A promising approach in laser vitrectomy executed by plasma-mediated removal of vitreous body via a diode-pumped Q-switched Nd:YAG laser

To examine the applicability of plasma-mediated vitreous body removal, a diode-pumped Q-switched Neodymium:YAG laser was used for a possible application in eye surgery/vitrectomy. On a total of 1500 porcine vitreous bodies, removal rates were evaluated by comparing different LaserVit-tip designs (Mark I/II Gauge 19 and Mark III Gauge 22). The Nd:YAG laser, operating at a wavelength of 1064 nm and a pulse duration of 4 ns, was utilized for vitreous body removal with respective settings of 2, 3 and 4 mJ and pulse repetition rates (cut rates) from 5 to 25 Hz (300–1500 /min) in 5 Hz-steps as well as for 100 Hz (6000 cuts/min). The exposure times were selected at 10, 20, 40 and 60 s, respectively. Comparative measurements were carried out with mechanical cutters (Gauge 20 and Gauge 23), applying a fixed cut rate of 800 /min (13.33 Hz) at identical exposure times. The LaserVit-tips showed successful vitreous body removal for all laser settings and exposure times (Mark I: 6.2 g/min, Mark II: 8.2 g/min at 1500 cuts/min and 3 mJ, Mark II: 10.1 g/min, Mark III: 3.6 g/min at 6000 cuts/min at 3 mJ). Similar tip-dimensions (Gauge 22laser and Gauge 23cutter) showed comparable removal rates of 3.6 g/minlaser and 1.3 g/mincutter with settings of 6000 cuts/min at 3 mJ (laser) and 800 cuts/min for the mechanical cutter. A diode-pumped Q-switched Nd:YAG laser can successfully and gently remove vitreous body. The efficiency of the laser was comparable to that of mechanical cutters in terms of quantity of material removed per time unit.

Sample preparation. Vitreous body from freshly slaughtered pigs (porcine model: German Landrace swine, 8 month old, mass 30-40 kg) was used for laser vitrectomy. The vitreous bodies of the pigs were surgically extracted from every pig eye within 1 up to 4 h postmortem. A total of 1500 pig eyes from 750 animals (two specimens from each animal) were used. Experimental investigations were carried out directly after the eyes were present in the lab (1-4 h postmortem).
Laser source. A diode-pumped Q-switched Neodymium:YAG laser (λ = 1064 nm, A.R.C. Laser GmbH, Nuremberg, Germany) was used for this investigation. The given pulse duration was 4 ns. Pulse energies of 2, 3 and 4 mJ, respectively, were chosen for irradiation. Selected pulse repetition rates were 5 Hz up to 25 Hz at increments of 5 Hz, and an additional frequency of 100 Hz was applied. Laser light was transmitted to the handpiece ( Fig. 1) by using an adapted 320 µm-diameter quartz fiber (NA = 0. 22).
For aspiration experiments of the vitreous body, handpieces with newly designed titanium applicator tips (LaserVit-tip, A.R.C. Laser GmbH, Nuremberg, Germany) were utilized in a stepwise modification (Mark I -III) during the investigations ( Fig. 1 and Table 1). Aspiration was carried out with a Phaco machine (Megatron, Geuder AG, Heidelberg, Germany). The chosen settings were 550 mmHg for the vacuum, a flow rate of 50 ml/ min and ramp time of 0.5 s for the Venturi effect.
Experimental setup (laser-induced aspiration). For I -III) regarding aspiration, suction rates were recorded at 10, 20, 40 and 60 s, respectively. Furthermore, the mass difference was determined by using an electronic balance (SBS-LW-2000A, Steinberg Systems, Zielona Góra, Poland) before and after removal. Depending on the suction time, 4 up to 7 vitreous bodies were used per measurement (mass of one porcine vitreous was approx. ≈ 2.9 ± 0.4 g). After removal from the pig eye, the vitreous bodies were collected in a beaker. The handpieces were positioned approximately 5 mm above the bottom of the beaker.
