Digital Analysis of Intraoperative Fluorescein Channelography in Human Eyes During Ab-Interno Canaloplasty

Summary

This study presents a standardized open-source workflow for quantifying intraoperative fluorescein channelography during ab-interno canaloplasty, which can be used to assess aqueous humor outflow patterns and potentially predict the postoperative outcome with regard to intraocular pressure.

Abstract

Ab-interno canaloplasty (AbiC) is a minimally invasive glaucoma surgery designed to enhance physiological aqueous humor outflow by circumferential dilation of Schlemm’s canal. While fluorescein channelography has been used intraoperatively during ab-externo canaloplasty to visualize outflow pathways, its quantitative application during AbiC has not yet been evaluated. This study introduces a standardized workflow for quantitative analysis of intraoperative fluorescein channelography during AbiC, with the aim of assessing aqueous humor outflow patterns and their potential predictive value for surgical outcomes. For this purpose, AbiC was performed with fluorescein-dyed viscoelastic and early and late phase channelography images were acquired intraoperatively. A step-by-step image analysis workflow using the open-source image analysis software was developed to quantify subconjunctival fluorescein staining. Three complementary approaches were applied: Measurement of circumferential staining intensity using concentric rings around the cornea, Quadrant-based analysis of segmental staining, and qualitative assessment of focal staining patterns corresponding to possible sites of trabecular meshwork alteration. Exemplary cases suggested that extensive and homogeneous fluorescein staining across quadrants was associated with sustained intraocular pressure reduction and medication independence, whereas limited and heterogenous staining were observed in eyes with less favourable outcomes. This workflow therefore enables a reproducible, quantitative assessment of aqueous humor outflow during AbiC, without requiring expertise in image analysis, and can be used to predict the success of minimally invasive glaucoma surgery.

Introduction

Glaucoma is a multifactorial disease with progressive optic neuropathy and one of the most common causes of blindness worldwide^1,2. Various factors have been identified that contribute to pathogenesis of glaucoma - including structural vulnerability of the lamina cribrosa, impaired ocular perfusion, genetics and autoimmune processes - but the only modifiable risk factor remains elevated intraocular pressure (IOP)^3,4,5,6,7,8. The most common subtype, primary open-angle glaucoma (POAG), is characterized by an increased outflow resistance leading to a reduced outflow of aqueous humor and thus to a long-term increase in IOP⁹. This leads to retinal nerve fiber loss and to visual field defects as the disease progresses.

Several treatment strategies are available, which are selected depending on the severity and duration of glaucoma and the measured IOP. If topical IOP-lowering therapy with eye drops is not effective or not tolerated by the patient, laser procedures or surgical interventions can be performed^10,11. In addition to traditional filtering surgeries such as trabeculectomy, there are now a variety of lower-risk minimally invasive glaucoma surgeries (MIGS)¹². These include ab- interno canaloplasty (AbiC), which, in contrast to ab-externo canaloplasty, does not require an incision of the sclera or the creation of a flap¹³. The ab-interno approach offers a significantly less invasive procedure in which no suture is left behind as foreign material. It has also been shown that the effect corresponds to that of the ab-externo approach, with significantly simpler handling and fewer complications^14,15. An additional advantage of AbiC is that the conjunctival tissue is preserved for future interventions, if necessary.

In this procedure, after assessing the anterior chamber via a corneal incision, a microcatheter is inserted into the Schlemm's canal with the aid of a gonioscope, incising the trabecular meshwork and passing through it in a 360° circle. A viscoelastic expands Schlemm's canal and the ostiae of the collector channels and may induce focal micro ruptures of the trabecular meshwork¹⁶. These effects are thought to facilitate aqueous humor outflow by reducing resistance within the conventional outflow pathway, in accordance with the anatomy and physiology of the eye and by improving access to the collector channels, although the exact mechanisms are not yet fully understood¹⁷. Furthermore, AbiC can be combined with cataract surgery, which provides an additional IOP reduction. In this context, intraoperative visualization of the outflow pathways may provide indirect information about the functional state of the trabecular meshwork and distal outflow system.

