from the multitined electrode strategy, simultaneously inserts several electrodes into the body, and a preprogrammed alternative current is then used to energize the electrodes. The deployment of these electrodes can be delicately designed to better conform to the target tissues and avoid some unnecessary injuries of other organs. Being minimally invasive and highly effective, radiofrequency ablation (RFA) is widely used for small-sized malignant tumor treatment. However, in clinical practice, a large number of tumors are found in irregular shape, while the current RFA devices are hard to control the morphologic appearance of RFA lesions on demand, which usually ends up with unnecessarily excessive tissue ablation and subsequently often brings irreversible damage to the organs' functions. Herein, active cannulas for each of the individually controlled subelectrodes to achieve an on-demand shape morphing and thus conformal RFA lesion are introduced. The target shape as well as the length of inserted subelectrodes can be precisely controlled by tuning the active stylets and cannulas. What's more, owing to independent movement and energy control of each subelectrodes, the electrode is shown to be not only efficient enough to accomplish accurate trajectory control to target tissue in a single insertion, but also adaptive enough to ablate target tissues with diverse morphologic appearances and locations. On-demand conformal ablation of target tissue is demonstrated as well under the guidance of ultrasound imaging with the device. Potentially, the RFA electrode is a promising minimally invasive treatment of malignant tumors in future clinical practice. An interactive preprint version of the article can be found at: https://www.authorea.com/doi/full/ 10.22541/au.164019293.38729522.
However, the classic single polar electrode, as well as the bipolar electrode, usually leads to multiple insertions into the body and increases the pain of patients. [24][25][26][27] What's more, the precise deployment of these electrodes is complex and the electrodes have to be supported by other tools to ensure the correct relative position. [26,27] Hence, an operationally simple RFA device is urgently needed to achieve on-demand conformal ablation of diverse-shaped target tissues in a highly efficient way.
Herein, we present an individually controlled multitined expandable electrode for on-demand conformal RFA of target tissues, as shown in Figure 1A. Our electrode is designed with three tines (subelectrodes), each of which consisting of active cannulas and precurved stylet. The independent tunability and following interactions of stylets and cannulas enable our electrode to morph diverse shapes, and therefore the trajectory of individual subelectrodes can be predicted according to the expanded length. Combined with rational energy control, various morphing shapes can result in diverse sizes and shapes of RFA lesions, as shown in Figure 1B. The ex vivo irregular RFA lesions ablated by our electrode is more conformal than that ablated by a commercial one. Ultrasound-guided experiments further validate that our proposed electrode can change the morphologic appearance (diverse ablation volumes in different shapes) of RFA lesions to match the target tissue on demand.

Results and Discussion
According to previous research, [28] RFA lesions are highly associated with biophysical properties of tissues and will conform to the shape of electrode. To match the target tumor with morphologic appearance of RFA lesions, and thus minimize unnecessary injury of normal cells, one of the most effective approaches is to properly design the electrode. [9] As shown in Figure 2A, a flexible precurved stylet and a relative stiff coaxial cannula are coupled to tune the final expanded-out electrode shape by adjusting the relative axial position of them. Three individually controlled subelectrodes are integrated in one electrode for the easy control of energy output, and hence 3D on-demand conformal ablation can be achieved in an efficient way (detailed specifications of our RFA electrode are elaborated in Supporting Information and Figure S1).

Design of RFA Device
Design and implementation of the individually controlled subelectrodes and their stylets and cannulas is illustrated in Figure 2B. Subelectrodes can be withdrawn/expanded inside the outer shaft of electrode, and the head of the electrode is designed to be a cone for better tissue insertion. The outer shaft is covered with an insulating polymer sheath for the electric insulation. To avoid cross infection, invasive parts of surgical tools are often designed as single use. However, as the control and driven module are essential parts in our electrode design, to balance cost and minimize cross infection simultaneously, our proposed RFA device consists of a disposable handle box where electrode and a reusable control box are integrated (photos in Figure S2, Supporting Information), which are both modularized components for convenient assembly and reassembly. In the handle box, each of the stylets and cannulas is fixed to a sider and they only have one degree of freedom, as shown in Figure 2C,D. Steel wires wound on the winders connect to the slider for tendondriven control of these stylets and cannulas. The winders are the interface of handle box that connect to the control box. Six servos in the control box actuate their winders accordingly via an array of couplings. These servos are preprogrammed to follow real-time instructions for various motions. In addition, our device is designed to be compatible with a commercial RF generator and the detailed operation method is elaborated in Supporting Information text and Figure S3.

