Needle bevel–guided standardisation of scleral tunnel formation in intrascleral haptic fixation: a quantitative geometric analysis

Abstract

Background

Scleral tunnel formation during Yamane double-needle intrascleral haptic fixation (ISHF) depends on proprioceptive estimation, producing variable tunnel lengths that contribute to intraocular lens (IOL) malalignment. The aim of this study was to evaluate a needle bevel–guided technique using the innate 2 mm bevel of a 26-gauge needle as a visual endpoint for standardising intrascleral tunnel length, and to assess postoperative IOL positional stability using a quantitative geometric landmark framework.

Methods

Retrospective consecutive case series of 24 eyes undergoing ISHF in the horizontal meridian (3- and 9-o’clock positions) by a single surgeon. During scleral needle advancement, the bevel disappearance point was used to standardise tunnel length to approximately 2 mm at each fixation site. IOL centration was quantified by extracting limbus centre (L) and optic centre (O) from intraoperative and 6-month images, and flange coordinates (F1, F2) from intraoperative images only using Fiji/ImageJ (National Institutes of Health, Bethesda, MD, USA; version 2.14.0/1.54f). The primary outcome was longitudinal Euclidean displacement of the optic centre.

Results

Mean best-corrected visual acuity improved from 1.28 ± 0.54 to 0.52 ± 0.47 logMAR (p < 0.001). Mean IOL decentration was 0.401 ± 0.240 mm at baseline and 0.397 ± 0.234 mm at 6 months (mean change − 0.004 mm; p = 0.457). Mean Euclidean displacement was 0.029 ± 0.018 mm; no eye demonstrated displacement exceeding 0.06 mm. No haptic extrusion, IOL dislocation, or sight-threatening complications occurred.

Conclusions

The needle bevel provides a simple, reproducible, and cost-free visual cue for standardising intrascleral tunnel length during ISHF and was associated with excellent longitudinal positional stability. Initial centration, governed by entry angle and chord length symmetry, remains a separate challenge requiring complementary strategies.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12886-026-04980-w.

Background

The Yamane double-needle intrascleral haptic fixation (ISHF) technique is widely adopted for sutureless intrascleral fixation of intraocular lenses (IOLs) in eyes without capsular support [1–4].

However, scleral tunnel formation remains dependent on proprioceptive estimation, which is influenced by surgeon experience, hand mechanics, and scleral biomechanics [5–7]. IOL tilt after flanged intrascleral fixation has been associated with measurable refractive error [2], and tunnel length has been shown to correlate directly with positional stability, with shorter tunnels producing less stable haptic flanging [3, 4].

Recent efforts to address this variability have focused on technology-dependent solutions. Digital microscope overlays [5], image-guided navigation systems [6], and high-resolution surgical video analysis [7] have each been proposed to standardise needle entry and tunnel creation during ISHF, but they require expensive equipment, specialised software, or post hoc analysis unavailable in most settings. A simple, cost-free method for standardising scleral tunnel length has not been described.

The bevel of a needle used in ISHF has an inherent length of approximately 2 mm. During scleral advancement, the bevel tip transitions from a sharp, visible profile to a visibly rounded contour under the operating microscope as it is progressively buried within the scleral stroma. This transition point—the disappearance of the bevel—represents a reproducible visual endpoint that can serve as an intraoperative cue for standardising intrascleral tunnel length. Unlike proprioceptive estimation, this visual landmark is independent of surgeon experience and patient biomechanics, and it requires no additional instruments or cost. Nevertheless, bevel visibility is not invariably optimal under the operating microscope; conjunctival haemorrhage, heavily pigmented sclera, oblique illumination, or reduced microscope magnification may render the bevel transition less distinct. In such situations, proprioceptive judgement and surgical experience remain indispensable complementary cues. In addition, stabilisation of the globe with a spatula placed at the lateral corneal entry point during scleral tunnel formation can provide a tactile reference for tunnel progression and improve the reproducibility of the bevel-disappearance endpoint; this manoeuvre is recommended as an adjunct to the visual cue.

