Intra-Tissue Pressure Measurement in Ex Vivo Liver Undergoing Laser Ablation with Fiber-Optic Fabry-Perot Probe

We report the first-ever intra-tissue pressure measurement performed during 1064 nm laser ablation (LA) of an ex vivo porcine liver. Pressure detection has been performed with a biocompatible, all-glass, temperature-insensitive Extrinsic Fabry-Perot Interferometry (EFPI) miniature probe; the proposed methodology mimics in-vivo treatment. Four experiments have been performed, positioning the probe at different positions from the laser applicator tip (from 0.5 mm to 5 mm). Pressure levels increase during ablation time, and decrease with distance from applicator tip: the recorded peak parenchymal pressure levels range from 1.9 kPa to 71.6 kPa. Different pressure evolutions have been recorded, as pressure rises earlier in proximity of the tip. The present study is the first investigation of parenchymal pressure detection in liver undergoing LA: the successful detection of intra-tissue pressure may be a key asset for improving LA, as pressure levels have been correlated to scattered recurrences of tumors by different studies.


Introduction
Thermal ablation (TA) has gained significant interest from clinicians and scientific community, for the mini-invasive treatment of tumors [1][2][3][4][5][6]. The principle of operation of TA relies on a miniature, typically percutaneous, applicator on active tip(s) positioned at the point of treatment, and connected to an electromagnetic source: a spatially confined heat field is induced in proximity to the active tip(s), and propagated to the adjacent tissue [1]. Methods using radiofrequency ablation (RFA) [2,3], microwave ablation (MWA) [4], high-intensity focused ultrasound (HIFU) [5], and emerging techniques based on gold nanoparticles and light-activated nanoparticles [6] have been reported for ablation of tumors. With respect to resection, TA methods provide minimal invasiveness for patients, often resulting in outpatient care, and enable repeatable treatments.
Laser ablation (LA) stands as the main alternative to electrical-based treatments [7][8][9][10]. LA makes use of a mid-power laser source, coupled in an optical fiber that serves as applicator, which is positioned inside the tumor. Applications of LA to hepatic [7], pancreatic [8], thyroid [9], and brain tumors [10] have been reported. Compared to mini-invasive RFA [2], LA yields an even smaller invasiveness, with a setup that mimics in-vivo operation. The obtained results are compared with the previous literature, and discussed in view of setting the basis for real-time intra-tumoral pressure measurement in clinical operation.

Experimental Setup
The experimental setup for LA and pressure detection is shown in Figure 1, whereas LA and pressure measurement operate as separate blocks. LA is performed with a solid-state Nd:YAG laser (Echolaser X4, Elesta s.r.l., Florence, Italy), activated through a laser controller that allows selecting the output power. The laser emits around 1064 nm, and is operated in continuous wave. The applicator is a large-core fiber, having 300 µm diameter. Ablations were performed on freshly excited hepatic porcine tissue, which were obtained from locally bred pigs and stored in a refrigerator prior to perform LA experiments.
Sensors 2016, 16,544 3 of 11 ex vivo on porcine liver, with a setup that mimics in-vivo operation. The obtained results are compared with the previous literature, and discussed in view of setting the basis for real-time intra-tumoral pressure measurement in clinical operation.