The ab-externo canaloplasty has already been combined with intraoperative channelography with fluorescein to visualize the outflow pathways^18,19,20. Eyes with more stained areas of the sclera and thus improved aqueous humor outflow showed a lower IOP one year after the procedure. Intraoperative channelography can therefore be used to predict the effect of the surgery¹⁹. To our knowledge, whether this also applies to AbiC has not yet been investigated. The aim of this study is therefore to evaluate a workflow for analyzing intraoperative channelography during AbiC. We developed this workflow to quantify the intensity and extent of subconjunctival fluorescein staining in an early and a late phase. This is the first evaluation of the effect of fluoresceinated viscoelastic in AbiC.

Subsequently, it can be investigated whether there is a correlation between the quantification of the stained areas and the extent of the individual aqueous humor outflow as well as the reduction of the IOP or IOP-lowering medication used postoperatively. It is hypothesized that a stronger staining and thus visibly higher aqueous humor outflow would also mean a greater reduction in IOP and medication use. Division into quadrants will be used to test whether the outflow is significantly altered depending on the anatomical localization.

In the long term, the presented workflow will be used to clarify whether the success of AbiC can be predicted on the basis of staining with fluorescein and can also be used for other necessary surgical treatments.

Protocol

Ethical Statement:

This study was conducted in accordance with the Declaration of Helsinki. It was considered non-interventional as it was a retrospective review of patient data collected as part of regular medical practice. No prospective therapeutic approaches were pursued, and all treatments were given routinely as part of standard care. The Bonn Ethics Committee reviewed the study protocol and waived informed consent given the non-interventional design (no. 2026-220). Patient data were de-identified prior to analysis.

1. AbiC and Intraoperative Channelography

NOTE: Perform AbiC in each eye using the microcatheter-based canaloplasty system. The following procedure describes the use of this device. Perform AbiC either as a standalone procedure or at the end of uneventful cataract surgery by phacoemulsification.

1. Position the lid speculum so that as much scleral tissue as possible is visible for subsequent analysis.
2. Apply topical anesthesia, such as proxymetacaine hydrochloride. Disinfect the eye to be operated on with iodine. Administer parabulbar anesthetics.
3. Standardize the microscope settings for all patients.
   NOTE: For AbiC, set the microscope to 20% LED light, 100% fundus reflex, and 0% surrounding light. For channelography, set the microscope to 100% LED light, 0% fundus reflex, and 100–0% surrounding light.
4. Apply the settings to the Zeiss Lumera (ophthalmic surgical microscope). Settings may vary for other microscope models.
   NOTE: Use a temporal approach, irrespective of whether the procedure is standalone or combined.
5. Create one or two 1.0 mm paracenteses 2 clock hours away from the temporal seat position, approximately 1 mm into the clear cornea.
6. Fill the anterior chamber with viscoelastic.
7. Tilt the patient’s head away from the surgeon’s side to an angle of approximately 30°. Tilt the microscope to an angle of approximately 30°. Focus on the peripheral iris tissue and bring the anterior chamber angle into view.
8. Mix the viscoelastic with fluorescein dye.
   1. Aspirate 0.05 mL fluorescein very slowly into the viscoelastic using a sterile cannula so that a single cavern is created inside the viscoelastic.
   2. Allow the fluorescein to diffuse for 5–10 min.
   3. Slowly press the plunger of the viscoelastic syringe to express the fluorescein-filled cavern completely.
9. Fill the microcatheter device with fluorescein-dyed viscoelastic.
10. Test the device by advancing the microcatheter beyond the tip of the injector device.
11. Apply hydroxy-methylcellulose to the cornea. Insert the AbiC microcatheter into the anterior chamber through the paracentesis. Place a gonioprism, such as a Swan-Jacob gonioprism lens, on the cornea.
12. Use the tip of the device to incise the trabecular meshwork. Use the blue actuator on the device to advance the microcatheter into Schlemm’s canal.
13. Advance the microcatheter 360° through Schlemm’s canal while delivering twelve 2.5 µL boluses of dyed viscoelastic by turning the visco-injector.
14. Deliver one 2.5 µL bolus per clock hour.
15. Retract the microcatheter slowly back 360° while delivering twelve 2.5 µL viscoelastic boluses.
16. Remove the microcatheter from the anterior chamber. Remove the gonioprism from the ocular surface.
17. Realign the patient’s head and the microscope to a straight position to ensure a standardized, straight view of the eye. Flush the cornea to remove dye and hydroxymethyl-cellulose.
    NOTE: If necessary, use a swab to remove any surface stains to avoid incorrect interpretation of superficial discoloration.
18. Switch to the blue light filter. Record the early phase of channelography using video or a photo series.
19. Zoom out to visualize as much scleral tissue as possible and ensure that the image is in focus.
20. Once done, switch off the blue light filter.
    CAUTION: Use bimanual irrigation/aspiration probes to remove viscoelastic from the anterior chamber. Prevent blood reflux by keeping the anterior chamber under irrigation with balanced salt solution until the paracentesis is hydrated.
21. Inject cefuroxime intracamerally.
22. Switch back to the blue light filter. Record the late phase of channelography using video or a photo series approximately 2 min after the image was taken for the early phase.
23. Zoom out to visualize as much scleral tissue as possible and ensure that the image is in focus.
24. Switch off the blue light filter and inject dexamethasone sub-conjunctivally.

