Multidimensional thermal mapping during radiofrequency ablation treatments with minimally invasive fiber optic sensors

Temperature mapping is a key asset in supporting the clinician during thermal ablation (TA) treatment of tumors without adding additional risk to the TA procedure. Herein we report our experiments on multidimensional thermal mapping during radio frequency (RF) thermal ablation treatments of an ex-vivo animal organ. The temperature was monitored using several arrays of fiber Bragg gratings properly positioned around the RF applicator. The results show the effectiveness of our proposed method at assessing the TA probe depth and demonstrating how the insertion depth directly influences the maximum temperature and the treated area of the radio frequency ablation. © 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement


Introduction
Radiofrequency ablation (RFA) for removal of solid tumors is a minimally invasive thermoablative technique applied to the field of surgical oncology [1]. Its main advantage over more invasive procedures, particularly in surgical resectioning (such as the Pringle Manoeuvre in cases of hepatocellular carcinoma), is its minor impact on surgical stress thus allowing for faster recovery which is most important in old or very weak patients. RFA procedures can be applied percutaneously in both laparoscopic and open surgery, and are applicable to several types of tumors such as liver, kidney and lung, as well as in treating metastases such as hepatic metastases due to colorectal cancer [2][3][4][5]. RFA consists of electrodes inserted into a lesion which cause an electrical current flow. This phenomenon leads to frictional agitation at the ionic level and a consequent heat generation known as the Joule effect [1]. Thermal treatment brings about a localized increment of the organ's temperature with consequent tissue dehydration and water vaporization. In this way coagulation necrosis takes place; a term used to describe the irreversible thermal damage which biological tissues undergo. Indeed, RFA's aim is to destroy cancerous cells by exposing them to cytotoxic temperature while sparing the adjacent healthy tissue.
The mortality rate of tumor cells is a function of the temperature value reached within the organ and of the time exposure [6]. In this way real-time temperature monitoring, which allows for adjustment of the RFA settings, can be considered valuable for achieving optimum clinical results. The knowledge of temperature during thermal ablation (TA) treatment gives the surgeon a way of controlling temperature distribution in the tissue surrounding the RF electrodes. This information reduces the need for repeated sessions to achieve complete necrosis for larger tumors [7] and the occurrence of potentially adverse effects to the surrounding healthy tissue [8]. Indeed, a deep insertion of the TA probe in the organ allows the heat to disperse to the blood vessels to better control eventual bleeding (a serious postoperative complication), but the risk of damaging the surrounding tissue increases. Therefore, the application time is eventually associated with an extended tissue damage which could be better controlled and limited when using an accurate temperature mapping.
RFA is generally carried out under ultrasound (US) guidance [9] which allows for precise positioning of the probe close to the target even during challenging procedures (e.g., intracavitary and endoluminal applications). However, accurate tumor localization alone is not sufficient for obtaining targeted treatment without damaging healthy surrounding tissue, as it does not allow quantifying cell's response to heat [10]. Indeed, it is still challenging to visualize the thermal dose delivered, and to delineate the damaged tissue margins with conventional US techniques because of the low intrinsic contrast between normal and ablated tissue and artefacts due to gas bubble formation [11]. It is for this reason that researchers have devised temperature maps by using a cross-correlation algorithm as applied to RF ultrasound echo signal data acquired at discrete intervals during heat treatment and caused by the speed of sound variation and thermal expansion with temperature. This thermometric technique however, requires calibration of the speed of sound variations and tissue expansion with temperature. Furthermore, US-based thermometry is affected by physiological motion and unexpected changes in acoustic properties of tissues [12]. Modern US systems are equipped with an image modality called Shear Waves Elastography (SWE), allowing to quantify tissue stiffness. Since a coagulated tissue is stiffer than a normal one, some studies are investigating the feasibility and accuracy of SWE for quantitative monitoring of thermal ablation [13]. Anyway, this technique is still in the early-stage, and cannot be expended to the monitoring of deep-seated lesions undergoing ablation therapy.
To overcome this limitation, several commercial RF probes are equipped with thermocouples in the tips of the electrodes which can measure the temperature of the adjacent tissue [12]. Nevertheless, this is not enough to accurately monitor temperature distribution of the tissue.
Other solutions aimed at visualizing three-dimensional temperature distribution in tissue during RFA procedures have been investigated such as the implementation of 3-D finite element models based on bio-heat equation whereby the results were corroborated with experimental data recovered through the use of thermocouples at the tip of the ablation electrodes [14]. However, the use of thermocouples may cause significant measurement errors due to their high thermal conductivity. Another solution to predict tissue destruction and cell death consists of a laparoscopic infrared camera capable of thermally mapping surface tissue temperatures [15]. The limitation of using a thermal camera however is that it is almost always limited to ex vivo analysis. Moreover, use of a laparoscopic thermal camera in clinical applications implies invasive surgical intervention and for this reason its level of invasiveness is incompatible with RFA which is a minimally invasive treatment [16].
The current landscape of temperature monitoring during RFA is facing with the challenges to obtain real-time and accurate temperature measurements by a minimally invasive approach. Therefore, multi-point temperature measurements performed with fiber Bragg grating sensors (FBGs) are a promising solution. FBG sensors have great potential in medicine, thanks to their biocompatibility, immunity to electromagnetic interference, nontoxicity, chemical inertness and small size [17]. Due to their unique features, several studies have been conducted on the use of these sensors for real-time temperature monitoring during RFA. Some studies focused on the monitoring of tissue temperature during RFA by means of linearly chirped fiber Bragg gratings (LCFBGs) [16,18]. This promising solution for distributed sensing presents the drawback of no reliable technique for their detection in the spectrum domain. Indeed, certain spectrum decoding methods require an overwhelming complexity and are not affordable in real-time operation, while simplified approaches substantially turn the LCFBG into a few-point sensor, which does not offer relevant advantages over uniform FBGs.
The current technology allows writing multiple FBGs in a single fiber enabling multipoint temperature measurements with high spatial resolution, inserting only one fiber optic within organs. Accordingly, other studies investigated the potential of FBG array for monitoring temperature in this field. Tosi et al. had directly mounted the FBG array on the ablation device, providing 5 points real-time temperature measurements (1 sensor/cm).
Saccomandi et al. performed temperature measurements during CT-guided RFA performed with the StarBurst XL Electrosurgical Device. They used two custom-made thermal probes embedding a total of 9 FBGs (2 sensors/cm), in addition to the 5 thermocouples embedded in the umbrella-shaped RF probe. The multipoint temperature monitoring by using FBG sensors at several distances from the applicator provided useful and additional information regarding the boundary of damaged volume. The following table summarizes the principal details of some of the FBG-based temperature monitoring systems solutions proposed in the literature (see Table 1). Building upon this research, we have recently proposed a FBG-equipped RFA probe for real-time monitoring of temperature during ablation, employing the Habib 4X RF commercial probe [25] measuring impedance between the electrodes as an indirect measure of temperature during ablation experiments. In previous reported experiments [24] researchers demonstrated that a linear FBG array was very useful for direct real-time temperature measure as it allows for potential optimization of the RF parameter (i.e. power and duration of the RF discharge).
Herein we present a significant improvement in the design of the sensorized RF probe based upon multidimensional temperature measurements. Its main strength is the housing of multiple FBG arrays with high spatial resolution, which provides increased valuable information about the temperature of the RF treated area. This approach is clinically relevant since the surgeon usually evaluates and establishes the RF probe's insertion depth with the sole aim of reaching the cancerous tissue, disregarding the substantial temperature increase of the organ as the electrodes' depth increases.
Our previous studies analyzed the temperature map and the distance from the RF applicator axis along the treatment without focusing on the above-mentioned insertion depth [23,24]. Herein we show how the proposed system is able to measure the temperature profile around the RF probe assessing that the TA probe's insertion depth directly influences the maximum temperature and the treated area. The obtained temperature mapping is a key asset that can support the clinician during treatment, without added invasiveness and ensuing risk to the procedure. Our proposed setup and results obtained are detailed, described and commented upon below.

