Open Access
Add to BibSonomy
Abstract
Retinal damage is a common intraoperative complication during vitrectomy, caused by a complex interplay between the suction of the vitrectome, the cut- and aspiration rate, and the distance of the instrument to the retina. To control this last factor, we developed two miniaturized fiber-optic distance sensors based on low-coherence interferometry for direct integration into the vitrectome. Both sensors have a diameter of 250 µm, which makes them compatible with a 25G vitrectome. The first sensor measures distance in the lateral direction. The second sensor is capable of simultaneously measuring distance in both the lateral and the axial direction. Axial and lateral directions correspond to the direction of the cutter port of the vitrectome and the direction along the vitrectome’s shaft, respectively. In both sensors, a free-form mirror deflects and focuses the beam in the lateral direction. In the dual-axis distance sensor, an additional lens is integrated into the free-form mirror for distance measurement in the axial direction. The beam-shaping micro-optics at the tip of the sensor fibers were fabricated through two-photon polymerization and are selectively gold coated for increased reflectivity of the mirror. Distance measurements were successfully demonstrated in artificial samples and in ex vivo pig eyes with a back-end that uses a current-tuned VCSEL as a swept-source. We experimentally demonstrate that the complete sensor system can attain a ${\rm SNR_{{\max}}}$ of up to 80 dB. The small dimensions of the developed sensors make them a potential solution for various other medical applications.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. INTRODUCTION
Age-related retinal diseases and disorders, such as retinal detachment, vitreous hemorrhage, macular hole, and epiretinal membranes, are becoming more prevalent, mostly due to the global increase in life expectancy [1,2]. Many of these vitreoretinal conditions require a surgical intervention including a vitrectomy before performing surgical maneuvers on the retina itself [3]. Even though generally considered safe, vitrectomy still poses a number of intraoperative risks including infections, bleeding, and retinal damage due to interaction with the vitrectome [4]. Among these, inducing iatrogenic retinal breaks is a severe complication that results from the interplay between the cut and aspiration rates used during vitrectomy and the relative distance between the vitrectome and the retina. Therefore, precise control of the position of the instrument with respect to the retinal surface is of utmost importance for a safe vitrectomy.
A 25G vitrectome, with a diameter of 0.51 mm, imposes strict size limitations on any integrated sensor. Fiber-optic low-coherence interferometry (LCI), where the measurement beam is delivered and collected through a standard optical fiber, is a promising approach for a miniaturized sensor. Additionally, LCI can provide distance measurements with micrometer accuracy over several millimeters. This approach has already been applied to various smart surgical instruments, including microinjectors, micro forceps, as well as vitrectomes for detecting retinal structures in front of the cutter port [5–8]. All of the probes developed so far can be classified into two categories: side-viewing and forward-viewing, allowing distance measurement in the lateral and axial directions, respectively. Forward-viewing probes are usually based on bare optical fibers [6], gradient-index (GRIN) fibers [7,8], GRIN lenses [9], or lensed fibers [10]. A common approach for side-viewing probes is the combination of GRIN lenses and mirrors [11–13]. As the technology for 3D nano-printing evolves, the latest approaches for miniaturized side-viewing probes utilize two-photon polymerization (2PP) to print a freeform mirror, simultaneously deflecting and focusing the beam [14,15]. However, since these probes are based on total internal reflection, they require the probe to be enclosed by a plastic catheter sheath to function in a medium with a refractive index similar to that of water, which increases the overall diameter of the probe [15]. Since the vitrectome can interact with the retina in both lateral and axial directions, a new approach is required for such a bidirectional sensor tip operating in an aqueous medium, as depicted in Fig. 1.
Fig. 1. Illustration of the vitrectomy procedure. The developed sensor is directly attached to the vitrectome and allows for distance measurement between the vitrectome and the retina in both the axial and lateral directions. This distance measurement can help to prevent retinal damage caused by the interplay between the suction force of the vitrectome and the distance to the retina.
Fig. 2. Schematic of the complete sensor system, which is a common-path low-coherence interferometer. 10% of the VCSEL emission is directed to a reference Mach-Zehnder interferometer that provides real-time frequency feedback. The remaining 90% is directed through a circulator towards the probe head. Light reflecting from the fiber-tip, and the light back-scattered by the retina are then directed back to the circulator, which channels this light towards the signal photodiode. In essence, this setup is analogous to a swept-source OCT system.