2. Image Export and Data Storage

1. Download the images from your microscope software, Zeiss Forum.
2. Open the patient folder corresponding to the patient number in the microscope software and use the “Export” button to store the images in a selected folder.
3. Divide the exported images into early-phase and late-phase folders.
4. Name each image using the format “Patient No._Eye_Image_No.” and save the images in JPEG format.

3. Quantification of Subconjunctival Outflow via FIJI

NOTE: Apply this workflow to both the early and late phases of channelography.

1. Load a channelography image in FIJI²¹(open-source image analysis software). Crop the image so that only the operated eye is visible. Draw a rectangle using the “Rectangle Tool” and select “Image > Crop” from the menu.
2. Draw the white-to-white corneal diameter on the image using the “Straight Line Tool”. Measure horizontally from the nasal to lateral limbus. Set the white-to-white corneal diameter as the scale using the known corneal diameter from preoperative biometry under “Analyze > Set Scale > Known Distance (units)”.
3. Use the “Polygon Tool” to mark the boundary between the cornea and the sclera using at least 24 points. If necessary, review the surgical video or the image from the other phase to better identify the corneal border. Use landmarks such as conjunctival vessels and the base of the trabecular micro-rupture in the early phase.
4. Press “T” on the keyboard or select “Image > Overlay > To ROI Manager” to open the ROI Manager and save the corneal marking as an ROI. Save the image for later analyses, if required.
5. Press “Del” on the keyboard or select “ROI Manager > More > Fill” to cut out the cornea.
   NOTE: Ensure that the cut-out area appears black. If necessary, adjust the color using the “Color Picker” tool.
6. Convert the still image to a binary black-and-white image using the “Color Threshold” tool under “Image > Adjust”. After opening the tool, select “Default” as the thresholding method, “Black & White” as the threshold color, and “HSB” as the color space. Remove the ticks from the “Pass” boxes for each setting and from the “Dark Background” box.
7. Ensure that white areas correspond to yellow staining by fluoresceinated viscoelastic. If necessary, adjust this using the “Color Picker” tool.
8. Standardize the binarized image using one of the following three setting options to enable better comparability between images.
   1. For fully standardized settings for the entire patient collective, set the upper bar to “Hue” = 140, “Saturation” = 205, and “Brightness” = 205. Set the lower bar to “Hue”, “Saturation”, and “Brightness” = 255.
   2. For partially standardized settings, choose the upper-bar “Hue” individually, and set the upper-bar “Saturation” and “Brightness” = 255. Set the lower-bar “Hue”, “Saturation”, and “Brightness” = 255.
   3. For fully customized settings, choose each upper-bar setting individually for each image. Set the lower-bar “Hue”, “Saturation”, and “Brightness” = 255.
9. Compare the binarized image with the original image using the “stack” function or direct comparison.
10. Choose a minimal setting for the partially standardized and fully customized settings that still represents lighter stained areas while avoiding overrepresentation of strongly stained areas. If necessary, delete light reflections or superficial dye. Mark the artifact with the “Polygon Tool”, and press “Del” or select “ROI Manager > More > Fill”.
11. Select black as the fill color under “Image > Color > Color Picker” so that the cut-out areas are not incorrectly counted as dye.
12. Change the image type to “32 bit” via “Image > Type”. Normalize the settings via “Process > Enhance Contrast”, and set “% of saturated pixels = 0.35”.
13. Quantify the amount of aqueous humor outflow using approach 1 or 2 or 3 as mentioned below.
    1. Quantification Approach 1: Percentage Fluorescein Staining of the Total Area Circularly Outside the Cornea.
       ​NOTE: Use this approach to record the percentage fluorescein staining of the total area circularly outside the cornea.
       1. Depending on the image section and the width of the eyelid opening, project three rings around the previously drawn corneal demarcation, with each ring measuring 1 mm wide.
       2. Select the corneal marking in the ROI Manager and choose “Edit > Selection > Make Band: 1”.
       3. Retrieve the percentage of stained area in the first ring. Select “Analyze > Set Measurements > Mean Grey Value > OK”, and press “Ctrl + M” on the keyboard or select “Analyze > Measure”.
       4. Record the values from the table that opens.
       5. Press “T” on the keyboard or select “Image > Overlay > To ROI Manager” to save the first ring in the ROI Manager.
       6. Repeat steps 3.13.1.2 - 3.13.1.5 for the second and third rings using 2 or 3 units for the extension of the rings, starting from the corneal mark or from 1 unit of the previously drawn ring.
