In Vitro Investigation of Stone Ablation Efficiency, Char Formation, Spark Generation, and Damage Mechanism Produced by Thulium Fiber Laser

Abstract

To investigate stone ablation characteristics of thulium fiber laser (TFL), BegoStone phantoms were spot-treated in water at various fiber tip-to-stone standoff distances (SDs, 0.5 ~ 2 mm) over a broad range of pulse energy (E_p, 0.2 ~ 2 J), frequency (F, 5 ~ 150 Hz), and power (P, 10 ~ 30 W) settings. In general, the ablation speed (mm³/s) in BegoStone decreased with SD and increased with E_p, reaching a peak around 0.8 ~ 1.0 J. Additional experiments with calcium phosphate (CaP), uric acid (UA), and calcium oxalate monohydrate (COM) stones were conducted under two distinctly different settings: 0.2 J/100 Hz and 0.8 J/12 Hz. The concomitant bubble dynamics, spark generation and pressure transients were analyzed. Higher ablation speeds were consistently produced at 0.8 J/12 Hz than at 0.2 J/100 Hz, with CaP stones most difficult yet COM and UA stones easier to ablate. Charring was mostly observed in CaP stones at 0.2 J/100 Hz, accompanied by strong spark-generation, explosive combustion, and diminished pressure transients, but not at 0.8 J/12 Hz. By treating stones in parallel fiber orientation and leveraging the proximity effect of a ureteroscope, the contribution of bubble collapse to stone ablation was found to be substantial (16% ~ 59%) at 0.8 J/12 Hz, but not at 0.2 J/100 Hz. Overall, TFL ablation efficiency is significantly better at high E_p/low F setting, attributable to increased cavitation damage with less char formation.

Introduction

In 2020, Olympus launched the Soltive SuperPulsed thulium fiber laser (TFL) [1], which has since then generated significant clinical enthusiasm while propelling technology advances in laser lithotripsy (LL) [2, 3]. Compared to Holmium (Ho):YAG laser (2.1 μm in wavelength), the TFL operates at a different wavelength (1.94 μm) with higher optical absorption in water, lower stone damage threshold, and smaller retropulsion effect [2, 4–6]. Importantly, while the clinical optimal settings for the Ho:YAG laser have been well defined, i.e., high pulse energy (E_p = 0.8 ~ 1.2 J)/low frequency (F = 4 ~ 10 Hz) for fragmenting and low E_p (0.2 ~ 0.4 J)/high F (> 40 Hz) for dusting [7], such optimal settings have not been established for the TFL [2, 8]. In addition, although clinical experience in the past few years has demonstrated that TFL is excellent in stone dusting [3, 9], there is concern regarding the fragmenting power of the TFL, especially for treating impacted calcium phosphate stones in the ureter [10]. Furthermore, stone carbonization (or charring) produced by TFL is much higher than Ho:YAG laser [10, 11], which may compromise its treatment efficiency, prolong the procedure time, and, in some cases, even limit its effectiveness in ureteroscopy.

Because of the fundamentally different technologies employed, TFL has a rectangular pulse profile with limited peak power (500 W) yet much longer pulse duration (up to 12 ms), compared to the typical shark-fin shaped pulse profile of the Ho:YAG laser with high peak power (up to 20 kW) yet much shorter pulse duration (< 1 ms) [2]. These dissimilar pulse profiles and characteristics generate utterly different bubble dynamics, and thus will profoundly influence the energy transmission and stone damage mechanisms of the two lasers [12, 13]. Accordingly, the optimal settings of TFL should be distinctly different from those for the Ho:YAG laser – a critical issue that cannot be resolved simply by relying on clinical trials, in which the same settings are often employed for side-by-side comparison [9, 14]. Therefore, a rigorous study should be conducted to comprehensively explore the stone ablation characteristics of the TFL and examine the associated mechanisms.

As a first step in this direction, we investigate stone ablation produced by the TFL using BegoStone phantoms over a broad range of E_p, F, and power (P = E_p · F) settings. Based on the results from this pilot study, two distinctly different E_p/F combinations (0.2 J/100 Hz and 0.8 J/12 Hz) are selected to compare stone ablation efficiency, bubble dynamics, pressure transients, char formation, and spark generation produced by the TFL at low vs. high E_p levels using human kidney stones of different compositions. Furthermore, the contribution of photothermal ablation vs. cavitation in stone damage produced by the TFL is assessed by using different fiber orientations and leveraging the proximity effect of a ureteroscope end to the fiber tip.

