Intraoperative optical coherence tomography in ophthalmology: Technologies and applications

Abstract

Intraoperative optical coherence tomography (iOCT) offers valuable real-time, depth-resolved visualization of ocular anatomy and during ophthalmic surgical maneuvers, which can be used to augment clinical decision-making, help verify surgical endpoints, enhance surgical precision, and facilitate the development of novel surgical techniques. Early iOCT demonstrations used perioperative devices, such as handheld and intraocular probes, which required pauses in surgery and disrupted clinical workflow. The advent of microscope-integrated systems addressed these limitations, allowing for iOCT imaging concurrent with surgical microscopy. iOCT image visualization has similarly progressed from external monitors, which require surgeons to divert their gaze, to heads-up displays integrated into microscope oculars, enabling direct overlays and improved ergonomics. Most recent advances have included increasing imaging speed to enable four-dimensional visualization of surgical dynamics and integration of automated surgical instrument tracking technologies. Clinical translation of iOCT has demonstrated utility across a range of procedures, including glaucoma surgery, corneal transplants, cataract extraction, vitrectomy, membrane peel, retinal detachment and macular hole repair, subretinal injection, and retinal prosthesis placement. As more advanced technologies are integrated into the conventional ophthalmic surgical workflow, iOCT has the potential to improve surgical performance and patient outcomes.

Introduction

Intraoperative optical coherence tomography (iOCT) offers benefits to ophthalmic surgery by providing valuable real-time, depth-resolved visualization of changes in ocular anatomy during surgical maneuvers. iOCT can enhance clinical decision-making, help verify surgical endpoints, and has the potential to improve patient outcomes across various ophthalmic procedures. Early iOCT systems required pausing surgery to acquire images, which disrupted clinical workflow. However, this limitation has been addressed by the latest-generation microscope-integrated iOCT systems, which allow for continuous, concurrent imaging of surgical dynamics. Potential applications for iOCT are broad and include glaucoma surgery, corneal transplants, cataract extraction, and vitreoretinal surgery. While current-generation iOCT systems still have limitations, such as fundamental trade-offs between field-of-view (FOV) and imaging speed, ongoing research is focused on optimizing their performance and further enhancing the clinical utility of the technology.

Perioperative Optical Coherence Tomography

The first demonstrations of iOCT imaging during ophthalmic surgery were performed perioperatively in 2009. These studies included the use of handheld optical coherence tomography (HH-OCT) and intraocular optical coherence tomography (OCT) probes during surgery, which represented significant technological advancements by overcoming the constraints of conventional benchtop OCT devices that are unsuitable for imaging supine patients. The portability of HH-OCT and intraocular probes allowed them to be readily translated into the operating room to capture some of the first iOCT images of ophthalmic surgery, but necessarily requiring changes to the standard clinical workflow such as pausing surgery and removing surgical instruments from the eye. Despite these challenges, early studies effectively showcased the value of iOCT for augmenting clinical decision-making and assessing the completion of surgical goals in real time.

The first perioperative demonstration during ophthalmic surgery utilized a commercially-available HH-OCT probe (Bioptigen Inc.) to obtain depth-resolved images during macular hole repair, epiretinal membrane peel, and vitrectomy for vitreomacular traction.[1] iOCT allowed for visualization of persistent inner limiting and epiretinal membranes, and en face volume projections enabled the identification of retinal landmarks during surgery. Postprocessing segmentation of macular hole images was used to quantify changes in hole size immediately following surgical repair.

Building on this work, a 2011 study implemented custom mounting hardware to attach the Bioptigen HH-OCT probe directly to an ophthalmic surgical microscope.[2] This integration provided two key advantages: facilitating the alignment of the OCT FOV using microscope controls and reducing motion artifacts. iOCT imaging of 24 macular hole and epiretinal membrane cases showed robust identification of residual membranes, accessible areas for peeling, and alterations in retinal thickness and macular hole dimensions during surgical repair. An increase in temporary detachments and subretinal fluid was also observed intraoperatively for the first time and provided valuable insight into how surgical manipulation impacts retinal membranes and surfaces.