       7. Check the original image to confirm that the rings contain only conjunctiva.
       8. Save the binarized image under “File > Save As > Tiff”, and name it using the format “Patient ID_phase_Ring”.
       9. Save the corresponding results via “File > Save As”, and name them accordingly.
       10. Save the corresponding ROI Manager by clicking “Deselect”, then “More > Save”, and name it accordingly.
    2. Quantification Approach 2: Quadrant-Based Analysis of Percentage Staining
       ​NOTE: Use this approach to analyse the percentage staining individually for each quadrant.
       1. Complete steps 3.1–3.13.1.10 before beginning the quadrant analysis.
       2. Draw a rectangle using the “Rectangle Tool” so that it represents exactly one half of the image.
       3. Press “Del” on the keyboard or select “ROI Manager > More > Fill” so that only one half of the image remains.
       4. Draw a second rectangle representing one half of the remaining image.
       5. Press “Del” on the keyboard or select “ROI Manager > More > Fill” again so that one quarter of the original binarized image remains.
       6. Even if the selected cornea is not perfectly round, attempt to identify a quadrant that measures as close to 0.25 as possible by first using two exact halves.
       7. “Deselect” the three rings from the first approach via the ROI Manager.
       8. Click “More > Multi Measure > One Row Per Slice” to determine the stained proportion.
       9. Record the results from the table that opens.
       10. Repeat steps 3.13.2.2 - 3.13.2.9 for the remaining three quadrants.
       11. Save all binarized quadrant images under “File > Save As > Tiff”, and name them using the format “Patient ID_phase_Quadrant”.
       12. Save the corresponding results via “File > Save As”, and name them accordingly.
       13. Save the corresponding ROI Manager by clicking “Deselect”, then “More > Save”, and name it accordingly.
    3. Quantification Approach 3: Analysis of Focal Staining Patterns Corresponding to Localized Trabecular Meshwork Alterations
       ​NOTE: Use this approach to assess focal staining patterns that may correspond to localized alterations of the trabecular meshwork. During advancement in Schlemm’s canal, the catheter may induce micro-ruptures in the inner wall of the trabecular meshwork and influence herniations of the inner wall of the trabecular meshwork toward collector channel ostia. Therefore, analyse the early-phase images using this approach.
       1. Complete steps 3.1 to 3.4.
       2. Click “Edit > Make Inverse”.
       3. Press “Del” on the keyboard or select “ROI Manager > More > Fill” to delete scleral tissue for the sole analysis of trabecular micro-ruptures lying inside the corneal ring.
       4. Ensure that the area outside the corneal ring is black. If necessary, adjust the black-and-white display using the “Color Picker” tool.
       5. Draw a slightly smaller ring on the inside of the corneal ROI by first selecting the corneal ROI in the ROI Manager.
       6. Click “Edit > Selection > Enlarge > -0.2” so that a 0.2 mm narrower ring appears inside the corneal ring.
       7. Press “T” on the keyboard or select “Image > Overlay > To ROI Manager”.
       8. Click “Edit > Selection > Make Band > 0.2” so that a 0.2 mm wider ring appears.
       9. Place this ring on the corneal ring to provide the outer boundary so that only the area between the two rings can be analysed.
       10. Press “T” on the keyboard to save the ring as an ROI.
       11. Open “Image > Adjust > Color Threshold” to check the threshold settings.
       12. Use partially standardized settings. Choose the upper-bar “Hue” individually, set the upper-bar “Saturation” and “Brightness” = 255, and set the lower-bar “Hue”, “Saturation”, and “Brightness” = 255.
       13. Choose color settings that allow the color to fill the areas in the inner ring that show a trabecular break.
       14. Ensure that stained areas appear white in the binarized image.
       15. If necessary, delete artifacts and the locations where paracenteses were made. Mark the areas with the “Polygon Tool”, and press “Del”.
       16. Select black as the fill color under “Image > Color > Color Picker” so that the cut-out areas are not incorrectly counted as dye.
       17. Click “More > Multi Measure > One Row Per Slice”.
       18. Record the results from the table that opens.
       19. Save the binarized images under “File > Save As > Tiff”, and name them using the format “Patient ID_phase_microruptures”.
       20. Save the corresponding results via “File > Save As”, and name them accordingly.
       21. Save the corresponding ROI Manager by clicking “Deselect”, then “More > Save”, and name it accordingly.
14. For data compilation, open an Excel spreadsheet.
15. Copy all previously collected result tables into one spreadsheet.
16. For the quadrant analysis results, multiply every value by four so that the values represent staining in a complete image and can be compared with the other approaches.