In this work, we present two miniaturized fiber-optic distance sensors that use 3D nano-printed catadioptric freeform micro-optics for multi-axis beam shaping and focusing. We first discuss a uniaxial, side-viewing fiber-optic probe based on previously published approaches [14,15]. Instead of relying on total internal reflection, the mirror surface has a metal coating, which enables operation inside an aqueous medium without an additional sheath surrounding the probe. To extend the sensor for measuring in both lateral and axial directions simultaneously, we combine our approach for the side-viewing probe with the approach of using a lensed fiber for forward-viewing measurements [10]. This novel catadioptric optical component splits the measurement beam into a lateral beam using the mirror, and an axial beam through an embedded microlens. The back-reflected light from both beams is collected with the same optical fiber and processed by a low-cost LCI engine that uses a current-tuned vertical-cavity-surface-emitting laser as a swept-source [16].
2. OPERATION PRINCIPLE AND SENSOR HEAD DESIGN
A. VCSEL-Based LCI
The operational principle of the common-path LCI scheme used in this work is depicted in Fig. 2. The common-path configuration forgoes a separate reference arm, eliminating the need for dispersion compensation or matching the sample arm length, as the reference beam is guided by the same optical fiber as the sample beam. The light source is a fiber-pigtailed VCSEL, which is tuned by self-heating through the adjustment of the drive current. The system operates at a center wavelength of 850 nm, chosen for its minimal absorption by the vitreous [17]. This “swept-source” LCI has a significantly longer measurement range compared to an SLED-based one, ultimately limited by the coherence length of the source [18]. The light output of the fiber-pigtailed VCSEL is directed through an optical circulator toward the sensor head comprising a 3D nano-printed free-form element accommodated at the fiber tip. A portion of the light field is reflected back at the interface between the fiber tip and the micro-optical element, providing the reference beam for the interferometer. The light field transmitted by the sensor head reaches the object and a portion of it couples back to the fiber. When it reaches the photodiode, both fields contribute coherently to the detected power, which depends on the distance $d$ of the object and the wavenumber of the laser. LCI estimates the distance by continuously sweeping the wavenumber of the laser and observing the homodyne interference at the photodetector. An additional reference Mach-Zehnder interferometer provides real-time frequency feedback necessary to register the recorded interferograms in the $k$-space. A detailed discussion on this system can be found in [9,16].
Fig. 3. Schematic depictions of the single-axis and the dual-axis distance sensors. (a) The single-axis distance sensor redirects and focuses the beam emerging from a single-mode fiber by means of a concave, gold-coated mirror into the lateral direction. (b) The dual-axis distance sensors focus the central part of the beam emerging from the fiber into the axial direction, while the remaining optical power is deflected and focused into the lateral direction. Scale bar is 100 µm.
B. Optical Design of the Sensor Head
The fiber-optic sensor uses a standard single-mode fiber (Thorlabs GmbH, P3-780Y-FC-5) with a cladding diameter of 125 µm. The 3D nano-printed freeform micro-optical elements at the fiber tip are responsible for redirecting and focusing the light at the desired distance, as depicted in Fig. 3. Through interviews with physicians regularly performing vitrectomy surgeries, we determined that a measurement range of 0 mm to 2 mm is considered appropriate. In this range, 0 mm indicates full contact between the vitrectomy tool and a potential obstruction. To meet this requirement, it is essential to ensure that the depth of focus covers the region from the probe head to a distance of at least 2 mm. Achieving this objective requires both beam expansion and focusing after the fiber [14,19]. Beam expansion is attained through the propagation within the probe head from the fiber tip to beam shaping surfaces. To maximize efficiency and minimize power loss, we set the distance between the fiber facet and the center of the optical element at 307.5 µm. With this geometry, 99.99% of the incident optical power is captured by the optical element. At a greater distance, the beam would expand more, resulting in significant portions of the beam, and thus the optical power, being truncated. Since we limit the size of the optical elements with the diameter of the fiber to keep the probe diameter as small as possible, this sets the upper limit for the distance between the fiber facet and the center of the optical element. A shorter distance between the fiber and the optical elements would still transmit more than 99.99% of the optical power, but it is less favorable as it would reduce the probe’s maximum working distance [14,19] for a given numerical aperture and Rayleigh range. Furthermore, placing the mirror as far as possible allows increasing its focal length, which in turn reduces its sensitivity to shrinking effects [20] and alignment tolerances.