A subsequent 24-month prospective study using the same microscope-mounted HH-OCT probe conducted in 2014 evaluated its utility across 531 eyes undergoing both anterior and posterior segment surgery [Figure 1].[3] Enrolled patients included those undergoing Descemet stripping automated endothelial keratoplasty, deep anterior lamellar keratoplasty (DALK), cataract extraction and intraocular lens (IOL) implantation, laser-assisted in situ keratomileusis, vitrectomy, scleral buckling, and other retinal procedures. The outcomes of the study highlighted the benefit of iOCT for improving surgeons understanding of ocular anatomy in over 40% of cases. Similar studies have also demonstrated comparable clinical utility using the iVue HH-OCT system (Optovue Inc.) during vitreoretinal and anterior segment surgery.[4,5]

Figure 1.

Perioperative intraoperative optical coherence tomography (iOCT). (a) Handheld optical coherence tomography probe mounted to surgical microscope provides increased stability during imaging. iOCT shows clear visualization of (b) full-thickness cornea (arrows, left and right) and bare descemet membrane (arrow, middle) following stromal dissection. Imaging during epiretinal membrane surgery shows (c) membrane before peeling (arrows) and (d) residual membrane after peeling (arrow). Reprinted with permission[3]

HH-OCT has also been used to intraoperatively image sedated infants undergoing screening for pediatric retinal diseases. A 2017 study showed that widefield HH-OCT angiography in neonates was able to resolve microvascular changes following laser treatment for retinopathy of prematurity (ROP).[6] Diagnostic HH-OCT has also revealed vascular irregularities in infants with ROP and detected regions of macular nonperfusion, inner retinal atrophy, and vascular abnormalities associated with pediatric incontinentia pigmenti.[7,8,9,10,11,12,13] Multimodal imaging combining HH-OCT with scanning laser ophthalmoscopy has enabled visualization of parafoveal cones in infants and children and provided novel insights into photoreceptor migration and density during retinal development.[14]

Although OCT is widely used as an ophthalmic diagnostic modality, iOCT imaging of the intact eye during surgery presents significant hurdles. These include reduced image quality due to ocular opacities, limited FOV in the peripheral retina, finite imaging depth, obscurations resulting from surgical instruments shadows or reflections, and constantly moving regions-of-interest (ROIs) requiring frequent FOV adjustments. Intraocular imaging probes are particularly well suited to overcome many of these challenges, and they have been used to achieve imaging in the far periphery and suprachoroidal space. Intraocular probes offer several key benefits for iOCT including intuitive FOV aiming and alignment, integrability with various ophthalmic surgical instruments and robotic actuators, and compatibility with surgical microscopes without needing hardware modifications. Prototypes of these probes have been integrated with surgical instruments and robotic systems to demonstrate novel surgical functionality.

Early work in 2009 demonstrated single A-line measurements through a microsurgical pick using a common-path OCT system.[15] While this system required an external metrology system and had a limited depth range, it was successfully used for tasks such as instrument position tracking, retinal imaging, and preventing tool-tissue collision with robotic assistance. The same probe design was integrated with surgical forceps and a motion sensor to provide OCT-based distal sensing and physiological tremor compensation, enhancing accuracy during vitreoretinal surgery.[16,17,18] Integrating an OCT-based distal sensing with a handheld microinjector system achieved micron-scale subretinal cannula positioning accuracy.[19] Distal sensing intraocular probes integrated with robotic systems have demonstrated automated needle insertion for procedures like hydrodissection during DALK.[20] Forward-viewing intraocular probes, designed for OCT B-scan imaging through small-gauge cannulas, were presented in 2013 for use in retinal surgery.[21] When combined with pulsed-laser delivery, these forward-viewing probes could provide real-time tissue ablation feedback during retinectomy.[22] A similar probe design was directly integrated into telemanipulated surgical forceps to demonstrate OCT guidance combined with surgical robotic assistance, which significantly reduced hand tremor and improved tool manipulation precision and accuracy.[23] Integration of these probes with subretinal cannulas enabled continuous imaging of subretinal bleb formation in simulations relevant to gene therapy delivery.[24,25] And side-viewing probes have also demonstrated their capability for imaging intraoperative maneuvers and ocular structures in animal models.[26]

Beyond static imaging, the real-time depth-resolved OCT data provided by these probes can be used for closed-loop applications, including robotic-assistive devices, technologies for tremor reduction, and image-guided needles. While current research probes primarily offer A-line and B-scan imaging, future advancements in miniaturized scanning technologies or integration with robotic scanning hardware may enable intraocular volumetric imaging. However, it is crucial that additional design considerations and biocompatibility criteria are met to ensure patient safety and compliance with surgical sterilization requirements.