Figure 1 provides an overview of the workflow.

Figure 1: Schematic overview of the workflow used to quantify aqueous humor outflow after ab-interno canaloplasty (AbiC). The procedure includes intraoperative fluorescein channelography image acquisition, image processing using open-source image analysis software software, binarization and standardization of images, and quantitative analysis of fluorescein staining patterns. Please click here to view a larger version of this figure.

Results

Patients over the age of 18 with a diagnosis of glaucoma who had undergone AbiC either standalone or in combination with cataract surgery were included. In addition, channelography images had to be available at least in the early or late phase. A previous selective laser trabeculoplasty was not an exclusion criterion, in contrast to other glaucoma surgical procedures. In addition, no patients with advanced glaucoma damage were included in the study.

So far, we have been able to include a total of 86 eyes from 62 patients in the analysis in the period from March 2022 to June 2024. Of these, 82 eyes underwent canaloplasty combined with cataract surgery. Here, we report exemplary results of two eyes, one with fluorescein staining and one with only limited staining (Figure 2). The first patient was 63 years old at the time of surgery and received topical IOP-lowering therapy with two active ingredients. IOP was 20 mmHg preoperatively. Fluorescein channelography immediately after AbiC showed pronounced staining in all three rings and all four quadrants. Analysis of the rings showed the highest percentage of staining in the second drawn ring with 88.1% (first ring at 87.9%, third ring at 85.2%). Examination of the individual quadrants revealed the highest percentage of staining at temporal-superior with 94.4% of dye (other quadrants at TI 93.6%, NS 93.9% and NI 66.4%). IOP was successfully lowered to 14 mmHg after 18 months and 15 mmHg after 24 months. In contrast, the second patient, who was 58 years old at the time of surgery and received 4 IOP-lowering agents with preoperative IOP of 14 mmHg, showed less staining in the two analyzed rings of fluorescein channelography with the first ring at 45.6% and the second ring at 35.1%. Furthermore, only slight and uneven staining in all quadrants was measurable (TI 64.4%, TS 23.2%, NS 31.8%, NI 42%). IOP was 14 mmHg after 18 months and rose to 25 mmHg after 24 months. In patient 1, topical therapy could be discontinued directly after surgery. Both patients did not receive any IOP-lowering medication up to 18 months post-surgery.