1. Single-Axis Distance Sensor Head
To measure the distance in the lateral direction, the sensor fiber is terminated by a 3D nano-printed micro-optical element with an on-axis concave mirror, which simultaneously focuses and deflects the beam by 90°. Unlike previous approaches that relied on total internal reflection (TIR), the gold-coated mirror functions independently of the refractive index of the surrounding medium. The concave shape of the mirror is defined by the following equation:
Table 2. Simulated Beam Waist ${w_0}$, Rayleigh Range ${z_R}$, and Working Distance W.D. in Water and Air, Based on Skew Rays and Huygens’ Principlea
To reduce back-reflections that lead to unwanted signal peaks, the exit window has a tilt with respect to the incoming beam by 6°. This tilt reduces the reflection back to the core of the optical fiber by 96% according to a simulation in ZEMAX OpticStudio. At the same time, the deflection of the beam due to refraction at the tilted interface is kept below 1° in water.
2. Dual-Axis Distance Sensor Head
The dual-axis distance sensor (DADS) expands the capabilities of the single-axis distance sensor (SADS) by monitoring the distance not only laterally, but also axially. To split the beam into two perpendicular directions, the DADS head includes a microlens positioned at the center of the optical axis, which focuses the central portion of the incoming beam in the axial direction. The lens profile is also in the form described by Eq. (1). Corresponding beam waist ${w_0}$, Rayleigh range ${z_R}$, and working distance in water and air, obtained through the same optimization process as for the mirror, can be found in Table 2.
It should be noted that for the DADS, the signal coming from both measurement directions is collected through the same fiber and processed by the same LCI engine. Thus, they appear in the same depth scan after data processing, and distinguishing between the two peaks is in its simplest manifestation not possible. From an application perspective, however, the role of this sensor is to monitor the relative distance between the vitrectome-tip/cutter window to the retinal surface and provide the surgeon with a notification if one of the measured distances falls below a critical distance. A number of possible approaches that would enable unambiguous identification of signal directions are explored in Section 5.
The beam splitting ratio between the axial and lateral directions can be adjusted by adjusting the diameter of the lens, taking into account potential sensitivity to alignment tolerances between the fiber facet and the DADS head. These tolerances allow for a maximum tilt error of 4°. Such an error shifts the intensity profile of the beam from its original position at the center of the lens towards the mirror, changing the effective splitting ratio. We have chosen a splitting ratio of 65/35, with 65% of the light directed axially and 35% directed laterally. This choice ensures that even if the input fiber tilts by ${\pm}\;{4}^\circ$ during assembly, 50% of the light will still be directed axially. In contrast, a 50/50 split would result in only 25% of the light being directed axially. To achieve this ratio, the lens radius is set at 15.25 µm.
C. Mechanical Design
To align the beam shaping surfaces with the optical fiber, the SADS and the DADS heads feature a hollow channel terminated at the proximal end by an insertion funnel. Chosen mechanical tolerances allow a maximum tilt error of ${\pm}\;{4}^\circ$ between the fiber facet and the optical axis of the sensor head. A promising alternative to reducing this tolerance while significantly simplifying micro-assembly is to print the sensor head onto the fiber [24], which would also reduce the probe diameter from 250 µm to 125 µm.
To selectively coat the mirror surface but not the microlens, the latter is displaced by 20 µm in the axial direction, creating an overhanging edge on the mirror, which serves as a shadow mask that protects the lens from gold deposition during evaporation. To completely remove the uncured photopolymer after 3D nano-printing, the developer must be able to penetrate all parts of the structure, including all cavities and gaps [24]. Therefore, we integrated four rinsing windows into the sides of the sensor head.
3. FABRICATION AND ASSEMBLY
The probes are printed onto ITO-coated fused silica substrates in a dip-in configuration using a commercial 3D nano-printer (Nanoscribe GmbH, Photonic Professional Gt+) equipped with a ${25\times/0.8\rm NA}$ objective (0.8 DIC Imm Korr, Carl Zeiss AG) using IP-S resist (Nanoscribe GmbH). The laser power is adjusted to 50%/25 mW at a scanning speed of $50\;{\rm mm}\;{{\rm s}^{- 1}}$. For the optical components, we set a slicing distance of 0.1 µm and a hatching distance of 0.2 µm.
