1. Introduction

In recent years, several studies have demonstrated the advantages of using lasers in the medical field of bone-cutting (osteotomy). Laser technology offers distinct advantages by enabling precise, sterile, and functional cuts to be performed remotely with a clear field-of-view, while preserving the porous structure of bone, promoting accelerated healing [1–5]. The contactless nature of laser osteotomy has the added benefit of incurring fewer mechanical effects in comparison to traditional bone surgery methods [6–9]. Photo-thermal ablation using mid-infrared wavelengths has shown promising results in laser osteotomy [7,10]. Bone tissue primarily consists of approximately 13 % water, 27 % collagen, and 60 % hydroxyapatite [11,12]. The Er:YAG laser (2.9 µm), with its high absorption peak in water molecules present in bone tissue, is an ideal choice for laser osteotomy [1]. Upon laser heating, the water molecules in the bone tissue rapidly evaporate, resulting in micro-explosions and subsequent removal of the bone tissue [13,14]. This ablation mechanism, commonly referred to as cold ablation in literature, offers high ablation efficiency while minimizing thermal damage [15].

Despite the added benefits of employing lasers in the medical field of osteotomy, it is crucial to note that the accuracy of lasers is contingent upon their level of control. Employing a real-time visualization and depth monitoring system during laser osteotomy can improve the accuracy of laser surgery, especially in minimally invasive procedures. For this reason, feedback systems are essential for providing surgeons with the information required during laser surgery [16]. To date, several methods for real-time monitoring and feedback have been introduced. To ensure that the laser-induced cut follows a pre-defined depth of cut, for example, monitoring techniques include photoacoustic tomography, linnik interferometers using a femtosecond laser, self-mixing interferometers, and cameras with high magnification optics [17–22]. There are some drawbacks associated with each of these techniques, such as angled views of the ablation cut, invasive detection, limited resolution, and limited application in miniaturized configurations. One technique that may overcome these drawbacks is optical coherence tomography (OCT), a non-invasive, high-resolution, and high-speed interferometric imaging modality with a wide range of applications in the medical field [23]. The first study to use an OCT system to monitor laser ablation successfully demonstrated its potential as a monitoring system [24]. Since then, several researchers have confirmed the potential of OCT as a real-time visual feedback tool during laser surgery [20,25,26].

With the advancement of OCT-assisted laser osteotomy in free-space, the next step is to develop a miniaturized integrated setup for use in minimally-invasive laser osteotomy [27]. Jivraj et al. demonstrated that a pulsed CO$_{2}$ (1064 nm) laser coaxially integrated with an OCT system using double-clad fiber could monitor a laser-induced cut up to 0.5 mm in cortical bone of porcine scapula specimens [25]. Bernal et al. have shown that a fiber-coupled Er:YAG laser (germanium oxide fiber) with a forward-cutting configuration can achieve a cutting depth of 6.82$\,\pm$ 0.99 mm in sheep bone [28]. Despite the promising outcomes of utilizing a germanium oxide fiber to deliver Er:YAG laser into an endoscope for laser osteotomy, certain limitations such as the high temperature of the fiber’s tip (approximately 70 $^{\circ }$C) and its limited working distance hindered its integration with a feedback system.

The present study builds upon our previous research, which focused on the integration of OCT and Er:YAG lasers in free space, where the challenges of this integrated system are investigated [5,10,20]. This study introduces a novel approach by integrating a miniaturized fiber-based Er:YAG laser, employing a sapphire fiber, with a long-range swept-source OCT (SS-OCT) system to enable real-time depth feedback during laser osteotomy. By optimizing the coupling mechanism, a high coupling efficiency of 75.7 % with a stable fiber’s tip temperature of $\sim$ 31 $^{\circ }$C at an input energy of 990 mJ is achieved. The combined systems were successfully incorporated into a compact probe with the dimensions of 13 mm $\times$ 42.1 mm x 10.13 mm for side ablation. In addition, a depth-controlled system is developed which could stop the ablative procedure by reaching the desired depth of laser-induced cut.