Figure 2: Representative workflow for quantitative analysis of intraoperative fluorescein channelography in two eyes. (A,D) Original intraoperative channelography images showing extensive fluorescein staining (A) and limited fluorescein staining (D) after AbiC. (B,E) Corresponding binarized images after image processing and threshold adjustment in open-source image analysis software. (C,F) Subdivision of the binarized images into quadrants for quantitative analysis of fluorescein staining. In panel C, the quantified staining areas were temporal inferior (TI) 94%, temporal superior (TS) 94%, nasal inferior (NI) 66%, and nasal superior (NS) 94%. In panel F, the quantified staining areas were TI 64%, TS 23%, NI 42%, and NS 32%. Please click here to view a larger version of this figure.

In addition to evaluating the staining according to quadrant, trabecular micro ruptures were also examined. As shown in Figure 3, 67.8% of micro ruptures can be clearly distinguished. In the corresponding patient, there was a reduction in IOP from 23 mmHg preoperatively to 18 mmHg postoperatively after 12 months. Furthermore, the IOP-lowering topical therapy could be discontinued.

Figure 3: Analysis of focal staining patterns corresponding to trabecular micro ruptures. (A) Original early-phase intraoperative channelography image showing focal fluorescein staining patterns at the anterior chamber angle. (B) Corresponding binarized images used for quantification of focal staining patterns. Binarization facilitates visualization and quantification of the stained regions. Please click here to view a larger version of this figure.

To test the reproducibility of the workflow, intra- and inter-rater reliability were examined. Intra-rater reproducibility was assessed by re-analyzing the second ring in a subset of 15 eyes after a 4-month interval using the same standardized workflow. Inter-rater reproducibility was assessed by having a second independent examiner analyze the same subset of 15 eyes using the identical standardized workflow in open-source image analysis software. Both examiners were masked to each other’s results.

Agreement was evaluated using the intraclass correlation coefficient (ICC) based on a two-way random-effects model for absolute agreement. The intra-rater reproducibility analysis demonstrated excellent agreement, with ICC value of 0.95. Inter-rater reproducibility demonstrated good agreement, with an ICC of 0.81.

Discussion

This workflow provides a step-by-step guide to use open-source image analysis software to quantify the amount of aqueous humor outflow after AbiC. The features of open-source image analysis software provide easy-to-use templates that do not require any coding knowledge and can be applied by physicians without technical support. This approach will help to evaluate aqueous humor outflow after AbiC to assess its potential predictive value for surgical outcomes. Rather than representing a direct functional measurement, the workflow is designed to systematically characterize intraoperative fluorescein distribution patterns and enable reproducible comparisons between cases. Particular attention should be paid to microscope alignment, standardized illumination settings, image centering, and accurate corneal border definition, as these factors strongly influence reproducibility of the quantitative analysis. Our aim on the long term is to determine whether a greater postoperative reduction in IOP is associated with higher fluorescein outflow during channelography.