Experimental Setup
The experimental setup for LA and pressure detection is shown in Figure 1, whereas LA and pressure measurement operate as separate blocks. LA is performed with a solid-state Nd:YAG laser (Echolaser X4, Elesta s.r.l., Florence, Italy), activated through a laser controller that allows selecting the output power. The laser emits around 1064 nm, and is operated in continuous wave. The applicator is a large-core fiber, having 300 μm diameter. Ablations were performed on freshly excited hepatic porcine tissue, which were obtained from locally bred pigs and stored in a refrigerator prior to perform LA experiments. The EFPI probe, depicted in Figure 1b, has been fabricated as an all-glass, biocompatible assembly. The structure has been fabricated by means of fiber splicing and polishing, both manually performed. In order to preserve the all-glass structure, the pressure-sensitive diaphragm is a glass optical fiber (multi-mode fiber 62.5/200 μm); the Fabry-Perot cavity is formed within the standard single-mode input fiber and the multi-mode fiber serving as diaphragm. This arrangement allows a good fabrication control, as the diaphragm and capillary have same outer thickness and a very similar composition, which allows to be treated in a glass splicing machine. Initially, the multi-mode fiber is spliced to a quartz capillary, having 130 μm inner diameter and 200 μm outer diameter. The capillary was chosen as in [25]. With a manual splicer (Siecor X77), a single mode fiber SMF-28 (10/125 μm) is pushed within the hollow capillary, leaving a gap of about 38 μm length, that will serve as Fabry-Perot cavity, obtained between the multi-mode and single-mode fibers. All fibers are λ/10-polished prior to being spliced. The obtained cavity structure is subsequently adjusted to pressure sensing by reducing the thickness of the multimode fiber: in first step, the diaphragm is shortened to ~10 μm by means of fiber polishing, and subsequently wet-etching is performed shrinking the diaphragm thickness to approximately 2 μm. The resulting spectrum is shown in Figure 1c, as detected by the spectrometer.
The EFPI interrogator has been assembled in house as a portable device. A superluminescent LED (SLED, Exalos 2100 series) is used as light source, having 1 mW power on 60 nm bandwidth, and a Gaussian spectral profile; the SLED is powered by its driver board (Exalos EBD5200). Light The EFPI probe, depicted in Figure 1b, has been fabricated as an all-glass, biocompatible assembly. The structure has been fabricated by means of fiber splicing and polishing, both manually performed. In order to preserve the all-glass structure, the pressure-sensitive diaphragm is a glass optical fiber (multi-mode fiber 62.5/200 µm); the Fabry-Perot cavity is formed within the standard single-mode input fiber and the multi-mode fiber serving as diaphragm. This arrangement allows a good fabrication control, as the diaphragm and capillary have same outer thickness and a very similar composition, which allows to be treated in a glass splicing machine. Initially, the multi-mode fiber is spliced to a quartz capillary, having 130 µm inner diameter and 200 µm outer diameter. The capillary was chosen as in [25]. With a manual splicer (Siecor X77), a single mode fiber SMF-28 (10/125 µm) is pushed within the hollow capillary, leaving a gap of about 38 µm length, that will serve as Fabry-Perot cavity, obtained between the multi-mode and single-mode fibers. All fibers are λ/10-polished prior to being spliced. The obtained cavity structure is subsequently adjusted to pressure sensing by reducing the thickness of the multimode fiber: in first step, the diaphragm is shortened to~10 µm by means of fiber polishing, and subsequently wet-etching is performed shrinking the diaphragm thickness to approximately 2 µm. The resulting spectrum is shown in Figure 1c, as detected by the spectrometer.
The EFPI interrogator has been assembled in house as a portable device. A superluminescent LED (SLED, Exalos 2100 series) is used as light source, having 1 mW power on 60 nm bandwidth, and a Gaussian spectral profile; the SLED is powered by its driver board (Exalos EBD5200). Light from the optical source is directed to the probe through a 50/50 fiber-optic coupler. Light reflected by the EFPI is collected with a high-speed spectrometer (Ibsen Photonics I-MON-512-USB), operated on the 1520-1596 nm bandwidth. A LabVIEW™ program has been developed for data acquisition (DAQ), recording data from the spectrometer and switch and displaying the estimated output pressure; the software adjusts also the integration time of the spectrometer. The software also compensates the uneven spectrum of the SLED source, and adjusts the sampling frequency to approximately 10 Hz.
EFPI calibration was performed in an air-filled tank, testing for pressure levels up to 100 kPa, and dynamically using a 60-cm burette filled with water, for fine detection of pressure up to 5.8 kPa.