2. Methods

2.1 Miniaturized integrated setup

Figure 1 presents the schematic of the integrated setup designed for controlled bone ablation. A programmable akinetic swept-source laser (SLE-101, Insight Photonic Solution, Inc., Lafayette, CO, USA) was utilized, featuring a central wavelength of 1289 nm, a bandwidth of 62 nm, and an A-scan rate of 104.17 kHz. For the ablation process, an Er:YAG laser (R4X100C2-ER, MegaWatt lasers, USA,) producing 350 µs-long pulses was employed. In Fig. 1(a), the OCT laser is directed to the reference arm (10 %) and sample arm (90 %) using a fiber optic coupler (TW1300R2A1, Thorlabs). Then, OCT laser is delivered to the reference and sample arms using an optical circulator (CIR1310-APC, Thorlabs). The reference and sample arm outputs are connected to pigtailed ferrules (SMPF0115-FC, Thorlabs) and collimated using a grin lens with 0.23 pitch (GRIN2313A, Thorlabs). In the reference arm, the collimated beam is focused on the reference mirror (M1, PF10-03-M01, Thorlabs) using an achromatic doublet (L1, f = 30 mm, AC127-030-C, Thorlabs). The reflected/scattered light coupled back to the fiber is sent to a balanced detector (PDB480C-AC, Thorlabs) using a 50/50 fiber coupler. With this configuration, the OCT system has an imaging range of 2.25 cm in the air, with an axial resolution of 11 µm. During laser ablation of the bone, depth information is recorded using M-mode data over the time of laser ablation.

 

Fig. 1. (a) Schematic of the integrated setup, (b) schematic diagram of the integrated Er:YAG laser in the sample arm of the OCT system, (c) 3D-printed housing of the integrated setup.

Download Full Size | PDF

In Fig. 1(a), the output of the Er:YAG laser was controlled by an optical shutter (SH1/M, Thorlabs) that received commands from the OCT system. The optical shutter has a 15 ms response time based on the information provided by the manufacturing company. Fiber coupling of Er:YAG laser was performed by a CaF$_{2}$ lens (L2, f=50 mm, LA5763, Thorlabs). The sapphire fiber (Photran, USA) had a core diameter of 425 µm, with a numerical aperture (NA) of 0.12. Sapphire fibers serve the purpose of maintaining low tip temperature and minimizing beam divergence. However, a major drawback of using sapphire fibers is their cost because drawing sapphire fibers is challenging due to their high melting temperatures, leading to the production of costly pieces that typically offer limited length options. During the fiber coupling process, a thermal camera (FLIR A655sc, accuracy = $\pm$ 2 $^{\circ }$C) was used to monitor the temperature rise at the tip of the fiber. We used a sapphire fiber in these experiments because of the low temperature of the fiber tip and the small NA of the fiber [29]. The length of the sapphire fiber was $\sim$ 1 m. During the bone ablation experiments, the repetition frequency and the output energy of fiber were set to 5 Hz and 1 J/pulse, respectively.

Figure 1(b), shows an enlarged version of the integrated Er:YAG laser in the sample arm of the OCT system. It is worth noting that the required size of the optics was not readily available off the shelf. As a result, the optical elements underwent manual grinding. The required diameter of the optical components is attained by grinding the optics, which are mounted on a milling head, using 1000 and 2000-grade sandpapers. This process is performed with respect to the calculated beam size and designed housing. The 3D-printed housing is designed with a proper grooved holder for each optics which aids the initial alignment of the system. The OCT laser was collimated using a grin lens (L3) and focused using a polished achromatic doublet (L4, f = 30 mm, AC127-030-C, Thorlabs) with a diameter of 3 mm, and directed to the dichroic filter using a right-angle mirror (M2, # 65-844, Edmund Optics). The spot size is calculated as 51 µm. Fine alignment of the OCT is performed by adjusting the M2. Induced dispersion mismatched between the reference and sample arm is compensated using a standard dispersion compensator algorithm. The Er:YAG laser was focused on the sample using a polished lens (L5, f = 20 mm, LA5315, Thorlabs) with a diameter of 6 mm. The distance between the laser’s fiber output and the lens was 2.3 cm, and the corresponding beam diameter was measured to be 0.9 mm. The waist location was 4.9 cm away from the lens. The working distance of the current design is roughly 2.2 cm away from the sapphire window. The spot size was intentionally kept relatively large to ensure that no damage was induced on the dichroic mirror and to provide a long depth of focus. OCT and Er:YAG lasers were combined using a custom-made dichroic filter (DM, 97 % reflection at 2.94 µm, and 90 % transmission at 1289 $\pm$ 75 nm, II-VI Coherent, USA). For aligning the sapphire fiber, we utilize a fiber optic positioner (FP-1A, FPH-J, Newport) with the fiber’s end-tip. This positioner is mounted on a linear translation stage (PT1/M, Thorlabs) that offers the necessary degrees of freedom for fiber alignment. We carry out this alignment process using a 532 nm laser diode coupled to the fiber through a flip mirror within the path of the Er:YAG laser in free space. Finally, A 0.5 mm thick sapphire window was used in the miniaturized setup to protect the optics from debris generated during laser osteotomy (Ultitech Xiamen, China). Figure 1 (c) shows a photo of the 3D-printed housing of the integrated setup and its design.