The clinical relevance of IOL position after scleral fixation has been well established; even modest decentration may impair visual quality through induced higher-order aberrations and residual astigmatism [8, 9]. Objective, quantitative assessment of IOL centration is therefore essential when evaluating any standardisation technique [10–12].

This study aimed to introduce a needle bevel–guided approach for standardising scleral tunnel formation during ISHF and to assess postoperative IOL positional stability using a quantitative geometric landmark framework over 6 months.

Methods

Study design and participants

This was a retrospective consecutive case series conducted at a single centre by a single surgeon. The study period extended from January 2024 to June 2025. The study protocol was approved by the Institutional Review Board (2025/08–017) and adhered to the tenets of the Declaration of Helsinki. Written informed consent was obtained from all participants.

During the study period, 55 eyes of consecutive patients underwent double-needle flanged ISHF using the bevel-guided technique. From this consecutive series, 24 eyes met all predefined image-quality and analytic-imaging eligibility criteria and constituted the analytic cohort for quantitative geometric analysis (Table 1); the remaining 31 eyes were excluded for reasons detailed in the “Image acquisition and eligibility” subsection below. Among the 24 analysed eyes, 10 were aphakic and 14 had pre-existing IOL subluxation or dislocation; indications were aphakia following posterior capsule rupture (n = 5), dense cataract aphakia managed without primary IOL implantation (n = 5), crystalline lens subluxation related to pseudoexfoliation syndrome (n = 6), anterior IOL dislocation (n = 4), and posterior IOL dislocation or drop (n = 4). Exclusion criteria for the analytic cohort, with the number of eyes excluded by each criterion shown in brackets, were: (i) intraoperative or 6-month images failing to meet the predefined quality criteria (poor focus, motion artefact, specular obstruction of > 30° of the limbal or optic boundary, or insufficient contrast for landmark tracing) [n = 22]; (ii) intraoperative images on which both flange centres could not be reliably outlined [n = 9]; (iii) loss to follow-up before the 6-month visit [n = 0]; and (iv) significant ocular comorbidity present at baseline that would confound geometric or visual outcome assessment (advanced corneal opacity, end-stage glaucomatous optic neuropathy, or active macular pathology unrelated to the indication for fixation) [n = 0]. All 31 excluded eyes were therefore excluded on the basis of image-acquisition or image-quality factors (criteria i and ii); no eyes were excluded on patient- or disease-level grounds (criteria iii and iv).

Table 1.

Patient Demographics and Baseline Characteristics (N = 24 Eyes)

Characteristic	Total (N = 24)	Aphakia (n = 10)	IOL Subluxation/Dislocation (n = 14)
Age (years), mean ± SD	68.9 ± 7.1	67.1 ± 6.3	70.1 ± 7.5
Range	55–80	55–77	57–80
Sex, n (%)
Male	12 (50)	5 (50)	7 (50)
Female	12 (50)	5 (50)	7 (50)
Laterality, n (%)
OD	12 (50)	5 (50)	7 (50)
OS	12 (50)	5 (50)	7 (50)
Axial length (mm), mean ± SD	23.3 ± 0.6	23.4 ± 0.6	23.3 ± 0.6
Mean keratometry (D), mean ± SD	43.7 ± 0.6	43.6 ± 0.6	43.7 ± 0.6
Pre-op BCVA (logMAR), mean ± SD	1.28 ± 0.54	1.80 ± 0.37	0.91 ± 0.27
Indication subtype, n
Post-PCR aphakia	5	5	—
Dense cataract aphakia	5	5	—
Subluxation (PXF)	6	—	6
Anterior dislocation	4	—	4
Posterior dislocation/drop	4	—	4
Comorbidities, n (%)
PXF syndrome	6 (25)	0 (0)	6 (43)
Diabetes mellitus	6 (25)	4 (40)	2 (14)
Hypertension	5 (21)	2 (20)	3 (21)

BCVA = best-corrected visual acuity; D = diopters; IOL = intraocular lens; OD = right eye; OS = left eye; PCR = posterior capsule rupture; PXF = pseudoexfoliation; SD = standard deviation

Surgical technique

Throughout this study, scleral tunnel length denotes the linear distance the needle travels through stromal tissue—the quantity standardised here to approximately 2 mm using the bevel endpoint—and is distinguished from scleral tunnel depth, which represents the perpendicular distance from the outer scleral surface to the needle at any given point along its oblique trajectory; because this depth varies continuously as a function of needle angle and local scleral thickness, it cannot be assigned a fixed intraoperative value. While tunnel length is standardised visually via the bevel, adequate mid-stromal depth must still be maintained via standard tactile and visual feedback to prevent late conjunctival erosion.