Experimental setup (mechanical-induced aspiration). The same aforementioned experimental conditions regarding the vitreous bodies were retained for standard guillotine driven mechanical-induced "cutter" removal. For this purpose, a Phaco machine (Megatron, Geuder AG, Heidelberg, Germany) was used. The chosen  Table 1. www.nature.com/scientificreports/ settings were adapted to the laser-induced aspiration, 550 mmHg for the vacuum, a flow rate of 50 ml/min and a ramp time of 0.5 s for the Venturi effect. The mechanical handpiece sizes were Gauge 20 and Gauge 23, the cut rate was 800/min (13.33 Hz) and the measuring times were 10, 20, 30, 40 and 60 s, respectively. The same aspiration experiments were carried out with and without laser activity using pure water as a reference material.
Statistical evaluation. Descriptive evaluation, including calculation of the mean values, standard deviations and error estimations were applied to show the principle functioning of the LaserVit-tips. The spreadsheet calculation program OriginPro 8G (OriginLab Corporation, Northampton, MA, USA) was used for data evaluation of the vitreous mass removal. Graphical evaluation of the results was carried out as well with OriginPro 8G. Exponential and linear fit functions were applied in Figs. 2 and 3 as well as in Figs. 5, 6 and 7 for visualizing the data behaviors (trend curves). An explorative statistical analysis was not yet appropriate.

Results
Mark I handpiece. By means of the first aspiration tests, the framework conditions were investigated whether this method principally meets the requirements for vitreous body removal. The pulse energies used here were 4 mJ und 3 mJ with cut rates ranging from 300 to 1500 cuts/min (corresponding to pulse repetition rates of 5-25 Hz). Figure 2a,b show the attained removal rates for pulse energies of 4 mJ and 3 mJ, respectively.
The increase in the aspiration rates correlates with the increase in the pulse energy, the pulse repetition rate and the irradiation time. The maximally attained aspiration rates at 1500 cuts/min (25 Hz und 60 s) were: 13.3 g/ min for E 4 mJ , and 6.2 g/min for E 3 mJ . For cut rates of ≥ 600 cuts/min (10 Hz), a decrease in the slope of the curves can be observed for both pulse energies, as well as a saturation behavior at a pulse energy of 3 mJ. Due to the enormous fluctuation margin of the determined aspiration rates, the standard deviation values are reported for pulse energies of 4 mJ and 3 mJ, exemplary for the 25 Hz measurements. The standard deviation values for all other measurement values can be represented in the same order of magnitude.
By directly comparing Fig. 2a,b, the difference between both pulse energies as well as between the reference liquid water can be seen (Fig. 3). Here, water shows (independently of whether with or without laser activity) a linear increase over the entire timeline of 60 s with attained aspiration rates of 95.4 g/min.
Furthermore, Fig. 2a,b show the distal end of the LaserVit-tips for the respective maximum cut rate after ending the measurement. At irradiation times of ≥ 40 s, there was a definite abrasion for both pulse energies observed at the absorber plate material in the application tip. Not only was this material destruction associated with increasing irradiation time, but also with increasing laser frequency (Fig. 4). Depending on the exposure time, the penetration through the absorber plate begins already after 20 s at a pulse energy of 4 mJ. At a pulse energy of 3 mJ, however, this effect appears later (> 20 s). Depending on the frequency, at a fixed pulse energy (3 mJ) and an irradiation time of 40 s, a material abrasion is shown from the very beginning and increases with the cut rate and, thus, with the pulse repetition rate.
The destructive processes, that testifies an indirect proof of plasma formation occurring in the LaserVit-tips, highlight the necessity to modify the stability, which was technically implemented in the Mark II handpiece by using a thicker material layer in the absorber plate (Fig. 1).
Furthermore, a comparative measurement was undertaken for estimating the status quo of the laser-induced aspiration in relation to the clinically applied cutters. Figure 5 illustrates the obtained aspiration rates of the mechanical system with two different applicators (Gauge 20 and Gauge 23), in comparison with the laser-induced aspiration at a pulse energy of 3 mJ using a Mark I-LaserVit-type (Fig. 2b).
Similar to the laser-induced aspirations in the lower cut-region, the aspiration rates of the mechanical cutter showed a linear increase with time. Due to the low aspiration rates of 1.3 g/min (Gauge 23) < 1.6 g/min (Gauge 20) < < E 3 mJ of 6.2 g/min (Gauge 19) at 900 cuts/min (15 Hz), measurements with a pulse energy of 4 mJ were waived for further studies with the Mark II handpiece.