Shape Morphing of Subelectrodes by Stylet and Cannula Tuning
Owing to the precurved shape and the bevel tip of the stylet (the detailed structure of stylet and cannula depicted in Figure S4, Supporting Information), the shape of subelectrode can be tuned on demand by adjusting the relative axial position of the expanded stylet and cannula. As illustrated in Figure 3A, www.advancedsciencenews.com www.advintellsyst.com the expanding of a subelectrode can be categorized into three morphing states that are featured by the length of stylet remained in its corresponding cannula (it has to be noticed that the expanded length of stylet is designed to be no less than that of cannula). State 1, the stylet is expanding with the cannula. In such a state, the expanded length of the stylet equals that of cannula. The cannula is deflected by the precurved stylet and thus creates a slightly deflected angle of the subelectrode.
(As illustrated in Figure 3A, the expanded length of the cannula is the only variable in this morphing state.) State 2, the stylet is expanded out of the cannula and the expanded part of cannula is occupied by the precurved part of stylet. In this state, the precurved part of the stylet is not fully expanded so the straight part of the stylet does not enter into the cannula. For a certain expanded length of cannula, the deflected angle is almost unchanged in this state and a circle can be used to fit the profile of the precurved part of stylet. [29,30] (As illustrated in Figure 3A, the expanded lengths of cannula and stylet are both tunable in this morphing state.) State 3, the stylet is expanded out of the cannula and the straight part of the stylet exists in the cannula.  www.advancedsciencenews.com www.advintellsyst.com Ideally, the straight part of the stylet will not deflect the cannula. Hence, for a certain expanded length of the cannula, the deflected angle in this state is smaller than that in state 1 and 2. (As illustrated in Figure 3A, the expanded lengths of cannula and stylet are both tunable in this morphing state.) All of these three states of a subelectrode can be observed in the air, as shown in Figure 3B. Of course, in the tissue, the damping force of nearby tissues can influence the shape as well during the expanding process. Hence, the influence of biological tissues is still worth being investigated, as shown in Figure 3C (original photos in Figure S5, Supporting Information). Due to the existence of tissue, the deflection of subelectrode is damped and results in relatively smaller magnitude of the deflection angle in morphing state 1, as shown in Figure 3D. In State 2, the stylet is expanded out of the cannula and begins to contact with the tissue. As mentioned earlier, the deflected angle is almost unchanged and the radius of the curved stylet also seldom changes. In State 3, not only the recovery of the deflected cannula is damped by the tissue, but also the bevel tip of stylet is further affected by the tissue. Therefore, in this state, the deflected angle of the cannula in tissue phantom is larger than that in ambient air, and the bending radius of the stylet in the tissue phantom is smaller than that in air.
Like other bevel tip needles that move inside bodies, [31] the morphing shape of our active cannula-based precurved stylet is also influenced by its insertion speed. By carefully analyzing the obtained data, it is found that the higher the insertion speed, the larger force the exerted on the subelectrode by the tissue phantom. Thus, a lower deflection angle is yielded in a reasonable range, as shown in Figure 3E. Meanwhile, the impact of speed-relevant force is more obvious in State 2 than that in State 3, and the higher the speed, the stronger the impact, as shown in Figure 3F. The quantitative analysis of the subelectrode provides a useful guidance for the shape morphing and trajectory control of our electrode in tissue phantoms.

Trajectory Control in Tissue Phantoms
To ablate the target tissue on demand, the accurate trajectory control of the electrode is necessary. Here, to reach for certain target points, the regions of insertion positions are quantitatively analyzed using the measured data in Figure 3. As shown in Figure 4A, the regions of insertion positions are plotted and compared with the total region. The insertion angle of our electrode is fixed to mimic the practical condition that might occur during an RFA. Owing to the shape tunability of our subelectrode, the electrode tip can theoretically reach the target within a wide range of insertion positions. The insertion position region is becoming smaller and moving closer to the target point with the increase in expanding speed. Such a speed-relevant characteristic might be utilized to compensate the insertion error. What's more, plenty of choices of insertion positions will help to avoid inserting into vital vessels or hurting important functional parts in the body and further reducing the possible risks that might be caused during ablation. The accurate trajectory to a certain target is verified experimentally, as shown in Figure 4B,C. Given a certain expanding speed, our electrode can successfully reach the target point at different insertion positions.