All procedures were performed using the double-needle flanged ISHF technique through a transconjunctival approach [1]. A 2.8 mm main corneal incision was created at the superior limbus (12-o’clock position), and the two diametrically opposed scleral fixation sites were placed in the horizontal meridian (3- and 9-o’clock positions). Vitrectomy approach was selected according to the clinical indication: in 19 (79%) eyes a 25-gauge transconjunctival pars plana vitrectomy was performed (EVA NEXUS phacovitrectomy system; D.O.R.C. International B.V., Zuidland, The Netherlands), with continuous balanced salt solution infusion supplied through a pars plana trocar throughout the fixation procedure to maintain stable intraocular pressure during scleral tunnel formation and haptic externalisation; in the remaining 5 (21%) eyes, in which the indication did not require posterior segment intervention, anterior vitrectomy was performed through the corneal incision with infusion supplied through an anterior chamber maintainer.

For scleral fixation, two 26-gauge hypodermic needles (26G 0.45 × 13 mm; Berika Teknoloji Medikal, Konya, Türkiye) were used. Although the original Yamane technique was described using 30-gauge thin-wall needles [1], 26-gauge needles were preferred in the present series because the larger lumen reliably accommodates the polymethyl methacrylate (PMMA) haptics of three-piece IOLs without the need for haptic trimming or aggressive shrinkage. The rationale for, and trade-offs of, this gauge selection are discussed further below. At each of two diametrically opposed fixation sites (separated by 180°), a first scleral reference mark was placed 2 mm posterior to the limbus. The 2 mm limbus-to-fixation distance was selected in accordance with the original Yamane description [1] to optimise the trade-off between iris-capture risk (which increases with shorter distances, particularly in eyes with concomitant vitreoretinal pathology) and conjunctival erosion risk (which increases with greater distances and may be exacerbated by haptic protrusion in thin or aged sclera). A 2.5 mm distance may be considered in eyes with prior pars plana surgery or shallow anterior segments, but evidence directly comparing 2.0 mm and 2.5 mm fixation in flanged ISHF remains limited. A second mark was placed approximately 2 mm from the first at an angle of approximately 100°, defining the needle entry point.

At each fixation site, the needle was introduced through the second mark and advanced within the scleral stroma to create an intrascleral tunnel. During scleral advancement, the surgeon used the needle bevel as a visual cue to standardise tunnel length. The bevel of the 26-gauge needle measures approximately 2 mm in length. As the needle advances through the scleral stroma, the bevel transitions from a sharp, clearly visible profile to a visibly rounded contour under the operating microscope. When the bevel was no longer distinctly visible—indicating that the full bevel length had been buried within the sclera—the needle was redirected toward the centre of the globe to enter the ciliary sulcus (Fig. 1). This bevel disappearance endpoint was applied at both fixation sites to achieve symmetric tunnel formation.

Fig. 1.

Needle bevel as a visual cue for standardising intrascleral tunnel length during ISHF. (A) Microscopic view of a 26-gauge needle (Berika Teknoloji Medikal, Konya, Türkiye) against a metric ruler, demonstrating a bevel length of approximately 2 mm. Scale bar indicates 1 mm divisions. (B) Intraoperative anterior segment image with geometric annotations showing the intrascleral tunnel length of approximately 2 mm (cyan line, scleral entry to tunnel exit) and the corresponding bevel length of approximately 2 mm (orange line). Scale bar = 1 mm. Abbreviations: IOL, intraocular lens

A Sensar AR40e three-piece intraocular lens (Johnson & Johnson Vision, Santa Ana, CA, USA) was inserted through the corneal incision in all cases. The leading and trailing haptics were externalised through the respective needle lumens using the needle-docking technique. Each externalised haptic tip was cauterised using low-temperature cautery to create a terminal flange, and the flanged haptics were tucked into the scleral tunnels to achieve stable intrascleral fixation.