Materials and Methods

Stone sample preparation

BegoStone samples (5:2 powder-to-water ratio, BEGO^™ USA, Lincoln, RI) were prepared and pre-soaked in water for 24 hours before treatment [15]. Surgically removed large-sized uric acid (UA) and calcium phosphate (CaP) stones (50 ~ 60 mm), and calcium oxalate monohydrate (COM) stones (5 ~ 10 mm) were individually cold-mounted, bisected, and polished using 1200-grit sandpapers to obtain a flat surface (Fig. 1a) [16]. The chemical compositions of the human kidney stones were determined by infrared spectroscopy at the Mayo Clinic Laboratories (Rochester, MN).

Fig.1.

(a) Images of human kidney stones, including COM(100%), Uric Acid(80%)/CaP(20%), which is the core part circled out by red dashed line, and CaP(100%) after cold-mounting, bisecting, and polishing to prepare a flat and smooth surface before laser treatment. (b) Different laser power (P), pulse energy (E_p) and frequency combinations used for BegoStone experiments. The rectangular pulse profiles with various pulse durations at different E_p levels are plotted. The internal signal (V) is proportional to the output power of the laser pulse, with a peak power output at the fiber tip limited at 500 watts.

Laser treatment

The BegoStone samples were spot-treated in water using an IPG TFL system (TFL-50/500-QCW-AC, IPG Photonics, Oxford, MA) with 200 μm core diameter fibers (numerical aperture = 0.22, MED-Fibers, Chandler, AZ). A wide range of E_p (0.2 – 2.0 J) and F (5 – 150 Hz) were tested under 10 ~ 30 W power settings (Fig. 1b). Different pulse numbers (PN) were delivered at each spot for various standoff distances (SD = 0.5, 1.0 and 2.0 mm) between the fiber tip and stone surface, using perpendicular fiber orientation. All the pulses were generated at 100% peak power of the TFL to achieve better ablation efficiency [17]. The resultant stone damage was quantified by optical coherence tomography (OCT, OQ Labscope, Lumedica, Durham, NC) [15].

In addition, spot-treatments were performed on the COM, CaP, and UA samples using high E_p/low F (0.8 J/12 Hz) settings, in comparison with low E_p/high F (0.2 J/100 Hz) settings commonly used in clinical TFL lithotripsy [8].

Laser-bubble-stone interaction produced by TFL

During the TFL treatment, the laser-bubble-stone interaction was recorded by a high-speed camera (Phantom v7.3; Vision Research, Wayne, NJ) at 50,000 frames per second. Concomitantly, the acoustic pressure transients generated by the vapor bubble collapses were measured by a needle hydrophone (HNC-1000, Onda) placed at 45⁰ angle and 20 mm distance from the fiber tip [15, 18].

Photothermal ablation vs. cavitation damage produced by TFL

Taking advantage of the large UA and CaP stone samples, two additional experiments were performed to compare photothermal ablation vs. cavitation damage in TFL. First, the end of a flexible ureteroscope (Dornier AXIS^™, 3.6 F working channel, Munich, Germany) was placed near the fiber tip at 0.1 mm offset distance (OSD) to alter the dynamics of bubble collapse and mitigate cavitation damage [19]. Second, the stones were treated with parallel fiber orientation at SD = 0.5 mm to eliminate photothermal ablation while isolating the damage produced by cavitation [15, 18, 19].

Statistical Analysis

Student’s t-test was performed for statistical analysis, with a p-value below 0.05 considered as significant.

Results

BegoStone

Figure 2 summarizes the stone ablation speed in mm³/s produced by the TFL in BegoStone phantoms at different E_p and P settings under various SDs and treatment durations. The treatment efficiency was determined from the data and the associated crater volume vs. PN curve with a logarithmic fit under each test condition (see Supplementary Fig. S1). In general, the stone ablation speed varied with E_p and P at each SD, with the peak shifting, for example, from E_p = 0.4 J under 10 W to E_p = 1.0 J under 30 W at SD = 0.5 mm. Overall, the treatment efficiency decreased with increased SD from 0.5 mm to 2.0 mm, except for E_p = 1.0 ~ 2.0 J under P = 30 W. Moreover, for each combination of E_p, P and SD, the stone ablation speed decreased with treatment time (from 0.2 s to 1.0 s), reflecting the progressively saturated crater volume increase with PN (Fig. S1), except for E_p = 0.8 ~ 2.0 J under P = 10 W.