2.2 Depth-controlled ablation

Depth-controlled ablation is performed by controlling the optical shutter with the OCT software. In free space, the depth-control software works based on the detection of the depth in the middle of the crater and comparing it with the reference lines (in the left and right of the ablation crater). This depth predication is enhanced by using the Kalman filter, which is based on a constant ablation rate model, and has the potential to stop the ablative laser before reaching the desired depth based on the prediction that the next pulse will ablate more than the pre-planned depth of the cut [30,31]. The reference point was set to be the surface of the bone before ablation. In addition, irrigation causes an accumulation of water inside the laser-induced cut, and the next pulse causes a water explosion which can dramatically affect the depth measurements. The detection of the water’s surface as depth of cut can cause an error. We have modified the depth detection algorithm to ignore the values which show a decrease in the depth of cut compared to the last measured depth.

3. Results and discussion

3.1 Er:YAG laser: fiber coupling

Coupling a high-energy laser into a fiber has always been a challenge, which has been clearly described in our previous study [29]. When employing fibers for surgical applications via endoscopes, the coupling mechanism must be highly robust. Thus, in this study, we implemented a thermal camera and an aperture to monitor the maximum temperature at the fiber tip and adjust the energy on the fiber through the beam size on the fiber. Before increasing the energy level, the aperture diameter was reduced to deliver approximately 50 mJ of energy into the fiber. At the next energy level, first, the tip position was optimized in the x, y, and z directions. Since the laser output consists of multiple longitudinal modes that may vary with different output energies and shift the focal plane, this step enhanced the repeatability of the fiber coupling process by observing the temperature increase of the fiber tip while repositioning the fiber, especially in the z-direction. The simultaneous temperature measurements helped to foresee the damage before it occurred. With the suggested method, once the fiber position was optimized, the setup required no further alignment for months. Concurrently, continuous monitoring of the tip temperature served as feedback to assess the risk of fiber damage, reinforcing the robustness of the coupling setup. For instance, tip temperature measurements around 31 $^{\circ }$C suggested "safe positions," while temperature readings above 35 $^{\circ }$C ended up damaging the fiber. Thus, this approach enabled the determination of the optimal position for the fiber tip, considering not only the transmitted energy but also the tip temperature. This outcome has paved the way for the development of a resilient fiber coupling system that holds the potential for future automation. Note that, diffraction can hinder the coupling efficiency with this method when the aperture blocks the beam. An alternative could be using a half-wave plate and a beam-splitter combination, however, a feedback signal may still be needed from the fiber tip to prevent any damage.

Table 1 presents the performance of the sapphire fiber at various energy levels. The sapphire fiber demonstrated an estimated maximum transmission of 80 % at the lowest input energy. The average coupling efficiency, which is closely aligned with the estimates provided by the manufacturer (Photran, USA), reached 76.5 %. The maximum input tip temperature recorded was approximately 31$^{\circ }$C. It was observed that the fiber tip suffered damage beyond this temperature threshold, potentially attributed to the high beam intensity at the center, considering a Gaussian beam profile for the laser. Damage due to dust during the experiments usually resulted in a momentary irreversible decrease in the transmitted energy. During the fiber coupling tests, however, as soon as the applied peak fluence was above 2.1 kJ/cm$^2$, the fiber tip was ablated. We performed the experiments at a safe level peak fluence level of 1.76 kJ/cm$^2$. Figure 2 (a) depicts the experimental configuration of the fiber coupling setup, where a thermal camera placed above the fiber’s tip monitored temperature changes. Figure 2(b) illustrates the maximum temperature rise experienced by the fiber’s tip during the coupling procedure. As previously mentioned, the spatial filter was gradually opened after the laser initiation, resulting in an increase in tip temperature. The vertical dashed line in Figure 2(b) indicates the time when the maximum laser energy was delivered into the fiber. The graph reveals that the fiber’s tip temperature fluctuated within the range of the maximum and minimum temperature values (indicated by horizontal dashed lines), confirming the stability of the temperature throughout the experiments.