Microscope-integrated Optical Coherence Tomography

iOCT using handheld and intraocular probes inherently causes disruptions to clinical workflow because it precludes imaging concurrent with surgical microscopy visualization. These perioperative OCT technologies require the surgeon to pause the procedure and remove surgical tools from the eye, which fundamentally means that they cannot be used for real-time OCT imaging feedback or guidance during surgical maneuvers. These handheld technologies are also highly susceptible to motion artifacts, which often limit the performance of functional imaging modalities, such as OCT angiography. Microscope-integrated OCT (MIOCT), first demonstrated in 2010, overcomes both the surgical workflow disruptions and stability limitations associated with perioperative OCT probes by eliminating the need to pause surgery or remove instruments from the eye when acquiring OCT images.[27,28,29,30] Simultaneous OCT and surgical microscope visualization avoids extending surgery times and, crucially, allows for real-time visualization of surgical maneuver dynamics and the interactions between tissue and instruments as they happen. Observing these dynamics in real time can directly inform surgical decision-making and potentially lead to improved postoperative outcomes.[31,32,33]

OCT can be optically combined with ophthalmic surgical microscopes through either a microscope camera port or by using a dichroic mirror placed immediately before the microscope objective lens.[34,35] Sharing the objective makes the MIOCT system inherently parfocal with the surgical microscope and, thus, also compatible with vitreoretinal visualization accessories such as binocular indirect ophthalmomicroscope attachments or contact lenses. Careful design of the OCT relay is needed to satisfy FOV specifications and optimally fill the objective aperture to achieve maximum lateral resolution. These research MIOCT systems are often designed as modular scan heads that can be mounted or unmounted between surgeries and connected to external OCT engines [Figure 2a and b]. However, MIOCT scan-heads necessarily increase the overall height of the microscope stack, which may affect surgeon ergonomics.

Figure 2.

Examples of microscope-mounted optical coherence tomography (MMOCT) systems. (a and b) Research prototypes showing modular scan-heads attached to ophthalmic surgical microscopes. (c and d) Commercial MMOCT with optical coherence tomography optical relays integrated into surgical microscope bodies. Reprinted under creative commons attribution license (CC BY).[34,35,36,37] MMOCT = microscope-mounted optical coherence tomography

Development and integration of state-of-the-art swept-source OCT (SS-OCT) engines have significantly advanced MIOCT capabilities, enabling four-dimensional (4D) imaging of surgical dynamics. OCT angiography has also been successfully implemented using MIOCT to provide functional contrast of retinal vascular perfusion intraoperatively without the need for exogenous dyes. Finally, multimodal technologies that combine OCT with scanning laser ophthalmoscopy or surgical video have been developed to enhance visualization and contrast or for surgical tool-tracking to enhance MIOCT functionality.

Current commercial MIOCT systems include the Zeiss RESCAN 700,[36] Leica Microsystems EnFocus iOCT,[37] and Haag Streit iOCT.[38] These commercial systems offer advantages such as optimized optical throughput and resolution performance as compared to research prototypes. Additionally, the RESCAN 700 and EnFocus systems have OCT scanning relays and engines that are fully integrated with the microscope body, which enhances clinical robustness and ergonomics [Figure 2c and d]. The primary drawback of current-generation commercial MIOCT is slow imaging speeds. These systems utilize spectral domain OCT (SD-OCT) engines with <50 kHz line rates, which are an order of magnitude slower than some SS-OCT-based MIOCT research prototypes.

The availability of commercial MIOCT systems over the past decade has successfully expanded clinical access to iOCT in ophthalmology. These technologies have been leveraged to visualize changes in ocular anatomy during surgery, assess impacts on surgical ergonomics, confirm surgical endpoints, and explore the potential benefits of iOCT on postoperative outcomes.[39,40,41,42,43] Future applications of MIOCT may include the development of novel OCT-guided surgical maneuvers and instruments, more robust clinical training, and integration with surgical robotics for teleoperated or automated ophthalmic surgery.