Experimental setup
FBGs consist of a segment of optical fiber along which a spatially periodic modulation of the core refractive index is permanently made. The index modulation leads to the reflection of light in a narrow range of wavelengths while other wavelengths are transmitted along the fiber. The reflected range of wavelengths are centered on a specific value known as the Bragg wavelength λ B , which is expressed as: λ B = 2n eff Λ, where n eff is the effective refractive index of the guided core mode and Λ is the grating period. The grating is intrinsically a strain and temperature sensor and the external physical parameters are detected through the measurement of the reflected wavelength. For our purposes, the FBGs were employed in a strain-free configuration in order to consider the temperature contribution only, which is linearly related to the Bragg wavelength shift with the following expression: ∆λ B /λ B = S T ∆T, where S T is the thermal sensitivity coefficient of the grating [26].
In our experiments the Habib 4x laparoscopic bipolar Device RF probe was employed and consisted of 10 cm long electrodes (6 cm of inactive part and 4 cm of active part) that allowed for RF current circulation in a pig liver) [27]. Figure 1 shows the RF probe outfitted with 5 arrays for a total 27 FBGs representing the measurement sites. This configuration allowed for the processing of two-dimensional thermal maps in planes perpendicular and parallel to the probe's electrodes with a resolution of 0.1 °C.
FBG commercial arrays with the following characteristics were used in conducting the tests: • Array A of 3 FBGs -length of each grating 1 mm and distance edge-to-edge 3 mm.
Total length of 9 mm at x = −0.70 cm; • Array B of 7 FBGs -length of each grating 1 mm and distance edge-to-edge 2 mm. Total length of 19 mm at x = 0.35 cm; • Array C of 10 FBGs -length of each grating 1 mm and distance edge-to-edge 2 mm. Total length of 28 mm at x = 0.00 cm; • Array D of 3 FBGs -length of each grating 1 mm and distance edge-to-edge 3 mm. Total length of 9 mm at x = 0.70 cm; • Array E of 4 FBGs -length of each grating 1 mm and distance edge-to-edge 3 mm. Total length of 13 mm at x = 1.05 cm. Four FBG arrays were positioned at x = 0.00 cm, x = 0.35 cm, x = 0.70 cm and x = 1.05 cm in order to measure a temperature profile starting from the electrode's center. The last array was fixed at x = −0.70 cm to observe the symmetry of the temperature profile. Due to the length of the electrodes, a plexiglass support for the FBGs sensors was also fabricated to facilitate their positioning and make the structure easily removable from the RF probe, Fig.  1(b). The FBG sensors were then inserted into carbon fiber microtubes with an inner diameter of 0.30 mm a plexiglass sup larger area. Fi probe into the Several te tissues, in ord FBGs inserted Furthermo and an outer d pport and posi inally, the ends e organ and avo est have been c der to verify th d into them. ore, in Fig. 1

Results a
Experiments w Naples Feder RF probe and to compare th previously the 6.4·10 −6 and 6   After the f room tempera discharge was by FBG 9 an respectively. F Figure 5 r sample. The t arrays on the increases acco 97.1 °C regist Figure 5(a temperature a experienced a Arrays A an temperature i similar behav respectively. positioned at t the entire dim the center of t depths a temp ΔT≠0).   In particul while for a de 1.20 cm (from observed for distance of 0. fourth dischar distance of 1.5

Conclusio
Herein we pr treatments. W temperature perpendicular gives a comp knowledge. The