 

Fig. 2. Er:YAG laser fiber coupling configuration. (a) Experimental setup, and (b) the temperature of the fiber tip during the fiber coupling procedure.

Download Full Size | PDF

Table 1. Evaluation of coupling efficiency and measured tip temperature of the sapphire fiber.

View Table

3.2 Depth-controlled laser osteotomy

The Er:YAG arm of the miniaturized integrated setup consisted of a focusing lens, dichroic mirror, and a sapphire window after the fiber. At the maximum input energy of 1000 mJ, the fiber was able to transmit 75.7 % of the input energy. After the optics, the transmitted energy was 420 mJ corresponding to 56 % of the delivered energy and a peak fluence level of 132 J/cm$^2$ on the sample. Optical instruments used in the miniaturized probe for the Er:YAG laser were prepared by polishing, grinding, and cutting, possibly decreasing the surface quality and increasing the losses. We believe the overall transmission can be improved up to 65 %. Figure 3 depicts the real-time monitoring of bone tissue ablation utilizing the developed OCT system. The experimental procedure was conducted on cow femur bone tissue. The acquisition of OCT images commenced prior to the initiation of the first pulse from the Er:YAG laser and ceased upon completion of laser ablation. In Fig. 3, the process of creating a laser-induced cut with a depth of 4.35 mm is illustrated. It is important to highlight that manual irrigation was employed during laser ablation to minimize bone tissue dehydration, while pressurized air was directed towards the cut to remove accumulated water. We employed manual irrigation by monitoring reduced ablation efficiency through sound generation and OCT depth detection [5,32]. With the help of manual irrigation, no visible thermal damage was induced in the bone samples, and clean line cuts were achieved. Furthermore, given the high water absorption of the Er:YAG laser, it is crucial to eliminate accumulated water within the laser-induced cut, as this water can act as a protective barrier and diminish ablation efficiency [29]. The vertical and horizontal lines observed in Fig. 3 correspond to the water explosion within the bone and the reflection from the protective window, respectively [20].

 

Fig. 3. Bone tissue ablation process as a function of time (M-mode). Vertical lines indicate splashes of the water accumulated in the laser-induced cut.

Download Full Size | PDF

As the next step, we created laser-induced cuts with pre-defined depths of 1 mm, 3 mm, and 5 mm. Real-time monitoring of the ablation crater using the OCT system allows us to provide this feedback and send a close command to the optical shutter when the desired pre-planned depth of the cut is reached. The depth of the cut is determined by comparing it to the bone’s surface before the start of the ablation. Figure 4 shows the corresponding results of the depth-controlled ablation using the miniaturized integrated setup. Performance was evaluated by making three laser cuts on the bone tissue at the three pre-defined depths (controlled using M-mode OCT image in real-time), and measurements were taken using the same OCT system with a free-space sample arm of OCT system which is described in our previous study [20]. This sample arm is empowered using a Bessel-like beam which extends the depth of focus up to 28.7 mm. For the pre-determined 1 mm, 3 mm, and 5 mm deep cuts, the mean values were 1.0487 mm (std = $\pm$0.0336 mm), 3.0873 mm (std =$\pm$ 0.1303 mm), and 4.972 mm (std =$\pm$ 0.2382 mm), respectively. The average ablation efficiency is measured as 0.338 $\mathrm {mm^{3}/s}$ in this experiment.

 

Fig. 4. Depth-controlled ablation of the bone tissue. (a), (b), and (c) present the pre-defined ablation depths of 1 mm, 3 mm, and 5 mm, respectively.

