Trabeculopuncture as a predictive test of distal outflow resistance in canal-based surgery

We investigated whether trabeculopuncture (TP) could detect distal outflow resistance to predict the outcome of canal-based glaucoma surgery such as ab interno trabeculectomy (AIT). These procedures have a high utilization in open angle glaucoma, but fail in eyes with an unidentified distal outflow resistance. We assigned 81 porcine eyes to two groups: trial (n = 42) and control (n = 39). At 24 h, four YAG-laser trabeculopunctures were placed nasally, followed by a 180° AIT at the same site at 48 h. The proportion of TP responders between both AIT groups was compared. Histology and outflow canalograms were determined. Both post-TP and post-AIT IOPs were lower than baseline IOP (p = 0.015 and p < 0.01, respectively). The success rates of TP and AIT were 69% and 85.7%, respectively. Sensitivity and specificity values of TP as predictive test for AIT success were 77.7% and 83.3%, respectively. The positive and negative predictive values were 96.6% and 38.5%, respectively. We conclude that a 10% reduction in IOP after TP can be used as a predictor for the success (> 20% IOP decrease) of 180° AIT in porcine eyes.

We investigated whether trabeculopuncture (TP) could detect distal outflow resistance to predict the outcome of canal-based glaucoma surgery such as ab interno trabeculectomy (AIT). These procedures have a high utilization in open angle glaucoma, but fail in eyes with an unidentified distal outflow resistance. We assigned 81 porcine eyes to two groups: trial (n = 42) and control (n = 39). At 24 h, four YAG-laser trabeculopunctures were placed nasally, followed by a 180° AIT at the same site at 48 h. The proportion of TP responders between both AIT groups was compared. Histology and outflow canalograms were determined. Both post-TP and post-AIT IOPs were lower than baseline IOP (p = 0.015 and p < 0.01, respectively). The success rates of TP and AIT were 69% and 85.7%, respectively. Sensitivity and specificity values of TP as predictive test for AIT success were 77.7% and 83.3%, respectively. The positive and negative predictive values were 96.6% and 38.5%, respectively. We conclude that a 10% reduction in IOP after TP can be used as a predictor for the success (> 20% IOP decrease) of 180° AIT in porcine eyes.
Intraocular pressure (IOP) reduction is the only treatment for glaucoma 1 demonstrated to be effective with high-quality level I evidence [2][3][4] . Selective laser trabeculoplasty, now a recommended first line of treatment for most open angle glaucomas 5 , and medications may achieve the desired IOP levels in many patients, but about 50% still require surgery eventually 6 . Traditional glaucoma filtering surgeries 7 lower the IOP by bypassing the trabecular meshwork (TM) and draining the aqueous humor into a newly created epibulbar space 8 , but they are associated with a high rate of complications and require intensive postoperative care 9,10 .
Canal-based minimally invasive glaucoma surgeries produce far fewer complications and allow intervening earlier 11,12 because they lower the IOP by bypassing or removing the TM to enhance the physiological aqueous humor outflow route 11,13,14 . Leading modalities are trabecular bypass stents (TBS) [15][16][17] and ab interno trabeculectomy (AIT), in which the TM is either ablated 18 , incised 19 , or excised 20 . However, approximately 30% of patients experience an insufficient IOP reduction 21 . One would expect the IOP to be close to the level of episcleral venous pressure, approximately 8 mmHg, but this pressure can rarely be achieved 21 due to increased post-trabecular resistance 22,23 . So far, there is no presurgical test that could assess the post-trabecular resistance predict the outcome of AIT.
A noninvasive procedure that allows aqueous to bypass the proximal outflow resistance at least temporarily is Nd:YAG laser-assisted trabeculopuncture (TP). Not unlike trabecular bypass stents, TP creates a focal opening through the trabecular meshwork (TM) and the inner wall of Schlemm's canal (SC) 24 . In 1985, Epstein et al. investigated TP as a treatment for glaucoma, but the subsequent IOP reduction was short-lived 24 .
Here, we hypothesized that TP could be used to assess the distal outflow tract function before AIT or TBS are considered. We deployed our porcine anterior chamber ex vivo perfusion model 25,26 to develop a predictive test and address this unmet need.