Fig.2.

Stone ablation speed (assessed by the ablated stone volume per second, mm³/s) produced by the TFL on the BegoStone phantoms at different pulse energy (E_p) and power (P) settings under different fiber tip-to-stone standoff distances (SDs = 0.5, 1.0 and 2.0 mm) and treatment durations (i.e., 0.2, 0.5 and 1.0 s)

It is worth noting that under thermally safe treatment conditions (i.e., P = 10 W) [20], the ablation speeds at 0.8 J/12 Hz and 1.0 J/10 Hz settings were found to be mostly higher than the other settings at SD ≥ 1 mm for 0.2 s treatment time. In contrast, under P = 20 ~ 30 W settings, the maximum ablation speeds could be achieved in the range of E_p = 0.8 J ~ 1.5 J with higher frequency (F = 16 ~ 40 Hz) regardless of SDs and treatment time, yet at increased risk of thermal injury [20].

CaP, UA, and COM Stones

For human stones, faster ablation speeds were consistently produced at high E_p/low F (i.e., 0.8 J/12 Hz) than low E_p/high F (i.e., 0.2 J/100 Hz) settings regardless of SD and stone composition (Fig. 3a). Quantitatively, by elevating E_p from 0.2 J to 0.8 J, the ablation speed could be significantly increased by 329 ~ 400% for CaP, 75 – 210% for UA, and 219 ~ 616% for COM stones (p < 0.01). The increased ablation speed could be primarily attributed to the significantly enlarged profile area, and, to a lesser extent, to the increased crater depth in most cases (see second and third rows in Fig. 3a).

Fig.3.

(a) Comparison of ablation speed, crater depth, and profile area produced by TFL at 0.2 J/100 Hz vs. 0.8 J/12 Hz settings after 1 s treatment on BegoStone, calcium phosphate (CaP), uric acid and calcium oxalate monohydrate (COM) stones. (b) Representative images of stone damage produced on BegoStone and human kidney stones under different laser settings. Char formations were mostly observed from CaP stones, and in one incidence, on UA stone surface treated at 0.2 J/100 Hz

At 0.2 J/100 Hz, the CaP stones exhibited the lowest ablation speeds at various SDs, followed by the COM and UA stones, with the corresponding values for the BegoStone comparable to the CaP stones. In contrast, at 0.8 J/12 Hz, the COM stones demonstrated the highest ablation speeds, followed by the UA and CaP stones. Interestingly, for the CaP stones char formation was observed in 8 out of the 15 damage craters (3/5 at SD = 0.2 mm, 3/5 at SD = 0.5 mm, and 2/5 at SD = 1.0 mm) produced at 0.2 J/100 Hz, but not at 0.8 J/12 Hz (Fig. 3b). The ablation speeds decreased substantially as SD increased from 0.5 mm to 1.0 mm (by 33% at 0.2 J/100 Hz and 28% at 0.8 J/12 Hz) and stayed relatively unchanged from SD = 1.0 mm to 2.0 mm under both settings.

For the UA stones, charring was only observed in one crater produced at SD = 0.5 mm. Yet, the drop in ablation speeds with SD was more significant and unceasing (i.e., 60% from SD = 0.5 mm to 1.0 mm and 22% from SD = 1 mm to 2 mm, at 0.2 J/100 Hz, and 27% and 45%, respectively, at 0.8 J/12 Hz). In comparison, no charring was observed in the COM stones, which showed a corresponding reduction in ablation speeds of 48% and 72% at 0.2 J/100 Hz, yet only 9% and 12 % at 0.8 J/12 Hz.

Bubble dynamics, pressure transients, char formation, and spark generation

To better understand char formation in the CaP stones, we compared features in the laser-bubble-stone interaction and associated pressure transients (Fig. 4–5). For the crater produced at 0.2 J/100 Hz with charring, two cycles of irregular and chaotic oscillations of an elongated vapor bubble were generated and collapsed in the first 4 pulses, resulting in pressure transients mostly below 5 bars (Fig. 4a). From the 5^th to 20^th pulse, strong spark generation was initiated that spread quickly in all directions, leading to explosive combustion (or material burning) over the irradiated area for about 400 μs. Thereafter, boiling bubbles emerged underneath the fiber tip and rose slowly without collapse. In the following few pulses, the combustion gradually died down with weak sparks of short duration (< 100 μs), and prolonged bubble collapse time (Fig. 4b). During this phase, the pressure spikes disappeared. After the 24^th pulse, only few dim sparks were observed near the stone surface, and the pressure transients generated by the bubble collapse reappeared (Fig. 4c–d).