Fig. 4. Assembled DADS and gold evaporation process. (a) DADS gold coated and glued to optical fiber. (b) DADS before gold deposition, partly covered with polyimide film tape. (c) DADS after gold deposition. (d) DADS after gold deposition with polyimide tape removed. Scale bars are 200 µm.
Fig. 5. Deviation of manufactured SADS and DADS surface profiles from the design. (a) Mirror of the SADS. (b) Lens of the DADS. (c) Mirror of the DADS.
In contrast, the mechanical parts are printed with a slicing distance of 0.6 µm and a hatching distance of 0.5 µm to reduce the overall print time. Following printing, the probes undergo a rinsing process in propylene glycol monomethyl ether acetate (PGMEA) to eliminate any uncured resist. Subsequently, they are rinsed in isopropyl alcohol (IPA) to remove the PGMEA. Notably, a post-curing step with UV or heat is omitted since the probes are already irradiated with UV to cure the adhesive during assembly.
A 100 nm thick layer of gold is vapor-deposited onto the mirror surface following the deposition of a 2 nm thick chromium layer that is used as an adhesion promoter [25,26]. To prevent coating of the input facet, we cover the lower part of the probe with commercially available polyimide film tape during evaporation, as depicted in Fig. 4. After the metal coating, the probe heads are manually assembled on 2.5 m long fiber patches with cleaved tips, and secured in place with an index-matched adhesive (Panacol Vitralit UC 1618) cured for 120 s using a UV-LED. Figure 4 shows the fully assembled DADS.
4. CHARACTERIZATION
A. Surface Characterization
We measured the shape accuracy and surface roughness of the printed elements with a 3D optical profiler (ZYGO NewView 9000). The root-mean-square (RMS) roughness is measured within a square area of 25 µm on each side. Within this area, we applied a Gaussian high-pass filter with a period of 25 µm, following the guidelines outlined in EN ISO 25178:2012. Sample deviations from the ideal shape are shown in Fig. 5 for an area covering twice the beam radius ($T\gt 99.9\%$). In the case of the microlens, it was not possible to capture the full radius because optical access was limited by the edges around it. The two mirrors and the output window show a peak-to-valley (PTV) deviation of about 200 nm, with no systematic deviation evident. In the case of the lens, we observed a systematic, rotationally symmetric deviation with a PTV deviation of about 800 nm in all fabricated probes. The RMS surface roughness measurements showed a mean value of ${18}\;{\rm nm}\;{\pm}\;{5.1}\;{\rm nm}$ (${ n} = {10}$) for the SADS mirror, and ${16}\;{\rm nm}\;{\pm}\;{1.6}\;{\rm nm}$ (${n} = {12}$) for the DADS mirror. The DADS microlens had a mean RMS roughness of ${30}\;{\rm nm}\;{\pm}\;{7.8}\;{\rm nm}$ (${ n} = {12}$), while the exit window of both probes had ${21}\;{\rm nm}\;{\pm}\;{4.8}\;{\rm nm}$ (${n} = {10}$). Since there is no apparent rotational symmetry in the deviation of the mirrors of both the SADS and the DADS, these deviations are likely to introduce aberrations affecting the shape of the focal spot without significantly changing the working distance. In contrast, the rotational symmetry aberration in the lens will in addition to aberrations lead to a shift in the working distance.
The reported shape deviations of the fabricated optical components are caused by shrinkage of the structure during the polymerization and the subsequent development stage [24]. In principle, it is possible to substantially reduce these deviations by means of compensation steps [27]. However, since the optical performance in our case still meets the sensor requirements, this additional step was omitted. The measured roughness values are consistent with published results and do not affect optical performance in the intended application [24].
B. Optical Performance
We used an image sensor (IDS Imaging Development Systems GmbH, UI-1552LE-C) with a pixel size of 2.8 µm translated along the optical axis of the individual beams to measure 3D beam profiles. By fitting a Gaussian function to the measured profiles, we derived the beam radius $w$ ($1/{{\rm e}^2}$) as a function of distance along the optical axis, allowing us to determine the Rayleigh range ${z_R}$ and the working distance. The corresponding results are listed in Table 3.
