1.1.

Current Ophthalmic Surgical Techniques

As of 2004, 4.1 million people in the United States suffered some form of diabetic retinopathy, and that number has steadily increased each year.¹ Preventive surgeries must be performed on these patients regularly to seal the leaking blood vessels on the retina. For this task, a KTP laser ($\lambda =532\mathrm{nm}$) and multimode optical fibers are used to create hundreds of therapeutic lesions by photocoagulation alongside a vitrectomy procedure to remove debris in the eye.² In advanced cases, like proliferative diabetic retinopathy (PDR), the debris inside the eye settles, forming encrusted layers as well as exudate deposits on the retina. As the density of these deposits increases, especially if they form on the macular region, vision becomes greatly impaired. The current surgical procedure for PDR requires the tedious use of micro-hook needles, forceps, and other instruments to remove deposits from the retina.³ Incomplete removal of these deposits is the primary reason for retinal re-detachment and surgical failure rates. These tedious techniques create prolonged stress on the surgeon’s steadiness and the patient’s eye, allowing for errors like accidental puncture, tearing, and detachment. For these reasons, the surgical procedure for PDR, advanced retinopathy of prematurity, and other similar intraocular surgeries could benefit from an ultraprecise contact laser probe for ablation of ophthalmic tissues. Conventional fiber-optic-based surgical laser probes are designed to operate either at a fixed working distance from the tissue or in contact with the tissue with relatively large spot diameters (e.g., $>100\mathrm{\mu m}$). These fibers and laser probes are not widely used in intraocular laser surgery, except for thermal coagulation, because they lose their ability to focus light in aqueous media.

In contrast, the novel laser probe discussed here may operate to ablate the tissue in contact mode. The highly focused beam at the probe’s surface combined with the short optical penetration depth of the erbium:YAG laser may allow better surgical results in more sensitive and thin areas like the macula, where detailed vision is achieved and surgical precision is most important. Replacement of mechanical intraocular surgical instruments with a laser probe that has a much smoother tip with a tractionless method for tissue engagement may also reduce mechanical retinal surface damage as well as retinal detachments.⁴

1.2.

Erbium:YAG Laser for Tissue Ablation

The Er:YAG laser is currently used in the medical fields of dermatology,^5,6 dentistry,^7,8 and ophthalmology^9–15 because the 2940-nm wavelength closely matches a major water absorption peak in tissue [Fig. 1(a)], resulting in an optical penetration depth (OPD) of approximately 5 to 10 μm for soft tissues. The ablation threshold needed to achieve soft tissue removal using the Er:YAG laser is $\sim 1.6\mathrm{J}/{\mathrm{cm}}^2$,¹⁶ and thermal damage depths at the bottom of the ablation craters created in cornea tissue at high fluence are between 10 and 50 μm.¹⁷ In the case of PDR surgery, the Er:YAG laser’s short OPD will mean ablating the exudate deposits while minimizing thermal damage to the underlying retina, which is critical. Since most current applications of the Er:YAG laser are in dermatology and dentistry, where the tissue is readily accessible, a waveguide or fiber, although preferred, is not required to deliver the laser energy to the tissue. However, for endoscopic applications, like intraocular surgery, a specialized fiber optic delivery system is required because common silica-based fibers do not transmit beyond $\sim 2500\mathrm{nm}$. Germanium oxide mid-IR fibers and hollow waveguides have been developed and tested to overcomethis issue, and they provide flexibility and high transmission. Previous studies using the Er:YAG laser during vitrectomy procedures have been reported by D’Amico, Joseph, and Hoerauf.^13,18–23 These studies did not use contact probes but fibers and devices at varying distances from retinal and membrane tissue with spot sizes greater than 100 μm.

Fig. 1

(a) Water absorption data plotted across mid-IR wavelengths (peak absorption occurs at $\lambda =2940\mathrm{nm}$);²⁴ (b) Er:YAG normal spiking mode, 75-μs pulse profile.

1.3.