The pressure sensitivity, measured as air-gap compression as a function of applied pressure, is estimated as 1.61 nm/kPa, very close to [23] (1.5 nm/kPa) and [25] (1.6 nm/kPa). Figure 2a reports the pressure calibration performed in the air-filled tank, with pressure ranging from 0 kPa to 100 kPa, measured with a reference MEMS sensor (reference 0.1 kPa). Thermal effects were recorded by placing the probe in a temperature chamber, in absence of pressure changes; a low thermal sensitivity has been recorded, estimated as´0.05 nm/˝C. Thus, a thermal detuning of´0.03 kPa/˝C was obtained, which is to the best of our knowledge one of the lowest values for miniature-sized fiber optic pressure probes in absence of temperature compensation (approximately 350 times smaller than [23]). Figure 2b reports the thermal calibration performed in a temperature chamber, from 18˝C to 82˝C; the slope of the curve is estimated as 52.2 pm/˝C. In absence of fluctuations of refractive index of the outer medium, pressure measurement yields accuracy better than 0.1 kPa [25]. The results of the calibration are in line with the previous pressure sensing system presented in [25] and developed by University of Limerick, which makes use of the same splicing method and capillary/diaphragm structure. from the optical source is directed to the probe through a 50/50 fiber-optic coupler. Light reflected by the EFPI is collected with a high-speed spectrometer (Ibsen Photonics I-MON-512-USB), operated on the 1520-1596 nm bandwidth. A LabVIEW TM program has been developed for data acquisition (DAQ), recording data from the spectrometer and switch and displaying the estimated output pressure; the software adjusts also the integration time of the spectrometer. The software also compensates the uneven spectrum of the SLED source, and adjusts the sampling frequency to approximately 10 Hz. EFPI calibration was performed in an air-filled tank, testing for pressure levels up to 100 kPa, and dynamically using a 60-cm burette filled with water, for fine detection of pressure up to 5.8 kPa.
The pressure sensitivity, measured as air-gap compression as a function of applied pressure, is estimated as 1.61 nm/kPa, very close to [23] (1.5 nm/kPa) and [25] (1.6 nm/kPa). Figure 2a reports the pressure calibration performed in the air-filled tank, with pressure ranging from 0 kPa to 100 kPa, measured with a reference MEMS sensor (reference 0.1 kPa). Thermal effects were recorded by placing the probe in a temperature chamber, in absence of pressure changes; a low thermal sensitivity has been recorded, estimated as −0.05 nm/°C. Thus, a thermal detuning of −0.03 kPa/°C was obtained, which is to the best of our knowledge one of the lowest values for miniature-sized fiber optic pressure probes in absence of temperature compensation (approximately 350 times smaller than [23]). Figure 2b reports the thermal calibration performed in a temperature chamber, from 18 °C to 82 °C; the slope of the curve is estimated as 52.2 pm/°C. In absence of fluctuations of refractive index of the outer medium, pressure measurement yields accuracy better than 0.1 kPa [25]. The results of the calibration are in line with the previous pressure sensing system presented in [25] and developed by University of Limerick, which makes use of the same splicing method and capillary/diaphragm structure.  Figure 3 shows a series of photographs of the experimental setup, fiber insertion, and ablation outcome. In particular, the insertion of the EFPI probe into the tissue is shown: a needle is firstly inserted to guide the fiber into the tissue, and subsequently removed to leave the fiber in place. With this setup, no fracture or damage to the probe, despite its miniature size, was observed. The setup mimics in-vivo insertion of a fiber-optic probe.  Figure 3 shows a series of photographs of the experimental setup, fiber insertion, and ablation outcome. In particular, the insertion of the EFPI probe into the tissue is shown: a needle is firstly inserted to guide the fiber into the tissue, and subsequently removed to leave the fiber in place. With this setup, no fracture or damage to the probe, despite its miniature size, was observed. The setup mimics in-vivo insertion of a fiber-optic probe.