Follow-up schedule

Postoperative examinations were performed at day 1, week 1, month 1, month 3, and month 6. At each visit, best-corrected visual acuity (BCVA), manifest refraction, intraocular pressure (IOP), slit-lamp biomicroscopy, and dilated fundus examination were performed. The 6-month visit served as the primary endpoint for quantitative image analysis. Adverse events were recorded at each visit and classified by severity, timing, and relationship to the fixation procedure (Table 4).

Table 4.

Adverse Events (N = 24 Eyes)

Event	n (%)	Severity	Timing	Management	Outcome
Uncomplicated	17 (71)	—	—	—	—
Any adverse event	7 (29)	—	—	—	—
Mild anterior uveitis	2 (8)	Mild	Day 3–5	Topical steroids	Resolved wk 2–3
Vitreous haemorrhage	2 (8)	Mild	Intraop	Observation	Cleared day 2–3
Microscopic hyphaema	1 (4)	Mild	Day 1	Observation	Resolved day 4
Transient corneal oedema	1 (4)	Mild	Day 1	Hypertonic saline	Resolved day 5
IOP spike (28 mmHg)	1 (4)	Moderate	Day 1	Timolol/dorzolamide	Normalised day 3
Cystoid macular oedema	4 (16)	Moderate	Month 3	Topical NSAIDs + steroids	Resolved month 5

No sight-threatening, IOL dislocation, or endophthalmitis events occurred. All adverse events resolved without sequelae. IOP = intraocular pressure; NSAIDs = non-steroidal anti-inflammatory drugs; wk = week(s)

Image acquisition and eligibility

Representative frames were extracted from intraoperative surgical microscope video recordings (Zeiss S88 operating microscope viewing system; Carl Zeiss Meditec AG, Jena, Germany; captured as PNG, 2048 × 2732 pixels, RGB) to serve as baseline images. Anterior segment photographs were obtained at 6 months using a slit-lamp–mounted camera (Topcon SL-D701 slit lamp with DC-4 digital imaging module; Topcon Corporation, Tokyo, Japan; images exported as JPEG, >=2048 × 1536 pixels, sRGB) under pharmacological dilation at equivalent zoom level. Biometric measurements (axial length, keratometry, anterior chamber depth, white-to-white) were obtained preoperatively using optical low-coherence reflectometry (Lenstar LS 900; Haag-Streit AG, Koeniz, Switzerland). Target IOL power was calculated using the Barrett Universal II formula with a slightly myopic refractive target of − 0.50 to − 1.00 D, selected to compensate for the more anterior effective lens position expected with intrascleral haptic fixation compared with in-the-bag implantation; the observed mean 6-month spherical equivalent of − 1.00 ± 0.75 D was consistent with this target.

Images were included in the geometric analysis only if they met all of the following predefined criteria: (i) full limbal circumference visible without specular obstruction; (ii) complete IOL optic margin visible; (iii) image in focus and free of motion artefact; (iv) sufficient contrast to trace both limbus and optic to their margins. Of 55 eyes operated during the study period, 24 (43.6%) met all eligibility criteria and were included in the analytic cohort. Reasons for exclusion of the remaining 31 eyes were: poor image quality precluding reliable landmark tracing (n = 22; itemised causes: significant specular reflection over the limbus or IOL margin [n = 9], motion blur or out-of-focus capture [n = 7], partial corneal oedema obscuring the limbal contour [n = 4], and inadequate pupillary dilation [n = 2]); and intraoperative-image flange visibility limitations precluding reliable F1/F2 outlining (n = 9; itemised causes: oblique microscope illumination [n = 5] and flange burial under conjunctival fold [n = 4]).