Morphologic Characterization of RFA Lesions
Through finite-element analysis, the mechanism of on-demand conformal RFA is further investigated. As the morphologic appearance of RFA lesions mainly contains size (or ablated volume) and shape, these two aspects are considered in our performance evaluation. As shown in Figure 5A, the heat will accumulate in the surrounding tissues and thus cause overheating, which prevent further enlargement of RFA lesions. As further revealed in Figure 5B, rapid impedance increase occurs with the slowdown of the increasing rate of ablated volume. The ablated area starts from the tip of the electrode and finally forms a shape that is similar to the shape of electrode (insets in Figure 5B and S6, Supporting Information). Owing to such an effect, we are able to adjust the expanded length of the electrode to adapt diverse shapes of RFA lesions and achieve different volumes of RFA lesions, as shown in Figure 5C. Apart from ablation time and expanded length, the applied voltage/power also plays an important part in the determination of RFA lesions. Within our investigated range, the ablated volume peaks at %12 V and it overpasses four times the volume at 8 V, as shown in Figure 5D. In fact, low voltage/power heats the nearby tissue slowly and the overheating of tissue does not occur within the investigated range (curves of 8 and 10 V in Figure S7, Supporting Information). In contrast, high voltage/ power heats the tissue rapidly, and the higher the voltage/power, the earlier the overheating occurs (curves of 12-18 V in Figure S7, Supporting Information). Therefore, it is also important to choose a proper voltage for RFA planning. As our subelectrodes are individually controlled, any subelectrodes can be expanded on demand to enlarge the ablated volume of electrode, as shown in Figure 5E. Moreover, by expanding all of the three stylets and cannulas individually in a different setting, it will consequently result in distinct shape and volume of RFA lesions, as shown in Figure 5F. According to our simulation investigation, there still exist some RFA lesions that are unachievable by single ablation. In principle, for all RFA devices, the higher the degree of freedom (DOF), the better the conformality. However, higher DOF also means higher operation/control complexity and hence higher expense. Thus, a trade-off between conformality and operation complexity finally results in current design that is capable of conformally ablating various target tissues through simple operations. Furthermore, several supplementary strategies can be adopted for improving conformality between the target tissues and RFA lesions. For instance, sequential combinations of axial, radial, and circumferential morphing (motions of cannula, stylet, and rotation of electrode, respectively) can be used to conform to target tissues (e.g., after the first ablation and withdrawal of subelectrode, the main electrode can rotate and the subelectrodes can be expanded for another ablation. Cylindrical coordinate system of our electrode refers to Figure S8, Supporting Information.) Last but not least, appropriate preoperative planning by either computers or physicians is also important to minimize the unnecessarily excessive ablations. [32][33][34][35] To sum it up, the RFA lesion of our electrode is influenced by three hierarchical factors. On the system level, the rationally controlled ablation time and applied voltage/power offer energy to www.advancedsciencenews.com www.advintellsyst.com the electrode. On the electrode level, three subelectrodes can individually expand to conform to target tissues and share delivered energy. On the subelectrode level, the stylet and cannula in each subelectrode further morph by adjusting their relative position and hence expanded length. Thanks to these factors, our electrode can thus have high potential to meet diverse demands of tumor treatments with various morphologic appearances.

Ex vivo On-Demand RFA
To further verify our proposed device, several ex vivo experiments are conducted on fresh porcine kidneys, together with a commercial ablation electrode (photo in Figure S9, Supporting Information) serving as a reference. Both electrodes are used to ablate a target tissue with same ablation parameters (ablation time, insertion depth, and ablation power). To better visualize and evaluate the RFA lesions, the ablated kidneys are sliced carefully at the surface where the (sub)electrode is. As in Figure 6Ai,ii, small RFA lesions with nearly circular and elliptic sectional contours are observed in almost identical shape and the measured sectional areas of RFA lesions are roughly the same. However, the stiffer commercial electrode is impossible to steer inside the tissue, which makes it hard to conform to a curved sectional shape, as shown in Figure 6Aiii. In contrast, our precurved electrode can steer on demand and ablate a curved sectional shape that is almost identical with the target tissue. In addition, such an on-demand conformal ablation is also applicable for other tissues, for example, a porcine liver in Figure S10, Supporting Information. Given the same expanded length but different morphing states of the subelectrode (which means the expanded length of stylet keeps constant), different RFA lesions with nearly the same sectional area can be achieved by changing expanded cannula length, as shown in Figure 6B, and it indicates that our electrode is able to ablate malignant tumors with the same size but different shapes. In addition, the insertion poses (which means the angle and position) of the electrode are restricted during the experiment to mimic the actual situation for the safety

Ex vivo Conformal RFA Under the Guidance of Ultrasound Imaging
An ex vivo ultrasound-guided RFA is conducted in accordance with the protocol of in vivo RFA, as shown in Figure 7A. Based on the morphologic information of target tissue, preoperative planning of electrode is done for our experiments. Following our planning, in-plane operation of our electrode is subsequently conducted to ensure the total visibility of the entire procedure, as shown in Figure 7B,C. Under the guidance of ultrasound imaging, our electrode accurately targets a small target tissue with an oval sectional view (major axis of 14 mm and minor axis of 7 mm) and ablates the targets tissue conformally, as shown in Figure 3B. With rational energy control, the ablated area is controlled to encompass the target tissue including a circumferential safety margin (the brightened area covers the dark region with a margin of 5-10 mm). [9] Apart from the small target tissue, an elongated bigger one (major axis of 24 mm and minor axis of 5 mm) is also successfully ablated with an ensured safety margin, as shown in Figure 7C and Movie S3, Supporting Information. It can be clearly observed that the ablation region starts from the tip of electrode, which agrees well with previous simulation results ( Figure S6, Supporting Information). In addition, due to instrument limitation, currently, our experiments provide 2D images of the whole RFA process, which is sufficient to validate the visibility and conformality of our electrode during the in vivo ultrasound-guided RFA. Although experienced physicians are able to clinically pinpoint 3D information of electrode and target tissues for successful conformal RFA, [26,27] we still strongly suggest computed tomography-fused (CT-fused) 3D ultrasound imaging for high-quality, real-time 3D guidance of the whole procedure if possible. [36]

