Lamellar keratoplasty (LK) is a transplant procedure used to treat corneal diseases or injuries with a lower rejection and higher integrity of the cornea after transplantation.188.8.131.52.–6 It enables the partial replacement of a recipient’s corneal tissue with that of a donor. In the case of deep anterior lamellar keratoplasty (DALK), dysfunctional stroma of cornea is replaced after being clearly separated from the Descemet’s membrane (DM). The success rate for DALK surgery has been critically influenced by the extent of the uniformity and clearness of remaining lower layers. Since a human cornea is thin, typically , the separation process needs a specialized surgical technique.
Anwar’s big-bubble technique has been proven to be the most successful for this purpose.78.9.–10 In the big-bubble technique, air is injected through a thin syringe needle that is inserted inside the patient’s cornea to make the uniform corneal segments by inflating the air bubble in the cornea. Here, inserting the needle in the cornea is a potentially deadly part of the procedure. The needle may produce a fatal perforation if inserted too deeply, or no big-bubble may be formed if the air is injected at a shallow position. So far, the needle insertion has been performed with no direct information given about the depth of the needle’s tip but just with help of en-face visual guidance from the surgical microscope. As a consequence of the required precision of the needlework in DALK, the surgeon could risk perforating the corneal tissue.11,12 Moreover, it is difficult to predict the visual outcome because of the presence of uneven residual stromal thickness.13,14
To assist the sophisticated big-bubble formation, the use of intraoperative optical coherence tomography (iOCT)1516.17.18.–19 or iOCT with a surgical microscope2021.22.–23 imaging system has been considered for the DALK procedures. Monitoring the needle’s manipulations in the cornea can safely handle a sharp needle and better determine the injection point with knowledge of the depth. However, its practical use is still hindered by many factors. Tracking the needle tip in the OCT systems is a cumbersome task for both a human operator and a fully computerized system. Note that real-time three-dimensional (3-D) data covering the entire surgical area are difficult to be presented in a comprehensible form without heavy processing. Moreover, opaqueness of the metallic needle for the DALK surgery makes underlying tissue invisible in the conventional OCT imaging schemes.24 Therefore, the acquired OCT information is still limited in contents, difficult to comprehend, and vulnerable to the environmental factors in the real-world operation room. We believe that there is still room to further improve the entire surgical procedure by OCT visualization means.25,26
In this article, we introduce a position-guided needle combined with M-mode OCT for the DALK procedures. We developed a needle that is internally equipped with an ultrathin fiber-based OCT probing optics. Our needle utilized its OCT principle that acquires the back-reflected OCT light through the needle’s bevel outlet. In addition, free-hand manual scanning was employed in axial direction to display the dynamically varying location of the needle tip over the time (M-mode). Whereas the OCT needles developed in the previous studies were mainly for diagnostic imaging,2728.29.30.–31 our needle was used for fluid injection just like the ordinary syringe needles as well as position sensing. The performance of our system was initially evaluated through a cornea mimicking phantom, and its applicability in real ophthalmic surgery was identified in the animal model ex vivo.
Materials and Methods
Position-Guided Surgical Needle and M-Mode Optical Coherence Tomography
We developed a position-guided surgical needle, which equips an embedded OCT that can acquire the positional contexts of the underlying tissue. For a compact and minimalistic structure, our surgical needle employed a stepwise transitional core (STC) fiber [Fig. 1(a)], which produces an output beam without using a lens element.32 It was composed of three different optical fibers with an outer diameter of , spliced in series to the standard single-mode fiber of light delivery. The large core fiber with its diameter of at the distal end directly determined the characteristics of the output beam while keeping the effective single-mode guidance through the STC structure. To bend the beam perpendicularly to the fiber axis, a 43 deg-tilted reflector plane was made by polishing the end of the fiber structure. This structure is advantageous in its compactness, ease of fabrication, and operational reliability, though with a moderate sacrifice of the imaging range. The STC fiber supported an effective imaging range of 0.9 mm with beam diameter of in corneal tissue (),32 so was sufficient for our application to corneal imaging. The fabricated fragile fiber structure was reinforced by a metal jacket with an outer diameter of . This fabricated fiber was finally integrated into a 26-G syringe needle as shown in Fig. 1(b). The inner and outer diameters of the position-guided needle were 260 and , respectively. The fluid channel was formed between the metal-jacketed fiber and the inner side of the syringe needle.