Enabling Technologies

Depth-resolved iOCT complements conventional en face surgical views and uniquely resolves subsurface anatomy in real time. 4D iOCT, first presented in 2011 and later refined in 2016, further enhances intraoperative visualization by imaging dynamic changes in tissue microstructure during and in response to surgical maneuvers. These data can be used to provide both qualitative visual feedback and quantitative metrics to guide ophthalmic surgeries and predict visual function outcomes. Imaging data on instrument-tissue interactions has the potential to improve the precision and reproducibility of surgical maneuvers, augment clinical decision-making in real time, and provide novel clinical insights that lead to the development of new surgical instrumentation and techniques. However, 4D iOCT performance is constrained by a fundamental trade-off between imaging speed, FOV, and sampling density. Current active areas of research are developing new raster scanning waveforms and instrument-tracking technologies to circumvent these limitations in an effort to improve the clinical utility and ergonomics of current-generation iOCT systems. If successful, these technologies may significantly reduce the technical burden and financial/time cost associated with the clinical adoption of iOCT.

The first demonstrations of video-rate iOCT feedback were performed in 2010 using early SD-OCT engines with modest line rates below 40 kHz. These systems were too slow for acquiring high-speed volumes and, thus, real-time video-rate imaging was limited to continuous B-scans.[28,31,32,34,35] Similarly, live B-scans of surgical dynamics remain the iOCT paradigm for current-generation commercial iOCT systems, which have line rates between 10 and 32 kHz.[36,37,38] iOCT images are acquired across either two orthogonal or several parallel B-scans over the ROI for real-time feedback and densely sampled are acquired during periodic pauses in the surgical procedure for postoperative analysis and visualization.[33,44]

Starting in 2011, advanced research-grade systems have leveraged high-speed SS-OCT engines and GPU-accelerated image processing achieve line rates of 100 kHz to 1.67 MHz, allowing for video-rate 4D imaging at >10 volumes-per-second (vps).[45,46,47,48,49] However, 4D imaging also presents its own set of challenges. Volume refresh rates are slow compared to continuous B-scan frame rates and do not address the trade-off between acquisition time and FOV.[50] Raster scanning limitations also introduce dead time to each B-scan but may be mitigated using novel scan waveforms.[49] Finally, real-time video-rate 4D iOCT feedback requires computationally intensive rendering and display algorithms that may be constrained by bottlenecks in data acquisition, processing, and memory transfer.[46]

While 4D iOCT promises to provide depth-resolved volumetric imaging of surgical dynamics, imaging still requires precise alignment of imaging FOVs to ocular ROIs. The need for manual alignment of the iOCT FOV, whether performed by an imaging technician or the surgeon using foot pedals, remains cumbersome and may extend procedure times.[33] Again, FOV size remains a barrier in use cases when the ocular anatomy changes significantly in response to a surgical maneuver (e.g. subretinal injection of large volumes).

The challenges of manual FOV alignment and 4D imaging of surgical dynamics may be overcome by the development and integration of automated instrument-tracking technologies. Given limitations on the maximum permissible exposure on the eye and the minimum clinically usable signal-to-noise ratio (SNR) required, iOCT imaging speed is constrained to <400 kHz.[51] To maintain image quality when performing video-rate 4D imaging, it is often necessary to restrict iOCT to small, densely sampled FOVs. Therefore, the capability to track anatomical features and/or surgical instruments and automatically reposition the iOCT FOV to span ROIs is crucial. Early iOCT systems required an imaging technician to identify ROIs on the surgical documentation video and manually offset the iOCT FOV using a mouse pointer.[32,47] A similar manual alignment is used in current-generation commercial iOCT. While manual alignment ensures the iOCT FOV spans the ROI, the resulting interruption to surgical workflow has been shown to increase operation times by as much as 25 min.[3]