Geometric landmark extraction

Quantitative image analysis was performed using Fiji/ImageJ (version 2.14.0/1.54f; National Institutes of Health, Bethesda, MD, USA), an open-source distribution of ImageJ widely used for biomedical image analysis [13–17]. Each image was individually calibrated before measurement using the known 6.0 mm IOL optic diameter as reference (Analyze → Set Scale). The limbal boundary was manually traced using the Oval selection tool and Fit Ellipse function to define the limbus centre (L). The IOL optic margin was independently traced to define the optic centre (O). On intraoperative images, visible flange edges were outlined to define flange centres F1 and F2, with the midpoint M = (F1 + F2)/2. The line connecting F1 and F2 defined the fixation axis (Fig. 2). Landmark tracing was performed in a fixed order (Limbus → IOL → F1 → F2) for all images, with each region of interest named and verified in the ROI Manager before measurement. The Fit Ellipse function automatically computed the geometric centroid coordinates (X, Y) in calibrated millimetres for both the limbus and IOL optic ellipses; flange centres were defined as the centroid of the visible flange-head outline. All coordinates were exported to a structured CSV file immediately after each image, with image identifiers matching the eye number to prevent cross-image contamination. On low-contrast images, edge enhancement (Process → Find Edges) was applied prior to manual tracing to improve limbal boundary delineation; any image requiring edge enhancement was flagged in the dataset. All tracing was performed by a single experienced grader (H.A.) in a randomised image order, with the grader masked to clinical outcome; the corresponding implications for inter- and intra-observer reliability are addressed in the Limitations. A detailed step-by-step protocol for the Fiji/ImageJ geometric landmark extraction procedure, including calibration steps, tracing guidelines, and coordinate export, is provided in Supplementary Material 1.

Fig. 2.

Quantitative geometric landmark framework applied to an intraoperative anterior segment image. The limbus ellipse (green), IOL optic ellipse (blue dashed), fixation axis (yellow dashed, F1–F2), decentration vector (orange, L→O), and scaffold vector (purple, L→M) are overlaid. Key measurements: decentration d = 0.177 mm, angular deviation = 2.3° from 180°. Landmark coordinates are displayed in the inset data table. Abbreviations: L, limbus centre; O, IOL optic centre; F1/F2, flange positions; M, midpoint of fixation axis

Quality control measures

The following quality control rules were applied consistently across all 24 eyes: scale was verified on every image individually and never carried over from a different image; the major axis of the limbus selection was checked against expected white-to-white diameter (~ 11–12 mm) and the IOL selection against the known optic diameter (6.0 mm ± 0.2 mm); any discrepancy exceeding 0.5 mm from expected values triggered re-calibration; if specular reflection obscured more than 30° of any landmark boundary, the image was excluded; and all measurements were exported to CSV immediately after each image to prevent cross-image contamination.

Outcome measures

The primary outcome was longitudinal positional stability of the IOL, defined as the Euclidean displacement |ΔV| = √(ΔX² + ΔY²) of the optic centre (O) relative to the limbus centre (L) between the intraoperative baseline and the 6-month visit. Secondary outcomes included IOL decentration |L − O|, fixation decentration |L − M|, optic–haptic mismatch |O − M|, and F1–F2 span—all measured at the intraoperative timepoint only for flange-dependent metrics. Clinical outcomes included BCVA, spherical equivalent, and IOP at preoperative, 1-month, and 6-month visits (Table 2).

Table 2.

Visual Acuity and Refractive Outcomes (N = 24 Eyes)

Outcome	Pre-operative	1 Month	6 Months	p-value*
BCVA (logMAR)
Mean ± SD	1.28 ± 0.54	0.69 ± 0.50	0.52 ± 0.47	< 0.001
Median (IQR)	1.15 (0.88–1.73)	0.57 (0.19–1.13)	0.40 (0.09–0.86)
Range	0.50–2.50	0.10–1.88	0.00–1.82
BCVA (Snellen equivalent)
20/40 or better, n (%)	0 (0)	7 (29)	10 (42)
20/200 or better, n (%)	6 (25)	16 (67)	18 (75)
≥1 logMAR improved, n (%)	—	2 (8)	5 (21)
Spherical equivalent (D)
Mean ± SD	11.5 ± 1.50	-1.25 ± 0.75	−1.00 ± 0.75	< 0.001
IOP (mmHg)
Mean ± SD	14 ± 4	15 ± 3	15 ± 3	0.12
IOP > 21 mmHg, n (%)	0 (0)	1 (4)	0 (0)