Fig.4.

Representative high-speed imaging sequences of the bubble dynamics, spark generation, and explosive combustion (or material burning) produced by the TFL near the surface of calcium phosphate stones at 0.2 J/100 Hz under SD = 0.5 mm after different pulse numbers (PN): (a) 1 ~ 4, (b) 5 ~ 23, and (c) 24 ~ 100. The corresponding pressure transients measured during the 1 s treatment are shown at the bottom and plotted vs. (t – t_propagation), where t_propagation is the propagation time of the acoustic wave from the collapsing site to the hydrophone tip. The pink arrows point to the spark generation with stone material burning. (d) Histogram of the number and peak pressures generated by the bubble collapse during the three treatment phases.

Fig.5.

Representative high-speed imaging sequences of the bubble dynamics and spark generation produced by the TFL near the surface of calcium phosphate stones at 0.8 J/12 Hz under SD = 0.5 mm after different pulse numbers (PN): (a) 1, (b) 2 ~ 3, and (c) 4 ~ 12. The corresponding pressure transients measured during the 1 s treatment are shown at the bottom and plotted vs. (t – t_propagation), where t_propagation is the propagation time of the acoustic wave from the collapsing site to the hydrophone tip. (d) Histogram of the number and peak pressures generated by the bubble collapse during the three treatment phases.

For the crater produced at 0.8 J/12 Hz without charring, multiple cycles of irregular and chaotic oscillations of an elongated vapor bubble were produced in the first pulse with confined sparks (lifetime ≈ 200 μs) and weak combustion found during its late stage (Fig. 5a). Starting from the 2^nd pulse, four successive bubble expansions and collapses were consistently observed mostly without the presence of sparks (Fig. 5b–c). In addition, pressure transients with relatively high peaks exceeding 10 bars were detected after the 4^th pulse (Fig. 5d), suggesting elevated propensity for cavitation damage [18, 19].

Role of cavitation in stone damage produced by TFL

By mitigating bubble collapse, the crater volumes produced at 0.2 J/100 Hz were insignificantly affected for the BegoStone and CaP stones yet reduced by 33% ~ 45 % for the UA stones (Fig. 6a–c). In contrast, the crater volumes produced at 0.8 J/12 Hz were greatly reduced for BegoStone and all three kidney stone compositions, suggesting a significant contribution of bubble collapse to stone damage at high Ep. Moreover, stone damage using the parallel fiber setup was only observed at 0.8 J/12 Hz, initiated from the fiber tip and extending forward, presumably caused by the collapse of the vapor bubbles [18]. This finding confirms that cavitation damage can be produced by the TFL at high E_p.

Fig.6.

Stone damage produced by perpendicular fiber without vs. with a flexible scope placed at an offset distance (OSD) of 0.1 mm, and parallel fiber placed at SD = 0.5 mm on (a) BegoStone, (b) calcium phosphate (CaP), and (c) uric acid (UA) stones at 0.2 J/100 Hz and 0.8 J/12 Hz settings. *: p < 0.05; NS: no significance, i.e., p > 0.05.

Discussion

In this work, we first examined systematically the stone ablation speed produced by the TFL in BegoStone phantoms over a broad range of E_p, F, and P settings. Our results have demonstrated that under each power setting, the stone ablation speed will initially increase with E_p (or pulse duration since the peak power is limited at 500 W, see Fig. 1) to a maximum before declining thereafter (Fig. 2). At the same SD, the peak ablation speed may shift from low E_p (0.4 J) to high E_p (0.8 ~ 1.5 J) as the power or treatment duration increases. Furthermore, we have also observed that the contribution of bubble collapse to stone damage was mostly produced at high E_p/low F settings (e.g., 0.8 J/12 Hz, see Fig. 6), but not at the low E_p/high F settings (e.g., 0.2 J/100 Hz) often used clinically [8], except for UA stones. These findings suggest that the improved ablation speeds at these high E_p/low F settings may result from the combined effects of thermal ablation and cavitation damage. However, as E_p exceeds 1.5 J, such combined effects are likely to diminish (see Fig. 2), suggesting a potential limit in stone ablation at extremely high E_p with long pulse duration (> 2.6 ms). Further studies are warranted to examine the underlying mechanisms, including the possible influence of fiber tip degradation at high E_p [21].