Download Full Size | PDF

4. Conclusions

The Er:YAG laser has demonstrated great potential in laser osteotomy, offering deep ablation with minimal thermal damage to surrounding tissues. Additionally, OCT-assisted laser ablation has proven to be a reliable system, providing real-time feedback on the depth of laser-induced cuts. In this study, we miniaturized the integrated setup by coupling the Er:YAG laser into a sapphire fiber. A maximum output energy of 750 mJ was achieved when ablating bone tissue, with a fiber tip temperature of 30$^{\circ }$C. However, the incorporation of an uncoated focusing lens, dichroic filter, and sapphire window led to a decrease in energy output to $\sim$420 mJ. Nevertheless, we were able to achieve a 5 mm ablation depth on bone tissue using the fiber-coupled Er:YAG laser at a repetition rate of 5 Hz. The miniaturized integrated setup exhibited a promising average ablation rate of 0.338 $\mathrm {mm^{3}/s}$, compared to previous studies involving the use of the Er:YAG laser in osteotomy [7,28,33]. As the width of the cut and the parameters related to irrigation and pressurized air play a crucial role in investigating the efficient ablation rate, further research is necessary to enhance its efficiency. We tested its performance with pre-determined ablation depths of 1 mm, 3 mm, and 5 mm. The depth-controlled ablation system errors were $\pm$0.0336 mm, $\pm$0.1303 mm, and $\pm$0.02383 mm, respectively. The Er:YAG laser has demonstrated the output energy of 1.056 J$\pm$61.22 mJ measured over 169 pulses. The main error in the depth-controlled ablation arises from the lack of real-time adjustment of the energy of the Er:YAG laser. For instance, in the depth-controlled ablation monitoring using the OCT system, the energy of the last laser pulse should be modified to remove the required depth indicated by the OCT system. In addition, relatively slow communication to the optical shutter can also contribute to the induced error.

Despite its distinct advantages, the miniaturized integrated setup does have certain limitations. Long sapphire fibers are costly and hard to produce mainly because of their high melting temperatures. However, because of their resilience to high energies, low divergence, and low tip temperature under laser illumination, they have great potential to be used in minimally invasive surgical applications of lasers. In our study, however, we used a 1-m fiber to test the integrated setup for bone ablation performance and to investigate challenges of integration, but in certain medical cases, this length could be limiting. To use sapphire fibers in clinical devices, the first step would be the extension of the length to at least 2 m. This could increase the losses and decrease the transmitted laser energy. In addition, since our primary goal is to achieve a robotically controlled minimally invasive surgical probe, we should consider placing the ablation laser and the fiber coupling setup close enough to be able to use this length. In real surgery, the miniaturized integrated system needs to be accompanied by suction and pulsed irrigation units. These additions would be necessary to avoid the blood layer on the bone surface blocking the laser ablation. Furthermore, the miniaturized probe would also require tissue-specific surgery, by incorporating miniaturized tissue detection systems.

The integration of the fiber-based Er:YAG laser with the OCT system in a compact configuration leads to increased housing temperature, which can hinder the use of scanning mechanisms (such as MEMS mirrors, micro motors, etc.) [34]. Due to the limited reflection of the dichroic mirror at 2.9 um, the leaking energy caused a temperature increase of up to 80$^\circ$C (at this point the laser beam was blocked before any damage). The damage to the 3D-printed material was prevented by doing the ablation experiment in discrete steps, allowing the material to cool down. This could be easily avoided by using a custom-made beam block. Furthermore, as demonstrated in a previous study, laser ablation of bone tissue generates debris that can accumulate on the protective window during laser osteotomy [20]. Although the Er:YAG laser can clean its path on the protective window, the surrounding area remains covered with debris. Since the accumulated debris layer consists of bone particles and water splashes, the ablation laser induces ablation on the surface at the center of the beam where we expect to have higher intensity. This was the reason behind employing the M-mode scanning system to capture images through the clean path on the sapphire window. The measured errors in the depth-controlled system indicate that the errors tend to increase with the depth of the laser-induced cut. These errors primarily arise due to the accumulation of water inside the laser-induced cut, which cannot be effectively removed by pressurized air. Consequently, depending on the volume of water accumulated inside the cut, either the surface of the cut or an incorrect depth of the cut (due to the refractive index of water) may be detected.

The miniaturized system has exhibited promising performance in creating laser-induced cuts on bone tissue. However, to facilitate its application in robot-assisted laser osteotomy, several enhancements are necessary. First, the designed 3D-printed housing lacked the degree of freedom required for aligning the mirror and dichroic filter. This can be achieved by designing a holder that can be adjusted with a screw. The implementation of a top-hat beam shaper holds the potential to improve the maximum energy coupled to the fiber. Furthermore, further exploration of the scanning methodology for both the OCT system and Er:YAG laser is warranted. This would enable the acquisition of two-dimensional OCT images and line cuts. Employing scanning for the Er:YAG laser can also offer the advantage of cleaning the protective window, thereby enhancing the quality of OCT images. By optimizing the system developed in this study, there is a promising outlook for its utilization in minimally invasive robot-assisted laser osteotomy. We foresee that with the advancement of high-energy custom-made optics and miniaturized holders, the size of the integrated system can be further reduced.