Automated tracking technologies, first integrated in 2014, may overcome these surgical workflow challenges and significantly improve iOCT ergonomics. Automated axial tracking adjusts the iOCT reference arm to keep samples centered within the available depth range.[36,44] Semiautomated axial-tracking is available in the Zeiss RESCAN 700 and several research groups have developed segmentation-based approaches to localize surgical instruments or anatomical surfaces in iOCT B-scans or volumes for automated axial-tracking.[52,53,54]

Automated lateral tracking adjusts the iOCT FOV to be centered on fiducials or surgical instrumentation. Methods include using binocular stereovision to triangulate fiducials mounted on surgical instruments and machine-learning approaches for object detection.[55,56,57] While conventional object detection works for lateral tracking in the front of the eye, tracking features in the posterior eye is more challenging due to variations in eye length and position that distort the iOCT FOV relative to the surgical view. To address posterior eye challenges, one study implemented multimodality imaging, object detection, and adaptive sampling to achieve automated 4D instrument tracking at a 16 Hz volume rate [Figure 3]. Adaptive sampling refers to using novel raster scanning waveforms to circumvent the inherent tradeoff between sampling density, imaging speed, FOV, and SNR.[47,49,57,58]

Figure 3.

Automated instrument tracking using multimodal imaging. (a) Closed-loop control schematic using machine-learning to automatically detect surgical instruments on en face spectrally encoded reflectometry (SER) images and center optical coherence tomography (OCT) field-of-view onto instrument tips. (b) Four-dimensional (4D) imaging with instrument tracked OCT showing cross-sectional B-scans and 4D render. Reprinted under creative commons attribution license (CC BY).[57] OCT = Optical coherence tomography, CNN = Convolutional neural network

In summary, tracking capabilities, whether manual and automatic, are essential for successful 4D iOCT visualization of surgical dynamics. Automated tracking, utilizing techniques like segmentation, stereovision, and machine learning, offers promising solutions to improve ergonomics and clinical utility over manual tracking approaches. Combined with advances in scanning waveforms, these efforts aim to overcome technological, ergonomics, and workflow hurdles and accelerate the clinical adoption of iOCT technology.

Visualization and Feedback Technologies

Conventional diagnostic OCT volumes are viewed using cross-sectional “flythroughs” or as three-dimensional (3D) volume renderings. While these formats are standard for analyzing static volumes, they are not ideal for providing real-time or intraoperative feedback during surgery. The challenge lies in effectively presenting dynamic OCT information in a way that complements the en face view through the surgical microscope and fits seamlessly into the clinical workflow. Therefore, the translation of iOCT technologies into the surgical suite has also necessitated the parallel development of new technologies for OCT data display and visualization.

One common method for displaying iOCT data in both research and commercial systems uses external monitors. While broadly adopted, the use of external monitors requires the surgeon to shift their gaze away from the surgical field, which requires a pause in surgery and for all instruments to be removed from the eye. This also restricts feedback to merely confirming the completion of surgical goals rather than providing true real-time intraoperative feedback. Early clinical studies using commercial iOCT with external monitors noted an increase in procedure times ranging from 2 to 6 min.[3,59,60]

The Leica TrueVision and Alcon NGENUITY 3D visualization systems, first released in 2012, offer stereoscopic monitors as an alternative to external monitors. These systems utilize stereoscopic cameras to capture the surgical field, which is displayed on a 3D monitor viewable using polarized glasses.[59,61] A key advantage of these systems is that they can simultaneously display the stereoscopic microscope view and any digital data, including surgical plans or iOCT images. However, since these monitors no longer require surgeons to look through surgical oculars, they may cause spatial disorientation. In addition, stereoscopic monitors only provide optimal performance across a narrow range of view angles and image quality can be impacted when viewed off-axis.

Head-up displays (HUDs) offer advantages over both external and stereoscopic monitors for integrating iOCT data into the surgical workflow. HUDs function by using beamsplitters to optically combine images from LED or LCD panels with the microscope ocular view. Advanced HUD designs can utilize separate panels for each ocular to deliver stereoscopic views. The most substantial advantage of HUDs is their ability to overlay iOCT data directly onto the surgical FOV and eliminate the need for the surgeon to divert their gaze away from the microscope.