Conclusion
We present an individually controlled multitined expandable electrode for the on-demand RFA lesions. The interaction between the precurved flexible stylet and relatively stiff cannula produces various shapes of our subelectrodes. By tuning the expanded length and expanding speed of the stylets and cannulas, diverse trajectories can be realized on demand. Along with a rationally applied energy, three individually controlled subelectrodes can achieve 3D on-demand conformal ablation efficiently. The ex vivo RFA experiments of porcine kidney and its comparison with a commercial electrode further demonstrate the ability of electrode to ablate RFA lesions on demand. Further ultrasound-guided RFA verifies the potential of our electrode to conformally ablate in vivo target tissue with various morphologic appearances. With appropriate preoperative planning and precise control of our electrode, our device might be an important part of future intelligent surgical systems.

Experimental Section
Preparation of Tissue Phantom: As a widely used tissue phantom, [37] agar (Shanghai Regal Biology Technology Co., Ltd, China) was adopted in this paper. 12 g agar as well as 8 g NaCl (Sinopharm Chemical Reagent Co., Ltd, China) were added in the 800 g boiled water and stirred for 12 min to be fully dissolved. Then, the solution was poured in a mold and cooled for 1 h to be cured and finally the tissue phantom was made.
Preparation of Tissue Phantom with an Opaque Target: The evenly mixed solution of agar was first poured in a mold and cured for 8 min until there was a thin layer of film formed on the surface. Then a droplet of ink was dropped on the film and we waited for 5 min until the ink was dried. Finally, the mixed solution was again poured in the mold and cooled for 1 h for curation. (The tissue phantoms with an Opaque target were molded for the visualization of our target in Figure 4.) Equipment: An RF generator (BanBianTian, China) worked at a power of 15 W in the experiments and the frequency of AC was nominally 550 AE 40 kHz. During our ultrasound-guided RFA experiments, the power of RF generator was kept at 45 W. A magnetic mixer (HS 7, IKA, Germany) was used for the mixing and heating of the solution. A color doppler diagnostic ultrasound system (S60 pro, SonoScape Medical Corp., China) was used for the ultrasound imaging of RFA process.
Image Processing: The captured colored pictures were imported into MATLAB 2019a to transform to grayscale images. After a threshold was selected for each set of images according to the brightness of the pictures, these grayscales images were then transformed to black and white images. By calculating the area of each pixel, the number of pixels enclosed by the ablated region was converted into area.
Finite-Element Analysis: COMSOL Multiphysics v5.6 (Stockholm, Sweden) was used for RFA analysis. All the electromagnetic and thermal properties of the tissues were provided by COMSOL Multiphysics. For the simplicity of our model, blood perfusion and heat flux with the external environment were not considered, and only the conductive cone head and subelectrodes were considered in this model. Arrhenius kinetics model was used to assess the damage of tissues. All the insets in Figure 5 are reprocessed for better visualization, and the original pictures exported from COMSOL Multiphysics are illustrated in Figure S11, Supporting Information.
Pork, Porcine Kidneys, and Liver Preparation: The experiments were carried out in accordance with the guidelines issued by the Ethical Committee of Huazhong University of Science and Technology for the National Key R&D program of China (grant no. 2017YFB1303100).
Ultrasound-Guided RFA: An ultrasound probe (type VC2-9) was used to examine the profile of target tissue. If possible, CT was suggested for the reconstruction of the 3D profile. According to the morphologic appearance of the target tissue, the insertion position of electrode and the expanded length of subelectrode was then determined. After that, the position of the ultrasound probe should be adjusted to make sure that the insertion position and the whole insertion trajectory of electrode could be observed during the RFA (that is to say, in-plane operation is suggested in the experiment). Before insertion, subelectrodes were withdrawn in the electrode. Subsequently, the electrode punctured tissue and penetrated to the target tissue under the guidance of ultrasound imaging. After the electrode was precisely deployed, subelectrodes were expanded as planned. Then, rational ablation power was selected for RFA. By observing the shape and size of the brightened area (ensure that a safety margin of 5-10 mm exists), the ablation was ceased accordingly. Finally, the electrode was withdrawn and removed.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.