Experimental Results
A set of four different experiments of LA, with simultaneous pressure monitoring, has been performed. Table 1 summarizes the parameters of the setup used in each experiment. Different levels of Nd:YAG laser optical power have been set, ranging from 3 W to 5 W; in each experiment, the laser was turned on for different durations, ranging from 52 s to 105 s. The tip-to-tip distance is defined as the geometrical distance between the LA applicator (center of the optical fiber connected to the Nd:YAG source) and the EFPI sensor (center of the tip-mounted diaphragm); thus, by increasing the tip-to-tip distance, the measurement is performed further away from the ablation point [8]. As the LA applicator and the pressure probe were positioned in a deep-seated position inside the tissue, the estimation of the tip-to-tip distance was made by measuring the geometrical distance between the probes from the pictures taken before and after each experiment. We positioned the fibers such that the tip-to-tip distance ranges from 0.5 mm to 5 mm, in order to verify pressure values and its temporal evolution in different parts of the tissue, with respect to the LA applicator tip. Three experiments were performed positioning the EFPI probe on the side of the LA fiber applicator, with the two fibers placed in parallel, as in Figure 3c-d, while a fourth experiment was carried out positioning the EFPI along the perpendicular direction to the applicator, as in Figure 3e. The results of the first experiment are reported in Figure 4. The first chart shows the evolution of the EFPI spectrum with time, showing the optical spectrum as detected by the spectrometer, each 3.4 s. We observe, during the initial part of the acquisition, a blue shift of the EFPI envelope, which progressively saturates until the procedure is finished. On the right chart, the estimated pressure is reported. As the laser is turned on, pressure rapidly grows for the first 20 s, approaching 60 kPa. In

Experimental Results
A set of four different experiments of LA, with simultaneous pressure monitoring, has been performed. Table 1 summarizes the parameters of the setup used in each experiment. Different levels of Nd:YAG laser optical power have been set, ranging from 3 W to 5 W; in each experiment, the laser was turned on for different durations, ranging from 52 s to 105 s. The tip-to-tip distance is defined as the geometrical distance between the LA applicator (center of the optical fiber connected to the Nd:YAG source) and the EFPI sensor (center of the tip-mounted diaphragm); thus, by increasing the tip-to-tip distance, the measurement is performed further away from the ablation point [8]. As the LA applicator and the pressure probe were positioned in a deep-seated position inside the tissue, the estimation of the tip-to-tip distance was made by measuring the geometrical distance between the probes from the pictures taken before and after each experiment. We positioned the fibers such that the tip-to-tip distance ranges from 0.5 mm to 5 mm, in order to verify pressure values and its temporal evolution in different parts of the tissue, with respect to the LA applicator tip. Three experiments were performed positioning the EFPI probe on the side of the LA fiber applicator, with the two fibers placed in parallel, as in Figure 3c,d, while a fourth experiment was carried out positioning the EFPI along the perpendicular direction to the applicator, as in Figure 3e. The results of the first experiment are reported in Figure 4. The first chart shows the evolution of the EFPI spectrum with time, showing the optical spectrum as detected by the spectrometer, each 3.4 s. We observe, during the initial part of the acquisition, a blue shift of the EFPI envelope, which progressively saturates until the procedure is finished. On the right chart, the estimated pressure is reported. As the laser is turned on, pressure rapidly grows for the first 20 s, approaching 60 kPa. In the subsequent 40 s, pressure slowly increases, almost saturating to 70 kPa. The total ablation time is 74 s, with a peak pressure of 71.6 kPa; after that, laser is turned off and pressure rapidly decreases. In this experiment, where we position the pressure probe as close as possible to the LA applicator (0.5 mm), we observe a nearly instantaneous rise of pressure in correspondence of the ablation start; pressure then stabilizes to a nearly steady value. the subsequent 40 s, pressure slowly increases, almost saturating to 70 kPa. The total ablation time is 74 s, with a peak pressure of 71.6 kPa; after that, laser is turned off and pressure rapidly decreases. In this experiment, where we position the pressure probe as close as possible to the LA applicator (0.5 mm), we observe a nearly instantaneous rise of pressure in correspondence of the ablation start; pressure then stabilizes to a nearly steady value.  Figure 5a shows the results of all four experiments, where pressure estimated by the EFPI is reported as a function of time. In the second experiment, having duration 41% longer and input power 40% inferior to the first one, we observe a pressure curve different from the previous experiment. In this case, pressure slowly increases for the first 58 s, up to 22 kPa; then, an abrupt rise is observed, reaching a steady-state value of 48 kPa after 80 s. The peak value is 48.4 kPa. In the last two experiments, as the tip-to-tip distance increases, pressure values are decreasing, as suggested in [25]: in the third experiment, pressure reaches a peak of 5.8 kPa, while in the last experiment, pressure changes appear to be minute (1.9 kPa peak).