The Zeiss RESCAN 700 and Haag-Streit iOCT, FDA cleared in 2014 and 2015 respectively, both incorporate HUDs into their microscope oculars [Figure 4]. Commercially available HUDs have been integrated with research-grade iOCT systems,[35] and custom HUDs have demonstrated stereoscopic display of iOCT data.[62] Early clinical studies suggest a preference for HUD technology over external monitors and is attributed to improved surgical ergonomics.[63] However, HUDs are often limited in their display quality as compared to external monitors. Their poor display contrast is a result of using highly asymmetric beamsplitter ratios to ensure sufficient surgical microscope light throughput. Consequently, display overlays are generally restricted to dark or unused portions of the surgical field. Another limitation is the use of backlit panels adds haze to the surgical view even when not displaying information.

Figure 4.

Head-up display (HUD) visualization in commercial Zeiss RESCAN 700 intraoperative optical coherence tomography (iOCT). HUD shows optical coherence tomography (OCT) field-of-view (white box) with two orthogonal OCT B-scans imaged along the red and blue lines. Surgical microscope and corresponding iOCT B-scan showing (a) anterior chamber intraocular lens and (b) forceps engaging the internal limiting membrane. Reprinted with permission[44]

Augmented reality (AR) technologies are an emerging approach to facilitate surgical feedback, guidance, and training. AR systems allow for the integration of iOCT imaging data or even pre-operative surgical plans directly into the surgical context. Examples of AR applications include the Micro VisTouch ocular surgical simulator, which has been used to enhance vitreoretinal surgery training by incorporating OCT scans from real patients,[64] and the first immersive VR-OCT viewer using a head-mounted VR system (HTC Vive), which was presented in 2018, enabled reviewing pre-recorded volumes and guiding mock surgical procedures.[65]

While AR offers the potential for a highly immersive experience, which could intuitively integrate comple × 3D OCT data, there are significant challenges to adoption. A major concern is the potential for motion sickness, which requires frame rates of >90 frames per second to mitigate.[66] Achieving these high frame rates often necessitates sacrificing rendering resolution, which can result in low-quality displays.[65] These systems can add additional complexity to the clinical workflow, often relying on experienced technicians to manipulate the display information or orientation in real time.

Advancements in processing and rendering are enabling real-time, enhanced feedback during surgery. GPU-based image-processing and rendering capabilities are crucial for achieving video-rate 4D iOCT throughput. 4D iOCT visualization has progressed steadily from 5 vps in 2011[45] to 60 vps using optimized ray casting models in 2016.[46] Crucially, this rendering technology has been successfully combined with an iOCT system to deliver live volume renderings through a stereoscopic HUD at 10 volumes/second during mock ophthalmic surgeries.[47] Increased processing power has also enabled the integration of segmentation algorithms to provide quantitative intraoperative feedback. The addition of color contrast to iOCT volume renderings improved visualization and significantly enhanced surgical performance in several key areas, including differentiating between epiretinal membranes and retina, identifying when the instrument came into contact with the membrane, and tracking retinal deformation.[67]

These enabling technologies in processing and visualization are essential and allow surgeons to interpret iOCT imaging data in real time. The development of novel stereoscopic displays and HUDs will be key to improving surgical ergonomics and facilitating broad iOCT adoption. Furthermore, as iOCT imaging speed and image-processing throughput continue to increase, the need for advanced methods for 4D visualization becomes even more critical. These advanced methods should aim to add contrast, improve data interpretability, or provide quantitative metrics. Ultimately, these enhanced visualization techniques may serve as invaluable feedback mechanisms that directly inform clinical decision-making during ophthalmic surgery.

Clinical Applications of Intraoperative Optical Coherence Tomography

iOCT imaging and feedback can improve surgical visualization during conventional maneuvers and even enable more complex procedures that might otherwise have a significant risk of failure. However, the clinical value must be weighed against factors such as cost, potential changes to surgical workflow, and increased procedure time.