Surgical Laser Microprobes and Microspheres

Over the past 20 years various optical devices including spheres,²⁵ hemispheres, domes, cones, slanted shapes,^26,27 cylindrical gradient index (GRIN) lenses,²⁸ and tapered fibers have been used as tips for ophthalmic laser scalpels. However, these devices are typically designed to operate in noncontact mode in air at a fixed working distance from the tissue. Contact mode probes can be created by placing a single sphere with a high index of refraction, between 1.7 and 2.0, at the end of a waveguide or optical fiber. Spheres at these indices will focus an incident plane wave on the back surface to a focal point near the front surface. Fast focusing through the sphere also creates a fast divergence after the sphere. This property can be used to decrease the effective penetration depth in tissue by allowing only the light near the surface of the sphere to have an energy density sufficient to ablate tissue. Working in contact mode eliminates any intervening water layer which would otherwise strongly attenuate the mid-IR laser radiation.

Recently, chains of microspheres in odd multiples at these indices have been shown to reduce the spot size of the laser beam through filtering of nonperiodically focused modes (PFMs).^29,30 Even-numbered spheres recollimate, whereas odd-numbered spheres refocus. Sphere chains are known to reduce transmission as the chain length increases on account of Fresnel losses and mode filtering. However, by reducing the beam size our applications also require lower laser pulse energy requirements than other fiber delivery probe devices. Furthermore, the transmission loss of a spherical lens is significantly less than a pinhole with a diameter equal to a sphere’s beam waist. The purpose of this study is to test a mid-IR laser and novel fiber optic probe for precise ablation of a thin layer of tissue, in contact mode, with minimal collateral thermal damage to underlying tissue, for potential application in ophthalmic surgeries.

2.1.

Laser Source and Fiber Delivery

An Er:YAG laser (Model 1-2-3, Schwartz Electro-Optics, Orlando, FL) operating at a wavelength of 2940 nm, in normal-spiking mode with a pulse duration of 75 μs [Fig. 1(b)], and repetition rate of 5 Hz was used in these studies. Pulse energies of 0.02 to 1.0 mJ were used. A 2-m-long, 150-μm-core-diameter germanium oxide mid-IR optical fiber (Infrared Fiber Systems, Silver Spring, MD) was used to guide the laser energy to the scalpel. The fiber core has a refractive index of 1.84 with an attenuation of $0.7\mathrm{dB}/\mathrm{m}$ at a wavelength of 2940 nm.

2.2.

Microsphere Scalpel Design

During preliminary studies, three types of spheres were tested to determine their candidacy for use in the surgical probe: sapphire, borosilicate, and barium titanate glass. The borosilicate glass spheres ($n\approx 1.4$) did not focus close to the surface of the sphere because of their low index of refraction, and they also failed to provide surface focusing in longer chains. Barium titanate ($n\approx 1.8$) provided good surface focusing but suffered from very low transmission at $\lambda =2940\mathrm{nm}$ and also had a low damage threshold of $\sim 3\mathrm{mJ}$. For our experiments we chose to use sapphire spheres with diameters of 300 and 350 μm (Swiss Jewel Company, Philadelphia, PA). Sapphire provides low attenuation at $\lambda =2940\mathrm{nm}$ and has a refractive index of 1.71, to provide focusing near the front surface. Geometrical ray trace simulations have shown that chains of spheres with a refractive index close to 1.7 provide the smallest spatial beam width at the surface of the end sphere.³¹ Sapphire is also biocompatible, has a high phase transition temperature, and a high laser damage threshold. The spheres used had tight diameter tolerances and a low number of inclusions. No damage to the sapphire spheres was noticed at pulse energies up to 10 mJ.

The structure of the scalpel was a semi-rigid and robust silver-lined $300\pm 20\mathrm{\mu m}$ inner diameter, $750\pm 25\mathrm{\mu m}$ outer diameter, silica hollow waveguide (HWG) (HWEA300750, Polymicro Technologies, Phoenix, AR). Current mechanical instruments used in ophthalmic surgery have overall diameters of $\sim 1\mathrm{mm}$, so our design can be easily integrated into existing surgical techniques. In these studies we used 3-cm HWGs in a straight configuration with a maximum loss of $2.0\mathrm{dB}/\mathrm{m}$ at $\lambda =2940\mathrm{nm}$. Since the HWG is the support structure of our probe, a length of 3 cm was determined to be the minimum length necessary to reach the retina from outside the eye, as illustrated in Fig. 2. To be a multipurpose ophthalmic scalpel used for surgeries throughout the eye, all corneal and retinal testing as well as ray trace modeling was performed under this required design constraint.