Over the years, a number of studies have focused on the imaging of the outflow system using fluorescein^19,22,23,24. Injected directly into Schlemm's canal (in our case during an AbiC) has the advantage that only small amounts of fluorescein are required and 360° of the outflow pathways can be visualized in contrast to a filling of the whole anterior chamber. This concept was first introduced by Grieshaber et al., who demonstrated intraoperative fluorescein channelography during canaloplasty and visualized episcleral venous drainage as well as segmental outflow patterns, highlighting marked interindividual variability of aqueous humor outflow²³. While these studies primarily provided qualitative insights, they established the foundation for interpreting fluorescein distribution as a reflection of outflow pathways^23,25. The potential role of fluorescein channelography as a functional biomarker rather than a purely descriptive imaging tool is largely unexplored to date. Previous studies combining channelography with provocative gonioscopy demonstrated that poor episcleral venous filling and limited collector channel recruitment were associated with higher preoperative IOP and less favorable postoperative outcomes, suggesting that functional outflow imaging may reflect the underlying drainage capacity of the eye²³. The present workflow extends these qualitative observations by providing a standardized quantitative image-analysis approach for intraoperative fluorescein channelography.

Quantitative studies of aqueous humor angiography have further demonstrated that outflow is highly segmented even in normal eyes. A distinct circular gradient has been described, with the highest signal intensity in the nasal quadrant and the lowest in the temporal quadrant, showing statistically significant regional differences²⁶. At a finer scale, sectoral analyses revealed up to a twofold difference between regions of high flow (nasosuperior) and low flow (temporoinferior)²⁷. Similar patterns have also been observed in glaucomatous eyes, supporting the concept that segmental outflow persists in disease states and may contribute to pathological resistance. In addition, surgical studies indicate that aqueous angiography can detect localized increases in outflow following intervention, although these changes have largely been described qualitatively to date²⁸. The use of image binarization and quadrant-based quantification enables objective comparison of staining distribution patterns between eyes and between different surgical phases.

More recent aqueous angiography studies used fluorescein introduced tracer-based imaging under near-physiological conditions and confirmed that aqueous humor outflow is segmental, dynamic, and patient-specific rather than uniformly distributed across quadrants^25,29. These studies further demonstrated that angiographically positive regions correlate with episcleral venous drainage and intrascleral lumina identified by optical coherence tomography (OCT), providing structural validation of the functional imaging signal. Vaiculiene et al. already pointed out that good collector's channel function is required for successful IOP lowering after ab-externo canaloplasty²⁰. Grieshaber also was able to show that the IOP-lowering effect was lower if no fluorescein appeared in the anterior chamber during canaloplasty using fluoresceinated balanced salt solution. He attributed this to impaired trabecular permeability¹⁸. Additional mechanisms such as structural alterations of the trabecular meshwork and inner wall of Schlemm’s canal following vasodilation have been proposed, although their functional relevance remains incompletely understood^22,30. Experimental aqueous angiography models support this hypothesis, as targeted surgical manipulation of the trabecular outflow pathway has been shown to recruit previously non-perfused outflow regions, resulting in increased fluorescein signal in formerly angiographically negative areas²⁹.

From a methodological perspective, the reliability of this workflow strongly depends on standardized image acquisition and processing. AbiC has led to a significant reduction in IOP and the postoperative need for IOP-lowering eye drops compared to before surgery^17,31. Standardized channelography analysis may therefore help to better understand interindividual differences in surgical response after AbiC.

Despite our efforts to establish a standardized image acquisition and analysis workflow, there are some limitations to consider. Firstly, images and videos recorded during the surgery are a two-dimensional representation of a real three-dimensional situation. Any perspective distortions caused by this cannot be ruled out. As already described in the protocol, the same microscope settings should be used for each eye in order to create the most uniform imaging conditions possible. Varying settings, especially with different surgeons, can lead to differing images, which limits the quality of the analysis. In particular, variations in illumination, focus, and camera alignment can significantly influence fluorescence intensity and spatial distribution. To minimize variability, the same microscope settings and acquisition angles should be maintained throughout all procedures and analysis sessions.