Methods
Study design. In total, 81 hemisected, perfused porcine eyes were assigned to one of two groups: trial (T) (n = 42) and control (C) (n = 39). Eyes in the trial group underwent trabeculopuncture using a Nd:YAG laser 24 h after incubation, followed by ab interno trabeculectomy a day later. The IOP was measured continuously for 72 h, with baseline values being recorded 24 h (IOP BL ) after the start of the experiment. Post-trabeculopuncture IOP (IOP TP ) was measured at 48 h, and post-ab interno trabeculectomy IOP (IOP AIT ) at 72 h. The estimated outflow facility (C) was calculated by dividing the medium inflow by the IOP value. The fluid inflow simulated aqueous humor formation and was constantly at 6 µl/min. Statistical analysis. Our sample size calculation indicated a minimum requirement of 35 eyes per group to achieve a testing power of 0.9. We analyzed data with SPSS Statistics (Version 26, IBM, New York, USA). Means and standard deviations were reported for all parameters. We tested for normal distribution with the Kolmogorov-Smirnov test and used a paired t-test or a Wilcoxon Signed Rank test to compare dependent means; an unpaired t-test or Mann-Whitney U test was used for independent means. One-way repeated measures multivariate analysis of variances (MANOVA) was used to compare more than two means. The Pearson Coefficient was reported for the correlation between continuous variables. TP and AIT success was defined as a decrease of 5% and 10% from baseline IOP, respectively. We used Fisher's exact test to compare the number of eyes that responded to TP and AIT to the number of eyes that responded to TP but failed AIT. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated. After TP and AIT success, a receiver operating characteristic curve (ROC curve) was plotted for IOP reduction. A p value of 0.05 or less was considered statistically significant for all tests.

Results
Pilot experiments with whole eyes showed that a transcorneal TP using a Ritch trabeculoplasty lens could not be accomplished reliably. Therefore, anterior segments were inverted to laser the TM directly. The TM could be readily identified ( Fig. 1, BL1). The procedure resulted in small, well-circumscribed pits of approximately 500 µm in length and 250 µm in depth ( Fig. 1, TP1). We detected no obvious damage to adjacent tissue upon inspection with a microscope. AIT removed the TM extensively, leading to a narrow continuous groove along the nasal quadrants ( Fig. 1, AIT1). After both procedures, the findings on histological sections ( Fig. 1, middle row) corresponded well to our observations through an operating microscope. These were further supported by canalograms (Fig. 1, BL3), which illustrated an improved localized outflow after TP (Fig. 1, TP3) and further increased sectoral outflow in the nasal quadrants and adjacent drainage segments after AIT (Fig. 1, AIT3).
Eighty-one eyes were analyzed: 42 trial eyes (T) and 39 controls (C). Table 1 depicts the IOP parameters of both groups. Baseline IOP (IOP BL ) was 16.4 ± 4.5 mmHg in T eyes and 15.2 ± 3.9 mmHg in C eyes. There was no difference between both variables (p = 0.37). In the experimental group, mean IOP TP and IOP AIT values were 14.6 ± 4.3 mm Hg and 11.3 ± 3.8 mmHg, respectively. Both values were lower than IOP BL (p = 0.02 and p < 0.001, for IOP TP and IOP AIT , respectively). We found the three IOP measurements (IOP BL , IOP TP, and IOP AIT ) to be different from each other (p < 0.001). Figure 2 illustrates the mean IOP levels of all three groups and their respective average post-procedure IOP drops. The average IOP reduction from IOP BL after TP and AIT was 8.7 ± 22.4% and 28.8 ± 22.8%, respectively. There was a weak positive correlation between the amount of IOP reduction from IOP BL after TP and AIT (r = 0.37, p = 0.015). Control eyes had an IOP that was 2.0 ± 1.3 mmHg higher at the end of the perfusion studies (p > 0.05 for both 24 h and 48 h). Figure 3 plots all 42 eyes of the trial group according to IOP change after TP and AIT, respectively. After TP and AIT success, it visualizes that four out of six non-responders to AIT show an IOP increase after TP. www.nature.com/scientificreports/  www.nature.com/scientificreports/    (Fig. 4). A subanalysis (Supplementary Information, Table 1) showed a mean IOP BL of 17.1 ± 4.4 mmHg and 11.9 ± 2.7 mmHg for AIT responders and AIT non-responders. These values differed significantly (p = 0.008). There was no difference in IOP TP and IOP AIT in both subgroups (p = 0.48 and p = 0.45, Supplementary Table 1). 1Out of all 13 TP non-responders, 76.9% (n = 10) showed an IOP increase after TP of at least 10%.