Fig. 2

Proposed ophthalmologic surgical configurations consisting of a 30-mm-long hollow waveguide of 0.75-mm diameter with microsphere focusing at distal tip. The diameter of a typical human eye is 24 mm.³²

These waveguides provide an optical conduit, allowing insertion of the germanium oxide fiber on the proximal end and gluing of a sphere or sphere chain on the distal end. The end of the scalpel must be sealed to prevent liquid leaking into the waveguide and altering the focusing properties of the sphere chain. This was achieved by placing a slightly larger diameter end sphere of 350 μm, compared to the inner diameter of the HWG (300 μm), on the tip of the HWG and then gluing it in place with UV-cured epoxy (Norland Optical Adhesive #61, Cranbury, NJ) using micromanipulation and a small wire, as shown in Fig. 3. Additional 300-μm spheres were then inserted into the proximal end and pushed into contact with the glued sphere to create multiple-sphere structures. Our experimental setup was in a vertical orientation so the spheres inside the HWG remained in contact. Future work will be necessary to provide a more robust and permanent method of fixing the internal spheres without compromising the optical path.

Fig. 3

Sealing method of the probe: gluing a 350-μm sphere to the tip of a 300-μm-inner-diameter HWG with UV-cured epoxy.

2.3.

Ray Tracing Simulations

Simulations of the multimode fiber output, HWG, and three different sphere configurations were modeled using Zemax (Radiant Zemax LLC, Redmond, WA). Figure 4(a) shows the light source modeled as the germanium oxide fiber’s 150-μm core with a full cone angle of 12 deg. For all presented simulations a total count of 20 million rays was used. To approximate a Gaussian distribution in the far field, a weighting factor (WF) was introduced for the irradiance of each ray, depending on its starting polar angle $(\theta ):\mathrm{WF}(\theta )=\exp (-2{\theta }^2/{\alpha }^2)$, where $\alpha =6\mathrm{deg}$. The calculated average WF was applied in ten equally spaced angular steps from 0 to 6 deg along the cone of directions centered with the normal to the fiber surface. The rays were randomly distributed in azimuthal directions for each point source to take into account skew modes. The source was placed at the proximal opening of a 3-cm-long, 300-μm-ID HWG lined with 100% reflective walls. At the distal tip three different $n=1.71$ sphere configurations were modeled in air, as shown in Fig. 4(b). The analysis plane was placed at the minimal beam waist beyond the end sphere corresponding to our beam characterization measurements in air.

Fig. 4

(a) Source: representation of the 150-μm-core germanium oxide fiber tip, which is inserted into the proximal end of the 3-cm-long HWG. (b) Beam Shaping: one-, three-, and five-sphere configurations consisting of 300-μm and 350-μm spheres with refractive indices of 1.71.

2.4.

Experimental Tissue Setup

Fresh porcine eyes were harvested and used immediately after sacrifice of pigs from unrelated experiments. The cornea was chosen in this preliminary study as a simple model for ophthalmic tissue ablation experiments for several reasons, including its ease of use concerning dissection, preservation, processing, and analysis after laser tissue ablation. Preliminary test results performed on retinal tissue are not shown because our current capabilities hinder us from recording measurements on retinal craters. This was due to the fact that the microscopic ablation craters were difficult to distinguish visibly, and our histological fixation methods induced a large amount of stress on the fragile retina, therefore destroying the samples.

The Er:YAG laser beam was coupled directly into a 2-m-long, 150-μm-core germanium oxide fiber. The fiber’s distal end was then inserted 1 mm into the proximal end of the vertically positioned 3-cm-long HWG with the glued 350-μm sphere on the bottom. Tissue ablation experiments were then performed by bringing the cornea tissue into gentle contact with the various probe configurations and translating the sample, ex vivo. The cornea was kept hydrated prior to each test by applying a single drop of Tetrahydrozoline HCl 0.05%, a common over-the-counter eye drop solution. The relatively low laser pulse rate of 5 Hz allowed creation of hundreds of single-pulse craters along this path over a period of 60 s.