As peak values are different and highly dependent upon the distance from the ablation tip, and ablation durations are different, Figure 5b proposes a comparison of all experiments, reporting the pressure in logarithmic units, and the time normalized between 0 (LA starts) and 1 (LA turned off). We observe the qualitative evolution of pressure throughout the experiments. In the first ablation, pressure rises quickly to reach a steady state approximately after 30% of the elapsed time. In the second experiment, a steady state is reached after more than half of the elapsed time. In the third experiment, with tip-to-tip distance of 2.5 mm, pressure has a steadier rise, and appears to settle to a steady-state value only at the end of the procedure. Finally, in the fourth experiment, in which the probe is perpendicular to the applicator, pressure exhibits a non-monotonic trend: after an initial rise, and a subsequent decrease, pressure rises slowly even after the laser source is turned off.   Figure 5a shows the results of all four experiments, where pressure estimated by the EFPI is reported as a function of time. In the second experiment, having duration 41% longer and input power 40% inferior to the first one, we observe a pressure curve different from the previous experiment. In this case, pressure slowly increases for the first 58 s, up to 22 kPa; then, an abrupt rise is observed, reaching a steady-state value of 48 kPa after 80 s. The peak value is 48.4 kPa. In the last two experiments, as the tip-to-tip distance increases, pressure values are decreasing, as suggested in [25]: in the third experiment, pressure reaches a peak of 5.8 kPa, while in the last experiment, pressure changes appear to be minute (1.9 kPa peak).

Discussion
The present work provides, to the best of our knowledge, the first measurement of intra-tissue pressure in hepatic phantom for LA. We compare this result with the current literature related to pressure measurement in mini-invasive thermal ablation. Table 2 and Figure 6 compare the main results of our investigation, with the previous experimental results obtained by Tosi et al., with a As peak values are different and highly dependent upon the distance from the ablation tip, and ablation durations are different, Figure 5b proposes a comparison of all experiments, reporting the pressure in logarithmic units, and the time normalized between 0 (LA starts) and 1 (LA turned off). We observe the qualitative evolution of pressure throughout the experiments. In the first ablation, pressure rises quickly to reach a steady state approximately after 30% of the elapsed time. In the second experiment, a steady state is reached after more than half of the elapsed time. In the third experiment, with tip-to-tip distance of 2.5 mm, pressure has a steadier rise, and appears to settle to a steady-state value only at the end of the procedure. Finally, in the fourth experiment, in which the probe is perpendicular to the applicator, pressure exhibits a non-monotonic trend: after an initial rise, and a subsequent decrease, pressure rises slowly even after the laser source is turned off.

Discussion
The present work provides, to the best of our knowledge, the first measurement of intra-tissue pressure in hepatic phantom for LA. We compare this result with the current literature related to pressure measurement in mini-invasive thermal ablation. Table 2 and Figure 6 compare the main results of our investigation, with the previous experimental results obtained by Tosi et al., with a similar EFPI probe, in RFA [25], and with results obtained for RFA by Kotoh et al. ex vivo [16] and in vivo [17]. Data from [21] is not included in the table and chart, due to the lack of information on the pressure sensor and temporal pressure charts. These, to the best of our knowledge, are the only results related to intra-tissue pressure measurement in mini-invasive thermal ablation designed for cancer care; all of them are related to hepatic tissue RFA. Normalized data are within [0, 1] range whereas 0 corresponds to ablation start and 1 corresponds to peak pressure, and its corresponding elapsed time. MEMS experiments have been digitized from [16] and [17].