Table 3. Mean Beam Waist ${w_0}$, Rayleigh Range ${z_R}$, and Working Distance with Corresponding Standard Deviations of the Meana
Fig. 6. ${\rm SNR_{{\max}}}$ measurements of the SADS and the DADS. (a) ${\rm SNR_{{\max}}}$ of five individual SADS. (b) ${\rm SNR_{{\max}}}$ of three DADS in lateral and axial directions.
Comparison of the SADS results with the simulated values presented in Table 2 reveals deviations. We observed variations in working distance between the ${x}$ and ${y}$ directions indicating astigmatism. The lateral beam of the DADS exhibits astigmatism as well, which is consistent with the simulations. Furthermore, the working distance of the DADS in the lateral direction is larger by approximately 0.3 mm, whereas the axial beam of the DADS shows a reduction in the working distance by approximately 0.4 mm compared to the simulation.
To explain the deviation between the simulated and measured working distances, we must consider the measured deviations in the actual shape of the optical component from the design. The asymmetric deviations observed in the focusing mirrors of both the SADS and DADS contribute to astigmatism and minor shifts in the working distances. The influence of the surface deviation becomes more prominent for the axial beam of the DADS. In this case, we observed a rotationally symmetric deviation with a peak-to-valley amplitude of approximately 800 nm. This deviation leads to an increased radius of curvature of the lens, ultimately diminishing the focusing power.
C. Accuracy and Resolution
The axial resolution of the LCI system is, analogous to an OCT system, defined by the light source, not by the focusing optics [28]. Based on the tuning range of the VCSEL ($\Delta {k_{\rm low {\text-} cost}} = 12\,\,{{\rm mm}^{- 1}}$), the theoretical resolution is calculated to be 231 µm. The actual resolution of about 285 µm in the biological specimen, shown in Fig. 7, is close to this theoretical value. Since the precise identification of the distance of an individual peak and not the differentiation of individual retinal layers is relevant for distance measurement, accuracy is more important than resolution. The accuracy of the low-cost OCT is improved by zero-padding [29], resulting in a theoretical accuracy of 77 µm.
D. Sensitivity and Measurement Range
To assess the sensitivity of the system, we measured the ${\rm SNR_{{\max}}}$ as outlined by Agrawal et al. [30], which gives the maximum signal-to-noise ratio (SNR) that can be achieved for a sample reflectivity of one. The ${{\rm SNR}_{{\max}}}$ is defined as
We conducted the measurements in water, as its refractive index closely matches that of human vitreous [23]. We mounted the probes on a translational stage and placed a borosilicate crown glass wedge with a thickness of 10 mm and an angle of 0.5° as the target. The large thickness combined with the tilt minimized the generation of multiple signals or internal reflections. The significantly lower reflectivity of the glass compared to a mirror ensured that the system was not saturated. We connected the sample to a kinematic mount using a mechanical extension for tilt correction and immersed it in a beaker filled with water. We then recorded the ${{\rm SNR}_{{\max}}}$ 200 times at various probe-target distances. The measurement range was determined based on the 12 dB intensity roll-off, following the criteria proposed by Cereda et al. [7].
Fig. 7. Demonstration of distance sensing in a biological sample. (a) Dual-axis distance sensor inserted into pig eye. The cornea was removed to ease the insertion of the sensor. (b) Sketch of the experimental setup. (c) Normalized depth scan demonstrating simultaneous distance measurement in two directions. The widths (FWHM) of the peaks indicate a resolution of about 285 µm.