iOCT is demonstrated benefits in glaucoma surgery aimed at relieving excess intraocular pressure. In trabeculectomy, iOCT can provide real-time feedback on scleral incision depth and thickness, which is critical for flap effectiveness.[68] For trabeculectomy revisions, it can help localize fibrotic adhesions.[69] During glaucoma drainage device placement and minimally invasive glaucoma surgery, iOCT can aid navigation to or creation of a canal through the anterior chamber angle. Image-guided placement may reduce surgery time, prevent corneal decompensation, and minimize hyphema and tube leaks.[70] Finally, iOCT may facilitate faster and more thorough clearing of debris from intratrabecular spaces during trabecular aspiration.[68]

In corneal transplant surgery, iOCT provides crucial feedback for precise layer-specific surgery. These surgeries require removing only diseased layers and have advantages over full-thickness grafts but depend heavily on surgeon experience.[71,72,73] In DALK, iOCT provides immediate feedback on dissection depth to help remove stroma without perforating Descemet’s membrane.[71,74] In Descemet’s stripping endothelial keratoplasty, iOCT can readily visualize fluid trapped between the host cornea and graft, which is difficult to detect on conventional surgical microscopy.[75] Finally, in Descemet’s membrane endothelial keratoplasty, iOCT can be used to confirm graft orientation, potentially reducing detachment and failure rates.[76]

Achieving optimal postoperative vision following cataract surgery requires precise placement of the IOL. Even small displacements can cause significant errors.[77,78] iOCT can provide 3D feedback on lens decenter, tilt, and axial position and has been shown to reduce variability in lens placement.[79] This quantitative feedback can directly inform clinical decision-making for improved postoperative outcomes.

Real-time depth-resolved visualization of the retina during vitrectomy allows surgeons to monitor for retinal traction. In turn, surgeons can make adjustments to vitrector cutting speeds, aspiration rates, and position to minimize tissue injury. iOCT has also been shown to reduce surgery time in proliferative diabetics and provide clinically relevant feedback on retinal abnormalities in patients with dense vitreous hemorrhage.[80,81]

iOCT can visualize retinal membranes without needing exogenous dyes, provides information on retinal integrity before and after peeling, and can be used to confirm whether there is residual epiretinal membrane and whether there is iatrogenic macular hole after membrane peeling.[82,83,84] It has been used to identify scattering changes in the photoreceptor layers that may indicate damage undetectable by conventional microscopy.[85] Thus, real-time iOCT feedback may help surgeons minimize retinal trauma during membrane peels.[86]

In retinal detachment and macular hole repair, iOCT can be used to identify the presence and quantify the amount of subretinal fluid, which can impact postoperative outcomes.[87] When treating macular holes using the inverted internal limiting membrane flap technique, iOCT can be used to confirm the macular hole is properly covered by the flap.[88] Trans-tamponade iOCT, though optically challenging, can even allow intraoperative measurement of subretinal fluid volumes.[89]

Precision is paramount during subretinal injection of stem cells or gene therapies.[90,91,92] While a pre-bleb is often created to initiate detachment, iOCT allows for direct visualization of the injection location and provides quantitative feedback on bleb dimensions.[93,94,95,96] This enables more precise and reproducible subretinal injections. The ability to quantify parameters like bleb dimensions highlights the benefits of quantitative iOCT feedback.

Finally, for retinal prosthetics to achieve optimal stimulation, the distance between the electrode array and retina must be minimized.[97] Visualizing this gap using conventional surgical microscopy nearly impossible. However, iOCT provides robust feedback on prosthesis apposition and improves array placement for devices like the Argus II system.[98]

Conclusions

The diverse applications of iOCT in ophthalmic surgery underscore the need for effective intraoperative visualization technologies. As imaging speed and processing throughput increase, the need for advanced 4D visualization methods that add contrast, improve interpretability, or provide quantitative metrics becomes even more critical, ultimately serving as invaluable feedback mechanisms to inform clinical decision-making directly during surgery. Continued development of novel iOCT imaging and display technologies will be key to improving surgical ergonomics, facilitating iOCT adoption, and developing next-generation image-guided surgical techniques that improve patient outcomes.

Data availability statement

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.

Conflicts of interest

Royalties: Leica Microsystems, Carl Zeiss, Patents: Vanderbilt University, Duke University, Cleveland Clinic Foundation.

Funding Statement

This research was supported by the US National Institutes of Health (Grant Nos. R01-EY030490, R01-EY031769, and R01-EY033969). The content was solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.