In some cases, the eye is not properly centered during surgery, resulting in incomplete capture of the fluorescein outflow areas. As a consequence, a full evaluation of the outflow is not possible. This issue can introduce bias and is further limited by the lid width and the presence of a name box at the bottom of the images, especially when assessing the temporal superior (TS) and temporal inferior (TI) quadrants. Unfortunately, this leads to the fact that a usable photo or video of the early and late phase is not available for every patient, which limits the informative value of the data. Images with tilted eyes cannot be analyzed, as the white-to-white (WTW) distance cannot be used reliably as a scale in these cases. An incomplete display of the quadrants leads to the fact that not all analysis rings can be placed when using the image. This reduces the informative value regarding fluorescein outflow distribution. The same applies to the analysis of the individual quadrants. If one phase shows an obvious superficial staining, while the other phase clearly shows a deeper staining, the affected area should not be excluded from the analysis but recognized as overrepresented for one phase (as it may show the effect of dilatation)²⁵.

Another problem is that in many images the cornea is not perfectly round, making it difficult to determine the exact center. This leads to slight deviations in the quadrant distribution, with the range for the (control) quadrants of the cornea being between 0.24 and 0.26. As a further consequence, the newly defined outer analysis rings are also not completely round, which could also affect the measurement accuracy. In addition, the corneal border cannot always be defined beyond doubt due to insufficient image quality. For better orientation, it is therefore advisable to use conjunctival vessels and early-phase focal staining patterns as landmarks²⁵. Comparing the early and late phases and, if possible, viewing the corresponding surgical videos can contribute to a more precise localization. Future automation of corneal border detection and image registration may further improve reproducibility and reduce observer-dependent variability.

Ultimately, strong staining or completely stained rings lead to calculated results exceeding 100%. This may be due to the hypothesis that stronger staining increases the risk of overestimation, particularly if the quadrant distribution cannot be set exactly to 0.25. At the same time, areas with complete staining may make it easier to identify bias compared to areas that are only partially stained²³. It should also be noted that the catheter releases a small amount of dye during insertion through the oblique paracentesis and the incision in the trabecular meshwork. This additional dye must be removed when analyzing the focal staining patterns, with the amount varying from eye to eye. This means that comparability is limited, as areas with actual staining could be mistakenly removed if they are covered by the additional dye. Therefore, careful interpretation of focal staining patterns is required, particularly in eyes with extensive superficial fluorescein distribution.

As this workflow is based on open-source software, we would like to encourage others to reproduce our work to determine aqueous humor outflow and expand our knowledge of how other minimally invasive glaucoma surgical procedures work. To summarize, we present a workflow for determining intraoperative fluorescein outflow patterns during AbiC, which in the future may allow prediction of the effectiveness of the procedure performed. By focusing on standardized acquisition, structured analysis, and transparent reporting of limitations, this approach may support the development of quantitative imaging biomarkers. This workflow paves the way towards a new biomarker for success or failure especially of minimally invasive glaucoma surgery. The presented workflow may also be adaptable to the evaluation of aqueous humor outflow in other minimally invasive glaucoma surgical procedures.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Balanced Salt Solution	Bausch & Lomb	Ref.: BL705004
Bi-manual irrigation/aspiration probes	Zeiss	Ref.: 303060-0165-000
Cefuroxim	Laboratoires Théa	PZN: 09755705	Pre-mixed by the pharmacy
Dexamethasone	Dr. Winzer	PZN: 02192162
FIJI (ImageJ)	n/a	n/a	Open-source image analysis software
Fluorescein dye	Alcon	PZN: 01467007
Healon Pro Viscoelastic	Johnson & Johnson	Ref.: 10-3100-12
Hydroxymethylcellulose	Johnson & Johnson	Ref.: 10310012
Iodine solution	Mundipharma	PZN: 03930490
iPrism S gonioprism lens	Glaukos	Ref.: OTB/LH-2
iTrack Advance Canaloplasty System	Nova Eye Medical	Ref.: Track-ADS	Microcatheter-based canaloplasty system
Lumera 700 Surgical Microscope + Calysto	Zeiss	Ref.: 6726521526	Ophthalmic surgical microscope
Proxymetacaine hydrochloride	PUREN Pharma GmbH & Co. KG	PZN: 11055866
Sterile cannula	Braun	Ref.: 4657683