The ${\rm SNR_{{\max}}}$ of both probes was measured in the range of 0 mm to 3 mm, defined from the relative probe surfaces. Figure 6 depicts the results. The SADS has a ${\rm SNR_{{\max}}}$ of 65 dB to 80 dB over the observed range, with an average of ${71.11}\;{\rm dB}\;{\pm}\;{4.98}\;{\rm dB}$ (${n} = {460}$). The highest ${{\rm SNR}_{{\max}}}$ is observed around 1 mm. In the lateral direction, the DADS ${\rm SNR_{{\max}}}$ ranges from 65 dB to 80 dB, averaging ${69.72}\;{\rm dB}\;{\pm}\;{5.43}\;{\rm dB}$ (${n} = {280}$), without a distinct maximum. For the axial direction of two DADSs, the ${\rm SNR_{{\max}}}$ ranges from 60 dB to 80 dB, with an average of ${65.15}\;{\rm dB}\;{\pm}\;{7.66}\;{\rm dB}$ (${n} = {280}$), and no distinct maximum. Errors correspond to the standard deviation of the mean. While the ${\rm SNR_{{\max}}}$ of each probe varies over distance with minor deviations, it can be observed that there is a more significant variation in terms of ${\rm SNR_{{\max}}}$ between individual probes. The SADS and DADS both have a measurement range exceeding 3 mm in the lateral direction. The mean measurement range in the axial direction for the DADS is ${2.45}\;{\rm mm}\;{\pm}\;{0.46}\;{\rm mm}$ (${n} = {3}$).
E. Demonstration of Distance Sensing in Biological Samples
Although ${\rm SNR_{{\max}}}$ is a good metric for comparison with other systems because of its reproducibility, the SNR in the actual application is highly relevant. To allow for the acquisition of the SNR in an environment as close to reality as possible, we characterized the probes in biological specimens (i.e., ex vivo pig eyes). The handheld probes were carefully inserted into the vitreous body and subjected to intermittent, non-continuous movements, which allowed the acquisition of data within the predetermined distance range. This procedure was performed after the cornea was removed. Figure 7 illustrates the experimental setup and shows an exemplary A-scan. The SNR in the range of 0 mm to 3 mm of the SADS was ${26.0}\;{\rm dB}\;{\pm}\;{4.0}\;{\rm dB}$ (${n} = {847}$) while the SNR of the DADS in lateral and axial directions is ${18.0}\;{\rm dB}\;{\pm}\;{3.3}\;{\rm dB}$ (${n} = {531}$) and ${20.0}\;{\rm dB}\;{\pm}\;{3.4}\;{\rm dB}$ (${n} = {931}$), respectively.
Fig. 8. Demonstration of distance sensing in the biological specimen using one DADS in the axial direction.
For all probes, the SNR decreases with increasing distance from the retina, as exemplarily shown in Fig. 8 for one SADS. In the range of 0 mm to 3 mm, the obtained SNR is in the range of 10 dB to 40 dB, depending on the probe and the measured distance. These results are comparable to published approaches where the SNR showed values of about 20 dB. [7]. The variation of the SNR at a given sensor position is due to the speckles, which translate to a large variation in light collection intensity. Comparing the results of the SADS with those of the DADS, the SADS showed a higher SNR, which is consistent with the ${\rm SNR_{{\max}}}$ measurements.
5. DISCUSSION
Individual probes of the same type exhibit significant ${\rm SNR_{{\max}}}$ differences. This can be attributed to limitations in adjusting the reference light intensity within the common-path SS-LCI system, as opposed to a configuration that utilizes a discrete reference arm. In a common-path LCI system, the intensity of the reference is determined by the Fresnel reflection at the reference specific interface. In this case, the interface is positioned between the optical fiber and the 3D nano-printed probe. The level of back-reflected light varies due to tolerances in passive alignment, which affects ${\rm SNR_{{\max}}}$. Furthermore, the assembly tolerances cause a variation in the collection efficiency due to a tilt of the optical components relative to the core of the optical fiber. The sensors could also be used in a system with a distinct reference arm. This would allow the reflectivity of the reference arm to be tuned, potentially increasing the ${\rm SNR_{{\max}}}$. However, this approach would require precise control of the lengths of both the reference and sample fibers. Additionally, the system would be more susceptible to drift, resulting in a deviation in the measured distance.
A current limitation of the DADS is its inability to distinguish between the two measured directions. An approach relying on data processing could function in tandem with an accelerometer connected to the handpiece of the instrument. By integrating data on the instrument’s movement and position in space, and the distances from the two measurement directions of the DADS, it would be possible to track the instrument’s position relative to the eye/retina. This virtual position would facilitate clear distance assignments, as the anatomy of the eye permits only specific combinations of the measured variables. A second, hardware-based solution could involve using a polarization-maintaining optical fiber in combination with a polarizer positioned at the exit window of the lateral measurement direction. In such a configuration, modulation of the polarization properties of the measurement light would only affect the lateral beam, while the axial beam would remain unaffected.