In this study, to acquire the depth information from the needle, we utilized the home-built swept source OCT (SS-OCT) imaging system as shown in Fig. 1(c). The swept source laser (Axsun Technologies Inc.) with 1310-nm center wavelength and 105-nm bandwidth was incorporated into the SS-OCT system, providing average 20 mW of output power and an axial resolution of . A 90:10 coupler splits the light emitted from the swept source laser into the mirror in the reference arm and sample arm. The position-guided needle was attached to the sample arm in our system to achieve optical signals from the inside of the tissue. The power from the sample arm was measured . Reflections from each arm were combined and generated the spectral interference signals at the 50:50 coupler. A balanced photo detector (PDB450C, Thorlabs Inc.) and digitizer (ATS9350, Alazar Technologies Inc., Canada) were utilized to detect the interference signals. A fast Fourier transform was computed to achieve the position profile and image was updated based on time order with of update speed (M-mode). Our customized SS-OCT system facilitates 50-kHz A-line acquisition.
Corneal-Mimicking Phantom Preparation
To verify the functionalities of the developed position-guided surgical needle, we prepared the corneal-mimicking phantom, which has two layers, representing the cornea and the anterior chamber, respectively. The top layer was prepared by blending an agarose powder and distilled water, and a scatterer was added to the mixture to increase the optical scattering while imitating the stroma texture. The bottom layer had a low concentration of mixture compared with the top layer to realize the transparency similar to that of the anterior chamber. Solidifying the subsequently poured mixtures of differently concentrated agarose develops the corneal-mimicking phantom. The top layer of corneal-mimicking phantom was fabricated with 1-mm thickness and then utilized to evaluate the functions of the developed position-guided needle.
Ex Vivo Animal Study Preparation
The rabbit eyes were enucleated from two New Zealand rabbits weighing 2.0 to 2.5 kg that were euthanized in accordance with institutional guidelines and approved Institutional Animal Care and Use Committee at the College of Medicine, the Catholic University of Korea. The rabbit eye was held in an anterior chamber manufactured by a 3-D printer for an optimal surgical procedure. The anterior chamber consisted of a plastic cradle with two nozzles, which controlled pressure using water and air to fasten the eye tightly. After setting the eye in the cradle, a position-guided needle was inserted into the cornea, which measured the remaining thickness of the cornea in real-time when the needle end was moving toward a border between the cornea and the anterior chamber. When the needle end reached the target location inside the cornea, air was blown through the needle to separate two layers.
Evaluation of Phantom Study
To identify the performance of the surgical needle as a position sensor, the corneal-mimicking phantom was utilized, as shown in Figs. 2(a) and 2(b). For this experiment, an OCT image was taken to monitor the needle insertion in the phantom and calibrate the position information acquired from the needle. Figure 2(c) shows an OCT image of the phantom without a needle and clearly delineates the two solidified agarose layers with different concentrations and scatters. When the surgical needle is inserted into the phantom, an OCT image and position information from the needle were simultaneously acquired, as presented in Figs. 2(d) and 2(e). The A-line profile indicating the position information of the needle is directly converted and visualized as an M-mode OCT image. Video 1 presents demonstration of M-mode OCT when the position-guided needle is applied to the cornea phantom and approaches the desired location. The M-mode OCT image, which was updated every 10 ms, denotes the time sequential signals along the depth from the end of the STC fiber. As the needle penetrated into the phantom, the thickness of the M-mode OCT image gets thinner while showing the residual thickness of the top layer. In this experiment, we confirmed that a position-guided needle enables us to clearly distinguish the two layers, and offers position location with high accuracy. The sensing speed was also fast enough to apply to real surgical settings while guiding the placement of a needle.
Evaluation of Animal Study
The M-mode OCT needle was applied to the DALK procedure with an ex vivo rabbit animal model with manual scanning. In this experiment, a position-guided needle with the bevel facing downward was carefully penetrated tangentially into a paracentral rabbit cornea. The entire DALK procedure was recorded by video and M-mode OCT as shown in Video 2. Figure 3 shows how the surgical needle is inserted and guided to approach near the DM. A time-sequential A-line profile acquired from the distal end of the needle was continuously monitored while updating the current location. The real-time feedback from the inserted needle was visualized on the display to help the clinician perceive the position of the needle. The whole cornea thickness was visualized at the epithelial layer [Fig. 3(a)] and became thinner when the needle approached the paracentral cornea [Fig. 3(b)]. The green line shown in the Video 2 indicates the current residual thickness between the DM and the needle end, which provide fast and intuitive information to the clinician.
We further evaluated the feasibility of bubble formation induced by our needle using SS-OCT imaging system. voxel of OCT images were captured immediately after the needle was placed at above the DM layer and after performing the air injection at the same position. Figures 4(a) and 4(b) show the DALK process including needle insertion and the various bubble formations. As shown in Figs. 4(c) and 4(d), the 3-D OCT image presents the whole structure of the cornea with the needle. In DALK surgery, a clinician usually lays the needle parallel to the corneal surface to avoid the risk of perforation under the DM layer. This method may also help to easily separate the corneal layer with less damage in stromal layer when the big-bubble technique is used. 2-D and 3-D OCT images clearly identify the needle location inside the cornea in both success and failure cases of big-bubble formation. The blue and red lines in Figs. 4(c) and 4(d) show the cross-sectional image perpendicular and parallel to the needle, respectively. After the big-bubble technique was applied, it was found that more than half of the stroma layer was detached and air entered underneath, as shown in Fig. 4(d). From Fig. 4, we realized that DALK was successfully operated without any tissue damage by means of the position-guided needle.