Discussion
In this paper, we used a porcine anterior segment model to assess the utility of TP in predicting the success of AIT. The lack of good predictive tests combined with the relatively high rate (30%) of canal-based surgeries and implants led us to explore simple options to avoid unnecessary surgeries. Porcine eyes have been used as a model to study glaucoma extensively and were used here for that reason 29,30 .
We expected TP to cause an IOP drop in our porcine anterior segment perfusion model, similar to that reported in human eyes. TP is not unlike trabecular bypass stents, which also cause a focal opening in the TM and increase outflow in our model 17 . The total energy used was higher than what is normally required in humans due to an at least three times thicker TM compared to human eyes 30 . After AIT, a further decrease in IOP was seen because of the comprehensive excision of the nasal TM. Compared to a study by Dang et al., which tested outflow enhancements of three different AIT devices on porcine eyes, we found similar baseline IOP values (16.35 ± 4.52 mmHg vs. 15.93 ± 2.08 mmHg) 26 . Still, we observed a post-AIT IOP decreased by only 31%, in contrast to the 48% reported in that study 26 .
The control group in our cohort experienced a small IOP increase of approximately 13% over 72 h. This is in line with Dang et al., who observed a 10% IOP increase in control eyes during 72 h of incubation 31 . To adjust for this, we chose 5% and 10% to be satisfactory post-TP and post-AIT IOP reductions, respectively. These values correspond to a 10% and 20% IOP reduction after these procedures. Clinically, a 20% post-AIT IOP reduction is often regarded as sufficient for mild to moderate glaucoma 32 . Being able to predict this outcome will help avoid unnecessary procedures and decrease the burden on the healthcare system.
Interestingly, our subanalysis ( Supplementary Information, Table 1) revealed that AIT responders had a higher baseline IOP compared to non-responders. This is perhaps to be expected, as AIT generally caused a greater decrease in eyes with higher baseline IOPs in clinical studies 33 . After TP, AIT non-responders also had a higher mean IOP resembling a decreased outflow facility, respectively, which did not reach significance with the number examined here. It is possible that the collapse of laser-induced trabeculopunctures temporarily decreases the outflow facility as described before 34,35 , which could similarly affect our TP non-responders if their TM was compromised in the area of the TP and AIT. Additionally, there was no difference between baseline and post-AIT www.nature.com/scientificreports/ IOP levels in AIT non-responders, which is indicative of a post-trabecular meshwork resistance in these eyes. These were not glaucomatous eyes, however. Ocular hypertension can be induced experimentally in pig eyes 36 but does not occur in pigs naturally. We suspect that inadvertent compression of key elements of the distal outflow tract in the nasal quadrants by the compression ring of the perfusion mount is responsible for this. This would not necessarily lead to an increased IOP because at least 3/4th of the outflow tract has to be compromised 37 , but it can explain the failure to respond to TP and AIT. However, one has to be careful interpreting the results of this subanalysis, as there were only six AIT non-responders in our study with a significantly lower baseline IOP than AIT responders. A simple and noninvasive predictive test for canal-based surgeries that ablate, excise, disrupt or bypass the TM is urgently needed because of the rapidly increasing demand for these procedures. The implementation of a Nd:YAG laser-TP for such a test is straightforward as this device is ubiquitously available in ophthalmology practices and clinics, and most ophthalmologists are familiar with its use 38 . Although the effect of TP is too short-lived to be useful for glaucoma treatment 39,40 , it is precisely this benign nature that may afford a low-risk test of distal outflow resistance.
The amount of IOP reduction after TP and AIT had a relatively weak direct correlation. This is not surprising and matches the clinical reality that AIT will lower IOP, not by a certain percentage that can be generalized regardless of baseline IOP. Instead, AIT lowers IOP to a pressure level defined by residual, post-trabecular outflow resistance and independent of baseline IOP. For instance, both an eye with a pressure of 35 mmHg and another eye with a pressure of 20 mmHg can have a similar postoperative IOP because the primary resistance at the level of the trabecular meshwork is removed. Therefore, the weak correlation in our study does not undermine the utility othanpredictive test with a high PPV for the success of AIT. Moreover, our sensitivity and specificity values of 77% and 83% are sufficient for a clinical test. However, our data did not show a high negative predictive value (NPV, 38.5%) in porcine eyes because these are non-glaucomatous eyes. The NPV in human eyes should be higher, matching the AIT failure rate caused by a presumed higher rate of post-trabecular resistance than in pigs.
One limitation of our study is the ex-vivo setting. Hence, wound healing of the TM and its effect on IOP cannot be observed. Another limitation is the anatomical difference between porcine eyes and human eyes. In porcine eyes, the outflow tract consists of an angular aqueous plexus, whereas humans have a Schlemm's canal, often with a single lumen 41,42 . We used four evenly spaced TPs over the nasal angle to cover the extent of an AIT and to account for the decreased circumferential flow compared to a human Schlemm's canal. Clinically, blood reflux from SC can normally be seen after a TP, a useful indicator of completion absent in an ex vivo model. Instead, we had to use the IOP decrease and an increased outflow of fluorescent beads as an indicator. We did not perform canalograms with bead tracers on eyes whose IOP was obtained because they can progressively obstruct distal outflow channels and alter IOP measurements. Therefore, we could not compare the outflow patterns of responders and non-responders. Our study was not designed, and hence not powered, to discover a cause of AIT failure. It could be possible that a post-trabecular outflow resistance in the form of the compression mount of the culture dish causes the inability to achieve an improved facility by TP and AIT. For instance, a decentered mount may compress a part of the distal outflow system in the area treated with TP and AIT but not affect other parts of the circumference. As a result, the outflow at baseline would not be impaired because more than 75% of the TM has to be blocked in this model for IOP to rise 37 . On the other hand, debris created by TP or AIT may slightly reduce outflow away from the treated site precisely as we have observed in those eyes but not in others. Investigating the mechanism is relevant to past and future studies using perfused anterior segments and requires a study specifically designed to test this hypothesis.
In conclusion, a 10% IOP reduction after trabeculopuncture can be used to predict a successful ab interno trabeculectomy in porcine eyes.

Data availability
Data is readily available from the corresponding author on reasonable request. www.nature.com/scientificreports/