Mark II handpiece. In contrast to the Mark I LaserVit-tips, the Mark II LaserVit-tips have a thicker material layer (0.3 mm) in the absorber plate (Table 1 and Fig. 1). The effect is clearly shown by the decrease in the slope, i.e. by an increase in the aspiration rate from 6.3 to 8.2 g/min at a pulse energy of 3 mJ and a cut rate of 1500 cuts/ min (25 Hz). The increase in the cut rate to 6000 cuts/min (100 Hz) rate revealed an aspiration rate of 10.1 g/min. With this improvement, for the first time a linear increase in the aspiration rates could be shown at high pulse repetition rates of 1500 cuts/min and 6000 cuts/min up to 40 s irradiation time (Fig. 6).
With potential clinical application in mind, the pulse energy was reduced to 2 mJ. This step reduced the aspiration rate to 2.6 g/min (Fig. 7) which, however, still exceeded the aspiration rates of the mechanical cutter (Fig. 5).
Mark III handpiece. By reconsidering the clinical application, the diameter of the handpiece (Mark III) was reduced to Gauge 22 (Table 1 and Fig. 1). At a pulse energy of 3 mJ and a cut rate of 6000 cuts/min (100 Hz), the maximum aspiration rate of 3.6 g/min was reached (Fig. 7). Until now, no clinically relevant aspiration rates could be obtained for the Mark III handpiece by reducing once again the pulse energy to 2 mJ.

Discussion
Method (choice of Q-switched Nd:YAG). Different methods were carried out for vitreous body removal.
Regardless of the different techniques used, the material destruction of vitreous body is required because of its viscosity properties and anatomical structure. The underlying problem, however, is the fact that the employed handpieces need to have sizes of Gauge ≥ 22 (∅ ≤ 0.76) in order to minimize surgical side effects. Therefore, Scientific Reports | (2020) 10:21710 | https://doi.org/10.1038/s41598-020-78878-y www.nature.com/scientificreports/ With regard to the technical implementation of the vitreous body processing in this current work, a 2-step interaction could be employed as the cutting process. Due to the use of a quartz fiber for transmitting light into the ophthalmological handpiece, and because of its size, no additional internal focusing optics could be utilized here. The combination of the design of the LaserVit-tips (side aspiration port) with the positioning of the optical fiber at a distance of 1 mm up to 2 mm from the front of the titanium absorber, generated a primary interaction with an accompanying plasma-and shock wave propagation (1st Step). Based on the fact that the refractive indices of fused silica and vitreous humor are approximated (n fused silica = 1.449@1.083 µm, n vitreous humor = 1.341@650 nm) 26,27 , the generated intensity of the emitted Nd:YAG laser light of 3.1⋅10 8 W cm −2 is in the order of magnitude of the titanium absorber threshold 28 . Plasma-and shock waves in this technique tear the vitreous body inside the LaserVIT-tip so that the vitreous body can be aspirated (2nd Step).

Vitreous body.
The vitreous body is a clear transparent hydrogel in various species and consists, more or less, of identical extracellular matrix components. It contains a very high amount of water (98-99%) and traces of hyaluronic acid, glucose, ions and collagen [(11-16)⋅10 −3 %) [29][30][31] . Hyaluronic acid and collagen are the major components of the fiber matrix and responsible for the gel-like structure of the vitreous body; these components are, however, not homogeneously distributed 32 . The vitreous body is classified as a viscoelastic material and exhibits non-Newtonian rheological properties [32][33][34] , stemming from the varying concentrations of hyaluronic acid and collagen in the anterior, central and posterior regions 35 . Moreover, differences in the matrix composition mentioned above are responsible for the particular rheologic state of the vitreous in different species 36 .