Discussion and Conclusion
In this study, we introduced the M-mode OCT needle, which enables us to improve the surgical yield of DALK while minimizing fatal perforations due to unexpected needle movement. The STC fiber, operating as an optical sensor in the surgical needle, successfully traced the position of the needle end in the real time. This was confirmed by the phantom study. Through the animal model, we identified that the position-guided surgical needle offered the accurate feedback of needle placement inside the cornea, which enabled stable air bubble formation. Furthermore, displaying the residual thickness from the desired position as an M-mode OCT image could deliver comprehensive information to the user throughout the entire surgery. Thus, experimental results suggest that our device and corresponding new surgical protocol have great potential for use in DALK as well as other microsurgeries.
However, there are several issues that need to be considered for further usage. No structural change was made, in this study, to the standard 26-G syringe needle’s body in keeping its original robustness for reliably penetrating soft tissue like corneal layers. However, there are possibilities of broken tip of the STC fiber due to the pressure during the insertion process into the harder tissues. Fortunately, these could be improved by blocking the end of needle partially to reduce the external forces while maintaining the passage for airflow. In addition, we have found smearing artifact due to nonuniform movement or abrupt stops encountered during data acquisition. This could possibly confuse users for finding the location of the needle. However, there are few methods reported in public to correct these artifacts in HW/SW wise.3334.–35
The current device can be enhanced and matured by adapting other advanced engineering techniques. Our surgical needle can readily be integrated into an existing intraoperative OCT with a surgical microscope, or implemented as a handheld surgical tool. Previous studies have reported successful adaptations of OCT as a surgical tool and its potential for monitoring the surgical procedures.15,20,28 Since OCT is a modular system, the imaging path of the OCT system can either be shared for intraoperative OCT imaging or a sensor. Few studies have described the feasibility of intraoperative OCT to DALK.18,19,23 Thus, the combination of existing intraoperative OCT and our device would make a great impact on microsurgery, which requires the precise manipulation of a needle as well as the continuous observation of a surgical region.
Our device platform offers real-time depth tracking sensor. We believe this can be transformed into a tremor compensating device36 by utilizing the motion feedback actuator and software based on our position information. In addition, current position information can be converted into a sound or alarm to make it a more user friendly surgical device. Thus, a fully developed and ergonomically engineered surgical needle would find numerous applications. The precise depth information of a needle with morphologic data could be required in microsurgery as well as in drug delivery.37
This work was supported by the Technology Innovation Program (No. 10047943) through the Ministry of Trade, Industry and Energy of Korea, the Basic Science Research Program (NRF-2014R1A1A1006183) through the National Research Foundation of Korea, and the 2017 Research Fund (No. 1.170009.01) of Ulsan National Institute of Science and Technology (UNIST). The authors would like to thank Jin U. Kang at Johns Hopkins University for his advice and comments related to this work.
Sungwon Shin received his MSc degree in control and measurement engineering from University of Ulsan in 2008. Currently, he is pursuing his PhD in the Department of Biomedical Engineering at Ulsan National Institute of Science and Technology. Since the spring of 2014, he has been with Dr. Woonggyu Jung’s Transitional Biophotonics Laboratory. His research interests include utilizing biomedical imaging for the detection of diseases and developing new surgical guidance tools for clinical applications.
Jung Kweon Bae received his BSc degree in electrical engineering from the University of Illinois at Urbana Champaign in 2013. Currently, he is pursuing his PhD in the Department of Biomedical Engineering at Ulsan National Institute of Science and Technology. Since the fall of 2013, he has been with Dr. Woonggyu Jung’s Transitional Biophotonics Laboratory. His research interests include biomedical imaging device, mobile-based imaging, point-of-care diagnostics, mobile health, and clinical applications.
Sucbei Moon received his MS and PhD degrees from Gwangju Institute of Science and Technology in 2002 and 2007, respectively. Since he joined the faculty of physics, Kookmin University in Seoul, South Korea, in 2011, he has been exploring various topics of applied optics, with particular emphasis on optical imaging technologies. His current research interest includes the optical coherence tomography, catheter-based imaging, and fiber-optic imaging techniques.
Woonggyu Jung is an associate professor of biomedical engineering at Ulsan National Institute of Science and Technology. He received his PhD in biomedical engineering from the University of California, Irvine, in 2008. He has published various papers in the field of biomedical optics and biophotonics. His research interests include medical imaging systems, point-of-care diagnostics, label-free imaging, optical stimulation, and clinical applications.