Advances in femtosecond laser refractive eye surgery

Abstract

The advent of reliable short-pulsed lasers in the 1980s and 1990s made it possible to perform refractive laser surgery that offers correction without glasses. In laser-assisted in situ keratomileusis (LASIK), argon fluoride excimer laser ablation at 193 nm wavelength was employed to ablate central corneal stroma that had been exposed by cutting a flap using a mechanical microkeratome. Later, the mechanical keratome was replaced by a femtosecond (fs) laser that can produce a corneal dissection through applying a grid of near-IR laser pulses at  1030 nm with small energy at high repetition rate. This way, flaps with flat bed and side cut can be produced that enable precise repositioning of the flap after surgery. It enables to perform wave front-guided ablation for correction of higher-order aberrations besides myopia or hyperopia. In the 2010s, small incision lenticule extraction (SmILE) was introduced, in which a lenticule is dissected out of the central stroma using only fs laser pulses and removed through a small side cut. This technique requires only one laser and optical delivery system and involves less dissection of corneal nerves and biomechanical weakening of the cornea than LASIK. The thesis analyses the mechanisms of corneal dissection and explores novel dissection concepts for improving precision, efficacy and gentleness of refractive surgery both in LASIK and SmILE. Dissection relies on a sequence of laser-induced plasma formation, shock wave emission, and cavitation bubble dynamics. Pulse repetition rate, energy and spot separation influence the interaction of events from subsequent laser pulses and, thus, the cutting dynamics. The fundamentals of plasma formation and its dependence on laser pulse duration, wavelength, focusing angle and focus shape are described using generic equations for photo- and avalanche ionization, thermal ionization, and recombination. Laser-induced bubble formation and its interplay with shock wave emission is described using an extended Gilmore model of cavitation bubble dynamics. This analysis and previous experiments show that UVA sub-nanosecond laser pulses with stable temporal shape and a wavelength around 350 nm offer the potential of creating more precise cuts with more compact devices than conventional near-IR fs laser systems. Furthermore, the introduction of a helical phase plate can transform the Gaussian laser beam with elongated focus into a vortex beam with a ring shaped focus of similar length. Such focus shaping promises to improve the dissection efficacy in direction parallel to the corneal surface, which should reduce mechanical side effects and increase precision. The dynamics of corneal dissection by 330 fs, 1030 nm laser pulses was investigated using stroboscopic photography with sub-micrometer spatial resolution, high-speed photography with up to 50 million frames/s and videography at 1 kHz. The cavitation bubble size was determined from stroboscopic photographs through digital image evaluation. The morphology of the cuts was investigated histologically, and the quality of the cuts evaluated through scanning electron microscopy. It turned out that the plasma extends through several corneal lamellae and its rapid expansion leads to irregularly shaped lobular cavitation bubbles. Lobes from neighboring bubbles do not always meet, which leads to tissue bridges hindering flap lifting and lenticule extraction. The cutting process is largely an addition of individual disruption/cleavage effects, which depend mainly on pulse energy and spot separation. However, it has also a dynamic component resembling crack propagation during fracture in solids, which implicates a dependence on pulse repetition rate. The use of a vortex beam was explored for IR and UV wavelengths and found to improve dissection efficiency and quality. Scanning electron microscopy showed that the cuts are smoother than with a Gaussian beam. The force distribution during the plasma expansion from a ring focus facilitates tissue cleavage in the cutting direction parallel to the corneal surface and minimizes tissue bridges. Therefore, the volumetric energy density in the focal volume required for dissection is smaller with a vortex beam than with a Gaussian beam, which reduces plasma pressure and mechanical side effects. The smaller plasma energy density results also in less gas generation through free-electron-mediated molecular disintegration. This is advantageous because long-lived gas bubbles impair pupil tracking during clinical procedures and distort the corneal morphology during dissection of the upper lenticule cut in SmILE. Altogether, dissection using a vortex beam is more precise and gentler than with a Gaussian beam. Experiments with fs pulses at 343 nm wavelength and sub-ns pulses at 355 nm showed that the precision of corneal dissection could be further improved by reducing the laser wavelength. However, although the damage potential of the investigated UVA wavelengths lies four orders of magnitude below the peak value around 260 nm and although no corneal damage has been observed in previous animal experiments, the radiant exposure required for flap cutting lies above the permissible dose. The pulse duration dependence of laser pulse energy required for dissection with easy flap lifting was investigated for IR pulses between 480 fs and 8 ps duration and UV pulses between 1 ps and 400 ps duration. A strong pulse duration dependence is found even for the absorbed laser energy needed for flap cutting, which indicates that the efficiency of the cutting process drops. For durations L > 10 ps, stress and inertial confinement of energy deposition decrease, which reduces the plasma peak pressure and delays or even inhibits the formation of a shock front. For L < 3 ps, the reduced cutting energy is related to an increase of plasma energy density with decreasing pulse duration. These findings confirm the high suitability of pulse durations in the lower fs range down to about 150 fs, which are used in state-of the art clinical laser systems. The improvement of dissection through use of a vortex beam instead of a Gaussian beam is the finding of this thesis with greatest clinical relevance and potential for LASIK and SmILE. It has already been demonstrated that a helical phase plate can be easily implemented into the beam delivery system of existing clinical devices and is compatible with focus scanning. The next step is the conduction of a clinical study.