Despite the fact that the vitreous body exhibits rheological properties like a non-Newtonian fluid [23][24][25] , values of the viscosity are not clearly defined in the literature. Some data are given as dynamic viscosity adopted to Newton's law as a Newtonian fluid, given in Eq. (1): www.nature.com/scientificreports/ whereby η denotes the dynamic viscosity, ν the kinematic viscosity and ρ the density of the substance. Other data are given as apparent viscosities adopted to non-Newtonian fluids. In that case, the viscosity η depends additionally on the shear rate δ or the shear stress σ 34 . This involves a spatial separation of the viscosity measurements within the anterior, central and posterior regions of the vitreous body 32,37,39 . Therefore, the inhomogeneous viscosity could explain the erratic values (see error bars) shown in Fig. 2a,b at 1500 cuts/min (25 Hz). Compared to the dynamic viscosity of water (1.0087 mPa⋅s at 20 °C), the viscosity of porcine vitreous body is much higher ( Table 2). The resulting effects are reflected in the differences in the aspiration rates shown in Fig. 3. Regarding the clinical applications, the viscosity of human vitreous body is approximately one order of magnitude lower than porcine vitreous body. Therefore, higher aspiration rates can be expected along with smaller handpieces (Gauge 22) and lower pulse energies (2 mJ).
LaserVit-tips-handpiece design. The objective of further developments of the LaserVit-tips is to implement a standardized application size of Gauge 23 for clinical use. Nevertheless, it is not only the size but also the lifetime and aspiration rate, i.e. the mass of vitreous body removed per time unit, that are decisive for a gentle, relatively non-invasive therapy, that likewise saves time. Based on the complex viscosity properties of the vitreous body, the ratio of the inner diameter of the LaserVit-tip (Gauge 23: ∅ outer = 0.68 mm, ∅ inner = 0.4 mm customized ) to the optical fiber diameter (∅ outer ≤ 0.28 mm) is particularly important for a gentle and simultaneously timesaving therapy. In order to create suitable aspiration conditions, the ratio ∅ inner, tip /∅ outer, fiber should lie in the magnitude of ≥ 2:1 42 .
However, reducing the optical fiber diameter causes higher intensities on the absorber plate which, in turn, reduces the lifetime of the LaserVit-tip. This can be compensated by a reduction in the pulse energy of the laser. Ideally, the extent to which these ideas can mesh with one another is the aim of future investigations. and frequency-dependence [3 mJ and 40 s in a range of 600-1500 cuts/min (10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)]. The interaction between the absorber plate and laser radiation (plasma formation) is needed for vitreous body cutting to ensure aspiration. The destruction of the absorber plates affected the cutting process, resulting in a reduced slope of the curve at ≥ 600 cuts/min (10 Hz) and therefore in a reduction of the aspiration rate (Fig. 2a,b

Conclusions
Within the scope of this study, it was possible for the first time to technically realize the efficient removal of vitreous body using a diode-pumped Q-switched Nd:YAG laser. The plasma-mediated laser vitrectomy represents a new and promising approach in this regard. The results show that laser-induced aspiration rates comparable to those of mechanical cutters could be obtained. However, for clinical applications and use, further investigations are necessary with respect to handpiece size, tip-design as well as optimized laser parameters. Nevertheless, a  www.nature.com/scientificreports/ transfer of this laser technology into the area of clinical ophthalmology is feasible. Based on these promising results, this technology is approved for testing in animal experiments on rabbit eyes. We strongly believe that a new branch has opend up for vitrectomy. Modifications of the LaserVit-tips enable an additional illumination and a variability in geometry. With a clinical application in mind even a one-handed vitrectomy is conceivable.
Received: 18 July 2020; Accepted: 23 November 2020 Figure 7. Mark II and Mark III handpieces at pulse energies of 2 mJ and 3 mJ at 6000 cuts/min (100 Hz) for all measurements. Comparable results were achieved for Mark II at 2 mJ and Mark III at 3 mJ. Table 2. Viscosities of human and porcine vitreous body from the literature. *Apparent viscosity: The vitreous body exhibits non-Newtonian rheological properties based on spatially varying concentrations of hyaluronic acid and collagen 32,34,37 . **Dynamic viscosity: The vitreous body exhibits rheological properties as a Newtonian fluid adopted to Newtons law 34,40 . # Values from 40 describe the viscosity of "aspirated" human vitreous body (Reference by Shafer 41