*Paired t-test (pre-operative vs. 6 months). BCVA = best-corrected visual acuity; D = diopters; IOP = intraocular pressure; IQR = interquartile range; SD = standard deviation

Statistical analysis

Statistical analyses were performed using IBM SPSS Statistics for Windows (version 29.0; IBM Corp., Armonk, NY, USA). Continuous variables were expressed as mean ± SD. Normality of the paired-difference distributions for all primary and secondary geometric variables (IOL decentration, Euclidean displacement, fixation decentration, optic–haptic mismatch, F1–F2 span) and for BCVA, spherical equivalent, and IOP was tested using the Shapiro–Wilk test; in every case the test failed to reject the null hypothesis of normality (p > 0.05), supporting the use of parametric paired analyses despite the modest sample size. Paired comparisons were performed using the paired t-test. Effect size was estimated using paired Cohen’s d. Agreement between timepoints was assessed using the Bland–Altman method. Vector displacement analysis computed the magnitude and direction of positional change for each eye. A two-sided p value of less than 0.05 was considered statistically significant. A formal a priori sample size calculation was not performed because the study was designed as a retrospective audit of consecutive cases meeting predefined image-eligibility criteria. A sensitivity analysis using G*Power 3.1 (Heinrich-Heine-Universität, Düsseldorf, Germany) indicated that 24 paired observations were sufficient to detect a paired difference of approximately 0.06 mm at α = 0.05 (two-sided) with 80% power, given the observed within-subject SD of 0.018 mm; this analysis is presented to contextualise the detectable effect size rather than as a formal post-hoc justification of statistical power.

Results

Patient demographics and baseline characteristics

Twenty-four eyes of 24 patients were included in the analytic cohort. The mean age was 68.9 ± 7.1 years (range 55–80), with an equal distribution of male and female patients (12 each, 50%). Ten eyes (42%) were aphakic and 14 eyes (58%) had IOL subluxation or dislocation. Comorbidities included pseudoexfoliation syndrome (n = 6, 25%), diabetes mellitus (n = 6, 25%), hypertension (n = 5, 21%), and glaucoma (n = 4, 17%). Mean preoperative BCVA was 1.28 ± 0.54 logMAR. Baseline characteristics are summarised in Table 1.

Surgical data

All 24 eyes received a Sensar AR40e three-piece intraocular lens using 26-gauge needles with a bevel length of 2 mm.

Visual acuity and refractive outcomes

Mean BCVA improved from 1.28 ± 0.54 logMAR preoperatively to 0.69 ± 0.50 logMAR at 1 month and 0.52 ± 0.47 logMAR at 6 months (p < 0.001, paired t-test, preoperative vs. 6 months). The proportion of eyes achieving 20/40 or better increased from 0 (0%) preoperatively to 7 (29%) at 1 month and 10 (42%) at 6 months. The proportion achieving 20/200 or better increased from 6 (25%) preoperatively to 16 (67%) at 1 month and 18 (75%) at 6 months (Table 2).

Mean spherical equivalent decreased from 11.50 ± 1.50 D preoperatively to -1.25 ± 0.75 D at 1 month and − 1.00 ± 0.75 D at 6 months (p < 0.001). Mean intraocular pressure was 14 ± 4 mmHg preoperatively, 15 ± 3 mmHg at 1 month, and 15 ± 3 mmHg at 6 months (p = 0.12) (Table 2).

IOL centration and positional stability

Mean IOL decentration was 0.401 ± 0.240 mm (median 0.351 mm; range 0.056–1.045 mm) at the intraoperative baseline and 0.397 ± 0.234 mm (median 0.350 mm; range 0.049–1.028 mm) at 6 months. Detailed centration and stability metrics are summarised in Table 3. The mean change in IOL decentration was − 0.004 mm (95% CI − 0.031 to 0.023; p = 0.457; Cohen’s d = 0.02), indicating no significant positional change over the follow-up period. At baseline, 8 eyes (33%) had decentration exceeding 0.5 mm and 2 eyes (8%) exceeded 0.8 mm. At 6 months, 7 eyes (29%) exceeded 0.5 mm and 1 eye (4%) exceeded 0.8 mm (Table 3).