3.1.

Microsphere Probe Beam Characterization

Characterization of the probe’s beam-shaping performance was conducted before tissue studies. One-, three-, and five-sphere structures were tested to determine their relationship to spot size and transmission. Figure 5(a) and the top row of Fig. 6 show the simulation results for intensity distribution at the minimum spot size after the end sphere, $\sim 10\mathrm{\mu m}$ from the vertex. The full width at half maximum (FWHM) measured from the simulation for each configuration of one, three, and five spheres is 8, 9, and 4 μm, respectively. For comparison, this was also measured experimentally using a 12.7-mm-focal-length calcium fluoride lens to magnify and image the minimum spot size near the sphere’s surface. An infrared beam profiler (Pyrocam III, Spiricon, Logan, UT) was positioned 50 cm from the magnification lens. To calibrate the magnification an empty HWG was used to reference 300 μm. Since the output of the multimode fiber has an unpredictable mode orientation depending on its length, bends, and coupling, the images captured by the beam profiler were averaged while gently flexing the fiber near its mid-length point. This motion generated random mode orientation patterns hitting the sphere chains, thus creating repeatable profile measurements. The averaged FWHM beams measured for one, three, and five spheres were 67, 32, and 30 μm, respectively, with representative beam profiles shown in Fig. 5(b).

Fig. 5

(a) Normalized simulation of the intensity measured at the minimum beam waist (FWHM) for one-, three-, and five-sphere configurations. (b) Three-dimensional beam profiles acquired by magnifying the minimum beam waist onto the IR beam profiler’s detector array. As the number of spheres increase, a reduction in FWHM spot diameter and an expanding low intensity mode-filtering ring can be seen in both simulation and experiment.

Fig. 6

(top row) Simulated intensity profiles for one-, three-, and five-sphere probes at minimum spot size occurring at 10 μm after sphere vertex. (middle row) Representative bird’s eye corneal crater images for one-, three-, and five-sphere probes at 0.1 mJ. (bottom row) Representative histology images for one-, three-, and five-sphere probe craters at 0.1 mJ with trichrome stain.

3.2.

Ablation Crater Measurements on Cornea Tissue

Various pulse energies between 0.02 and 1.0 mJ at a duration of 75 μs were tested to experimentally determine the ablation threshold of the cornea. With an average FWHM beam diameter of $\sim 50\mathrm{\mu m}$, craters began to form at a pulse energy of 0.1 mJ corresponding to an energy density of $\sim 5\mathrm{J}/{\mathrm{cm}}^2$. Single-pulse craters were formed on hydrated cornea tissue using this near-threshold pulse energy of 0.1 mJ. Pulse energy was measured with a pyroelectric detector (ED-100A, Gentec, Saint-Foy, Canada) near the tip of the probe. Single-, three-, and five-sphere structures were tested, and the tissue was imaged and measured with an FS70 Mitutoyo microscope equipped with a CCD camera (Edmund Optics, Barrington, NJ). Since each sphere chain configuration (one, three, or five spheres) attenuated the energy differently, attenuation of the laser output was adjusted to normalize each configuration to an incident energy of 0.1 mJ on the tissue surface. Over 50 craters for each structure were measured, and the average crater widths are shown in Table 1 along with representative images for each in Fig. 6.

Table 1

Corneal ablation crater dimensions at pulse energies of 0.1 mJ (S.D.=10  μm).

Histologic analysis was performed on the corneas to quantify ablation crater and thermal damage depths. The crater diameters were measured by light microscopy immediately after the procedure, and then the eyes were placed in formalin. Eyes were sectioned and the extra material was discarded, leaving only the polar cornea region. The corneas were trimmed for histological processing, and various samples were stained with either trichrome or hematoxylin-eosin. The craters were viewed and measured with the Mitutoyo microscope. Over 10 trichrome-stained histology craters for each structure were measured, and the average crater depths are shown in Table 1 along with representative images for each shown in Fig. 6.