In Table 1, the comparison among results achieved by current and previous studies is performed considering the peak-pressure value (i.e., the maximum value of pressure measured at the end of the ablation procedure), and the 75% of peak (i.e., normalized time for which the pressure value is 75% of the peak pressure). Moreover, the comparisons of pressure/time charts are reported  Figure 5; experimental results obtained with a similar EFPI probe by Tosi et al. [25] on RFA ablation of liver, at 0.1 cm (1) and 0.5 cm (2) distance between the probe and the applicator; experimental ex-vivo results of Kotoh et al. [16], obtained with a MEMS sensor at 3 cm distance between the ablation device and probe; in-vivo study by Kotoh et al. [17] for hepatic RFA on animals, recorded with a MEMS sensor. Data are reported in three different formats: (a) Pressure (logarithmic units) as a function of time; (b) Pressure (logarithmic units) and normalized time; (c) Both pressure and time are normalized. Normalized data are within [0, 1] range whereas 0 corresponds to ablation start and 1 corresponds to peak pressure, and its corresponding elapsed time. MEMS experiments have been digitized from [16] and [17]. In Table 1, the comparison among results achieved by current and previous studies is performed considering the peak-pressure value (i.e., the maximum value of pressure measured at the end of the ablation procedure), and the 75% of peak (i.e., normalized time for which the pressure value is 75% of the peak pressure). Moreover, the comparisons of pressure/time charts are reported in Figure 6, including: (a) original data; (b) time-normalized data; (c) normalization of both pressure and time.
In first place, we observe a disagreement in terms of methodology. In [15,16] the MEMS sensor does not withstand the four criteria for pressure sensor in TA, as in Section 1: the sensor has cylindrical shape having 1 cm diameter and 3 mm thickness, and the thermal error is estimated as 0.5 kPa/˝C. In addition, in [15] the phantom in enclosed in a sealed capsule that contributes to artificially increase the detected pressure: as a consequence, the peak pressure reported in [15] is 833 kPa, while the same sensor records 9 kPa pressure in vivo [16]. The methodology proposed in the present study and in [23] is more suitable, as the probe has 0.2 mm thickness, yielding a negligible alteration of the ablation pattern [14], and the estimated thermal detuning is 0.03 kPa/˝C: the low thermal error allows measuring the pressure at the point of treatment, in close proximity to the ablation peak, while [15,16] measured pressure in far field. The measurement is also more confined to a small volume, as the probe has 0.002 mm 2 active volume, compared to 235 mm 2 in [15,16]. On the other side, the EFPI probe is vulnerable to variations to refractive index outside of the active cavity, which often occurs during the pressure decrease after the TA procedure is discontinued.
The obtained results are in agreement with [25], and in partial disagreement with [16,17,21] due to the different methodology. In LA we observe a peak pressure of 71 kPa at 0.5 mm distance (tip-to-tip), and 48 kPa at 1.2 mm; in RFA we observe a peak pressure of 162 kPa at 1 mm distance, and 32 kPa at 5 mm distance. It is possible to conclude that pressure has a steep spatial gradient, which follows the steep thermal gradients often in excess of 3˝C/mm [2,8,14,15]. Hence, such differences between maximum pressure values during LA and RFA may be related to the different ablation pattern that the two sources induce: in fact, LA performed with one bare fiber is used to produce focal thermal lesion (maximum diameter of about 2 cm), whereas RFA allows obtaining larger ablation volumes (3 cm to 5 cm of diameter). These results appear in disagreement with [16], in which the pressure is recorded at 3 cm distance and 833 kPa peak is observed, but in reasonable agreement with [17] which records a peak value of 9 kPa at 15 mm distance. On the other side, recent studies on heat propagation in TA [26] appear to agree that the stress induced in the tissue is higher near the ablation tip(s), where heat is maximum.