For the implementation of the system for use in vitrectomy, it should be noted that at the given resolution of the system, a distance peak in the depth scan does not necessarily indicate the distance to the boundary between the vitreous and the retina. Instead, the peak denotes the distance to the four hyperreflective outer retinal bands of the retina, where the backscattered intensity is largest [31]. This discrepancy must be taken into account in the final system by including an anatomical offset value.
6. CONCLUSION
The primary objective of this study was to design and implement a novel miniature catadioptric sensor attached to an optical fiber for distance measurement during vitrectomy. This goal was realized through the combination of 3D nano-printing by two-photon polymerization and gold evaporation, leading to a sensor head with refractive and reflective optical functionality. The resulting sensor, with a diameter of only 250 µm, demonstrated the ability to simultaneously measure distance in two directions within an aqueous medium, covering a range of approximately 2 mm. The feasibility and comparability of the results obtained with existing knowledge validate the effectiveness of the 3D nano-printed catadioptric sensor.
Funding
Horizon Europe Excellent Science-Marie Skłodowska-Curie Actions (860185).
Acknowledgment
We acknowledge the support from the European Union’s Horizon 2020 research and innovation program, which provided funding for Aybuke Calikoglu’s involvement in this study.
Disclosures
The authors declare no competing interests.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
REFERENCES
1. K. Christensen, G. Doblhammer, R. Rau, et al., “Ageing populations: the challenges ahead,” Lancet 374, 1196–1208 (2009). [CrossRef]
2. T. J. Wubben, N. Talwar, T. S. Blachley, et al., “Rates of vitrectomy among enrollees in a United States managed care network, 2001–2012,” Ophthalmology 123, 590–598 (2016). [CrossRef]
3. A. Chandra, W. Xing, M. R. Kadhim, et al., “Suprachoroidal hemorrhage in pars plana vitrectomy: risk factors and outcomes over 10 years,” Ophthalmology 121, 311–317 (2014). [CrossRef]
4. J. T. Thompson, “Advantages and limitations of small gauge vitrectomy,” Surv. Ophthalmol. 56, 162–172 (2011). [CrossRef]
5. J. U. Kang and G. W. Cheon, “Demonstration of subretinal injection using common-path swept source OCT guided microinjector,” Appl. Sci. 8, 1287 (2018). [CrossRef]
6. C. Song, D. Y. Park, P. L. Gehlbach, et al., “Fiber-optic OCT sensor guided “SMART” micro-forceps for microsurgery,” Biomed. Opt. Express 4, 1045–1050 (2013). [CrossRef]
7. M. G. Cereda, S. Parrulli, Y. Douven, et al., “Clinical evaluation of an instrument-integrated OCT-based distance sensor for robotic vitreoretinal surgery,” Ophthalmol. Sci. 1, 100085 (2021). [CrossRef]
8. A. Abid, R. Duval, F. Rezende, et al., “A smart vitrector equipped by a fiber-based OCT sensor mitigates intentional attempts at creating iatrogenic retinal breaks during vitrectomy in pigs,” Transl. Vis. Sci. Technol. 10, 19 (2021). [CrossRef]
9. S. Vilches, C. Ataman, and H. Zappe, “Interferometric fiber-based real-time distance sensor for endoscopic applications,” Proc. SPIE 10868, 1086819 (2019). [CrossRef]
10. Y. Qiu, Y. Wang, K. D. Belfield, et al., “Ultrathin lensed fiber-optic probe for optical coherence tomography,” Biomed. Opt. Express 7, 2154–2162 (2016). [CrossRef]
11. Y. Wu, J. Xi, L. Huo, et al., “Robust high-resolution fine OCT needle for side-viewing interstitial tissue imaging,” IEEE J. Sel. Top. Quantum Electron. 16, 863–869 (2010). [CrossRef]
12. D. Lorenser, X. Yang, R. W. Kirk, et al., “Ultrathin side-viewing needle probe for optical coherence tomography,” Opt. Lett. 36, 3894–3896 (2011). [CrossRef]
13. D. Lorenser, X. Yang, and D. D. Sampson, “Ultrathin fiber probes with extended depth of focus for optical coherence tomography,” Opt. Lett. 37, 1616–1618 (2012). [CrossRef]
14. J. Li, P. Fejes, D. Lorenser, et al., “Two-photon polymerisation 3D printed freeform micro-optics for optical coherence tomography fibre probes,” Sci. Rep. 8, 14789 (2018). [CrossRef]
15. J. Li, S. Thiele, B. C. Quirk, et al., “Ultrathin monolithic 3D printed optical coherence tomography endoscopy for preclinical and clinical use,” Light Sci. Appl. 9, 124 (2020). [CrossRef]
16. S. Vilches, H. Zappe, and C. Ataman, “Multi-point fiber-optic distance sensor for endoscopic surgery monitoring,” IEEE Photon. Technol. Lett. 35, 883–886 (2023). [CrossRef]
17. E. A. Boettener and J. R. Wolter, “Transmission of the ocular media,” Investig. Ophthalmol. Vis. Sci. 1, 776–783 (1962).