Table 3.

IOL Centration Stability: Baseline vs. 6-Month Outcomes (N = 24 Eyes)

Parameter	Baseline Mean ± SD	6 Months Mean ± SD	Mean Δ (95% CI)	p-value	Cohen’s d	Interpretation
IOL decentration (mm)	0.401 ± 0.240	0.397 ± 0.234	−0.004 (− 0.031, 0.023)	0.457	0.02	NS — Stable
Fixation decentration (mm)*	0.391 ± 0.212	—	—	—	—	—
Optic-haptic mismatch (mm)*	0.252 ± 0.138	—	—	—	—	—
F1–F2 span (mm)*	8.42 ± 0.65	—	—	—	—	—
Dec > 0.5 mm, n (%)	8 (33)	7 (29)	—	—	—	—
Dec > 0.8 mm, n (%)	2 (8)	1 (4)	—	—	—	—

NS = not significant; Dec = decentration; Δ = change (6 M – BL); CI = confidence interval. *Flange-dependent metrics (fixation decentration, optic-haptic mismatch, F1–F2 span) were measurable at the intraoperative baseline only; haptic flanges were not visible on 6-month slit-lamp photographs

Longitudinal trajectories of IOL position are shown in Fig. 3. The Bland–Altman plot (Fig. 4) demonstrated a mean difference of − 0.0045 mm with 95% limits of agreement of + 0.052 to − 0.061 mm. Vector displacement analysis revealed a mean Euclidean displacement of 0.029 ± 0.018 mm (range 0.004–0.058 mm); no eye demonstrated displacement exceeding 0.06 mm, as illustrated in Fig. 5A–C.

Fig. 3.

Individual IOL decentration trajectories from intraoperative baseline to 6-month follow-up (n = 24 eyes). Each grey line connects the baseline (blue circle) and 6-month (red diamond) decentration value for one eye. Lines highlighted in salmon indicate eyes with decentration exceeding 0.8 mm. The green dashed line marks the 0.5 mm clinically significant threshold. Box plots show median, interquartile range, and 95th percentile range at each timepoint. Baseline: 0.401 ± 0.240 mm; 6 months: 0.397 ± 0.234 mm; p = 0.457

Fig. 4.

Bland–Altman plot of IOL decentration stability between intraoperative baseline and 6-month follow-up (n = 24 eyes). The solid blue line represents the mean difference (−0.0045 mm). Red dashed lines indicate 95% limits of agreement (+0.052 mm to −0.061 mm). One outlier (red dot) falls at the lower limit of agreement

Fig. 5.

Vector displacement analysis of IOL positional change over 6 months (n = 24 eyes). (A) IOL vector displacement map. Blue arrows represent the positional change vector for each eye. The red arrow indicates the mean displacement vector. The dashed grey circle marks a 1 mm reference boundary. (B) Polar plot of IOL displacement direction and magnitude. Each point represents one eye; colour encodes displacement magnitude (mm) per the colour bar. Clustering in the 180–225° quadrant indicates a predominant inferonasal displacement pattern. (C) Distribution of IOL displacement magnitude across 24 eyes. Vertical dashed lines mark the mean and ± 1 SD boundaries. All eyes demonstrated displacement below 0.06 mm

Intraoperative fixation geometry

Mean fixation decentration was 0.391 ± 0.212 mm (range 0.105–0.881 mm). Mean optic–haptic mismatch was 0.252 ± 0.138 mm (range 0.047–0.596 mm). The mean F1–F2 span was 8.42 ± 0.65 mm.

Adverse events

Seven eyes (29%) experienced adverse events (Table 4). All resolved with conservative management. No severe or sight-threatening complications, IOL dislocations, or endophthalmitis occurred.