We next compared the TFL ablation speed in kidney stones of various compositions and found that similar to the observations from BegoStone, the high E_p/low F (i.e., 0.8 J/12 Hz) can produce faster ablation speed than the low E_p/high F (i.e., 0.2 J/100 Hz) settings (Fig. 3). In particular, the ablation speed at 0.8 J/12 Hz is less sensitive to SD compared to 0.2 J/100 Hz, extending the effective treatment range of SD for TFL under high E_p (see Fig. 3a). This will allow TFL lithotripsy to be performed in non-contact mode, giving the urologist more freedom to scan the fiber tip over the stone surface during LL. Moreover, CaP stones exhibited the lowest ablation speeds under both settings presumably due to its extremely high melting temperature, compared to other stone compositions (Table 1). Interestingly, UA stones showed higher ablation speeds than COM stones at 0.2 J/100 Hz, which is consistent with previous studies utilizing lower E_p (0.035 – 0.07 J) [4, 22] and may be related to the low elastic moduli of UA stones [23] and their propensity to cavitation damage under such settings (see Fig. 6c). However, the ablation speeds of COM exceeding those of UA stones at 0.8 J/12 Hz is in distinct contrast with a previous study, in which a similar setting (1 J/10 Hz) was generated by a modulated continuous-wave TFL but with a much longer pulse duration (20 ms) [24]. Altogether, these results highlighted the need for further research to optimize TFL settings for different stone compositions.

Table 1.

Melting temperature, occurrence rate of char formation and spark-generation from 15 damage craters produced under 0.2 J/100 Hz and 0.8 J/12 Hz settings at SD = 0.5, 1.0 and 2.0 mm of calcium phosphate (CaP), uric acid (UA), calcium oxalate monohydrate (COM) stones [26–28].

	Melting Temperature	Char Formation		Spark Generation
		0.2 J/100 Hz	0.8 J/12 Hz	0.2 J/100 Hz	0.8 J/12 Hz
CaP	1670 °C	53%	0%	60%	60%
BegoStone	1449 °C	0%	0%	100%	100%
UA	360 °C	6.7%	0%	6.7%	40%
COM	206 °C	0%	0%	20%	100%

Importantly, CaP stones show higher propensity for stone carbonization (i.e., charring), accompanied by strong spark generation and resultant explosive combustion with material burning at 0.2 J/100 Hz than other stone types (Fig. 4, also see Table 1). This observation suggests that the low E_p levels with short pulse durations may not be sufficient to reach the high melting temperature of CaP stone for effective thermal ablation and material removal. Moreover, once the char is formed on the stone surface, it may shield the energy transmission of subsequent laser pulses, and ultimately, generate multiple boiling bubbles near the stone surface with diminished pressure transients as a result of significantly weakened bubble collapses. Conversely, sparks will be more frequently observed at high E_p in the absence of charring. These sparks could potentially result from either plasma formation [25] or the interaction between laser light and stone debris within the bubbles (see Fig. 5c). Further studies are warranted to better understand the mechanism of spark generation and char formation during TFL lithotripsy.

Regarding the limitations of the current study, the spot treatment in a water tank neglects the impacts of stone size, geometry, movement, fiber scanning, and tissue environment on the treatment efficiency. Future studies are warranted to conduct TFL treatment under clinically relevant conditions so that the various treatment outcomes, including fragment size and dust distribution [29] and thermal injury risk [30, 31] produced by different E_p/F combinations can be systematically investigated.

Conclusions

We have comprehensively examined the stone ablation speed of TFL in BegoStone phantoms, covering a broad range of E_p, F, and P. The results clearly demonstrate that high E_p/low F combinations (0.8 J/12 Hz) at 10 W power level (with minimal risk of thermal tissue injury) consistently outperformed the low E_p/high F settings (0.2 J/100 Hz) at 20 W commonly used in clinical LL, under a range of SD (0.5 to 2.0 mm). This observation was further confirmed for three types of human kidney stones. Notably, because of their high melting temperature, CaP stones possess high propensity for charring at 0.2 J/100 Hz, which, however, can be significantly mitigated at 0.8 J/12 Hz, despite the increased propensity for spark generation (especially for COM and UA stones). Moreover, this study confirms the significant contribution of bubble collapse to stone damage at high pulse energy levels. Further in vivo and clinical studies are warrantied to confirm the benefits of high E_p/low F combinations using TFL for urinary stone treatment.