18. G. Barteselli, D.-U. Bartsch, R. N. Weinreb, et al., “Real-time full-depth visualization of posterior ocular structures: comparison between full depth imaging spectral domain OCT and swept source OCT,” Retina 36, 1153–1161 (2016). [CrossRef]
19. Y. Mao, S. Chang, S. Sherif, et al., “Graded-index fiber lens proposed for ultrasmall probes used in biomedical imaging,” Appl. Opt. 46, 5887–5894 (2007). [CrossRef]
20. Y. Liu, O. Stein, J. H. Campbell, et al., “Three-dimensional printing and deformation behavior of low-density target structures by two-photon polymerization,” Proc. SPIE 10354, 103541U (2017). [CrossRef]
21. P. D. Colbourne, “Generally astigmatic Gaussian beam representation and optimization using skew rays,” in Classical Optics (Optica, 2014).
22. M. Schmid, D. Ludescher, and H. Giessen, “Optical properties of photoresists for femtosecond 3D printing: refractive index, extinction, luminescence-dose dependence, aging, heat treatment and comparison between 1-photon and 2-photon exposure,” Opt. Mater. Express 9, 4564–4577 (2019). [CrossRef]
23. D. A. Atchison and G. Smith, “Chromatic dispersions of the ocular media of human eyes,” J. Opt. Soc. Am. A 22, 29–37 (2005). [CrossRef]
24. T. Gissibl, S. Thiele, A. Herkommer, et al., “Sub-micrometre accurate free-form optics by three-dimensional printing on single-mode fibres,” Nat. Commun. 7, 11763 (2016). [CrossRef]
25. S. Pradeep Narayanan and S. Raghavan, “Fabrication and characterization of gold-coated solid silicon microneedles with improved biocompatibility,” Int. J. Adv. Manuf. Technol. 104, 3327–3333 (2018). [CrossRef]
26. A. D. Lantada, R. Kumar, M. Guttmann, et al., “Synergies between surface microstructuring and molecular nanopatterning for controlling cell populations on polymeric biointerfaces,” Polymers 12, 655 (2020). [CrossRef]
27. S. Ristok, S. Thiele, A. Toulouse, et al., “Stitching-free 3D printing of millimeter-sized highly transparent spherical and aspherical optical components,” Opt. Mater. Express 10, 2370–2378 (2020). [CrossRef]
28. S. Aumann, S. Donner, J. Fischer, et al., “Optical coherence tomography (OCT): principle and technical realization,” in High Resolution Imaging in Microscopy and Ophthalmology (2019), pp. 59–85.
29. X. Jinzhi, S. Qing, and C. Wei, “A novel single tone frequency estimation by interpolation using DFT samples with zero-padding,” in IEEE 13th International Conference on Signal Processing (2016), pp. 277–281.
30. A. Agrawal, T. J. Pfefer, P. D. Woolliams, et al., “Methods to assess sensitivity of optical coherence tomography systems,” Biomed. Opt. Express 8, 902–917 (2017). [CrossRef]
31. R. F. Spaide and C. A. Curcio, “Anatomical correlates to the bands seen in the outer retina by optical coherence tomography: literature review and model,” Retina 31, 1609–1619 (2011). [CrossRef]