An important factor for our analysis is the determination of the peak pressure value. To account for this parameter, which is shown in Figure 6c, we report in Table 2 the normalized time for which the 75% of the peak pressure is obtained. For LA, pressure reaches a steady-state relatively early, for normalized time of 0.24 at 0.5 mm tip-to-distance, and 0.63 at 1.2 mm distance; in RFA, instead, pressure is observed as a steep rise at the final ablation stage, when peak temperature approaches 100˝C, when the elapsed time is 83%-94% of the total ablation time.
As a challenge of mini-invasive TA is preventing scattered recurrences [18,19], the possibility of predicting the maximum intra-tissue pressure is therefore an asset. In RFA, highlighted as well in Figure 6b, a high peak pressure is observed; this is typically observed as a steep pressure rise in proximity of the water-to-vapor transition and such behavior makes it very hard to predict the final pressure value. Conversely, in LA our investigation shows that pressure reaches a steady value relatively early during the procedure; the closer the pressure probe is to the LA applicator, the earlier pressure reaches a regime condition. The possibility to better control pressure may give an additional advantage for LA, over other more established methods such as RFA, in the treatment of tumors. Furthermore, our investigation shows that in order to perform a predictive task, the probe has to be positioned as close as possible to the ablation tip, strengthening even more the requirements for minimum footprint and thermal insensitiveness.
This study sets the basis for further investigation of the parenchymal pressure measurement in LA, and its influence on the long-term lifetime expectance due to better control of scattered recurrences. A possible extension of this study in the field of LA is the investigation of pressure values during short-pulse laser-induced explosion of gold nano-particles injected within tumor, which allow selective nanophotothermolysis. The cellular damage is entailed by both temperature and vapour pressure produced by both water and gold vapours, therefore a fine control of pressure values can allow good performances of the treatment [6]. Moreover, this analysis can be also extended to other applications of laser ablation. For instance, lasers can be considered a valid alternative to scalpel in many different situations where it is crucial to limit collateral damage caused by electro-cautery [27]. In this scenario the abrupt heating of the tissue entails the generation and propagation of thermoelastic stresses. These pressure waves must be confined close to the cut (stress confinement), to optimize the healing process and cutting precision. Therefore, an analysis of tissue pressure may be useful to figure out the best laser settings (laser power, power, duration of pulse, and wavelength) to optimize the cutting.

Conclusions
In this work, we performed for the first time, to the best of our knowledge, a pressure measurement in laser ablation of hepatic tissue. Pressure is recorded with a miniature, biocompatible, and thermally insensitive fiber-optic probe based on EFPI. Experimental results have been obtained with different ablation parameters, and probe-applicator distance ranging from 0.5 mm to 5 mm. Measurements show peak values of 71.6, 48.4, 5.8, and 1.9 kPa at 0.5, 1.2, 2.5, and 5 mm tip-to-tip distance. In addition, pressure reaches a steady-state condition earlier, at shorter distances from the ablation device. A comparison with previous literature, based on fiber-optic sensors and MEMS for hepatic RFA, has been included. This investigation shows the premises for intra-tissue pressure measurement in LA, for improving tumor ablation by means of reduction of scattered recurrences.

Acknowledgments:
The authors acknowledge the Social Policy Grant "Optical fiber technology for supporting thermal ablation of tumors".
Author Contributions: D.T. and E.S. conceived and designed the experiments; P.S. and D.B.D. performed the experiments; D.T. analyzed the data; D.T., S.P. and G.L. contributed to materials and analysis tools; D.T., P.S., E.S. and E.L. wrote the paper.

Conflicts of Interest:
The authors declare no conflict of interest.