Discussion

The use of the innate 2 mm needle bevel as a visual endpoint during double-needle ISHF was associated with clinically negligible longitudinal IOL displacement over 6 months, supporting the hypothesis that a reproducible visual cue for tunnel formation yields consistent positional stability. Vector displacement analysis demonstrated that IOL optic centers remained tightly clustered around their intraoperative positions, with no preferential direction of migration—a finding that distinguishes random micro-movements from systematic mechanical drift. The visual and refractive rehabilitation observed was consistent with published outcomes for secondary IOL implantation [18, 19]. Consistent with this, Xu and colleagues [20] recently reported long-term outcomes of a modified Yamane technique in 75 eyes over a median follow-up of 3.3 years (up to 7 years), with no IOL dislocation and no progressive peri-haptic scleral thinning beyond 5 years. The same series, however, observed flange exposure in 14.0% of flanges, with localised conjunctival thinning at the exposed sites, identifying white-to-white distance as a protective factor and larger flange diameter as a risk factor. These findings support the durability of correctly constructed scleral tunnels and suggest that the 6-month stability observed here is likely to persist with longer follow-up, while underscoring the importance of complete flange burial and continued long-term conjunctival surveillance. Recent modifications to ISHF, including alternative IOL designs [21] and modified double-flanged approaches for managing dislocated posterior IOLs [22], further highlight the evolving surgical landscape and underscore the need for standardized tunnel construction. The relatively wide range of postoperative BCVA in this series (logMAR 0.00 to 1.82 at 6 months) deserves comment, as the optically excellent IOL centration outcomes contrasted with the modest proportion of eyes achieving 20/40 or better (42%). This dissociation reflects the heterogeneity of the underlying indications: 14 of 24 eyes had pre-existing IOL subluxation or dislocation (often after complicated cataract or vitreoretinal surgery), and several eyes carried concomitant comorbidities—diabetic maculopathy, glaucomatous optic neuropathy, longstanding amblyopia, and corneal endothelial compromise—that limit visual recovery independently of the fixation procedure. Postoperative cystoid macular oedema, observed in 4 of 24 eyes (16%) and resolving with topical anti-inflammatory treatment, also accounts for a portion of the moderate visual outcomes despite optically favourable IOL centration.

The rationale for standardizing intrascleral tunnel length to 2 mm is strongly supported by recent biomechanical and clinical evidence. Khan et al. [23] identified tunnel construction variability as the primary technical challenge of sutureless fixation in their comprehensive review but described no simple, cost-free standardisation method—a gap that the bevel-guided technique directly fills. Tokuhisa et al. [2] demonstrated that IOL tilt after intrascleral fixation contributes to refractive prediction error, and Jun et al. [3] established that absolute tunnel length is a critical prerequisite for baseline geometric stability; 1.0 mm tunnels produced significantly more tilt variation than 2.0 mm tunnels. While the bevel-guided approach cannot entirely dictate three-dimensional tunnel trajectory, it ensures the minimum 2 mm length required for adequate stromal engagement, circumventing one critical source of proprioceptive variability—namely inconsistent tunnel length. In this study, the mean Euclidean displacement was 0.029 mm and was substantially below the 0.5 mm clinically relevant threshold [1, 8], supporting adequate stromal engagement.

Several technology-dependent approaches have been proposed to address the same challenge. Sotani et al. [24] reported that intraoperative OCT-guided adjustment reduced IOL tilt to 2.10° ± 1.66° at 3 months, and Lo Giudice et al. [25] confirmed that iOCT can assess IOL centring across fixation techniques in real time. Other reproducibility strategies for scleral tunnel construction described in the literature include preoperative limbal-corneal marking systems, intraoperative angle-guided trocars or callipers, digital microscope overlays [5], and image-guided navigation platforms [6]. While each has demonstrated stepwise improvements in tunnel symmetry, they share the disadvantages of proprietary instrumentation, dedicated digital infrastructure, surgeon retraining, or additional intraoperative time. However, these methods require specialised hardware and capital investment that limits availability, particularly in resource-constrained settings [5, 6]. The bevel-guided technique utilises the intrinsic geometry of the surgical needle and requires no additional instrumentation or cost. In this series, longitudinal displacement of the IOL optic centre remained below 0.06 mm in all eyes during the 6-month follow-up. Comparisons with studies evaluating intraocular lens tilt using intraoperative OCT should be interpreted cautiously, as the present analysis assesses planar positional displacement rather than angular tilt. Previous studies have also demonstrated the value of quantitative image-based methods for evaluating IOL position, including digital measurement techniques for postoperative centration analysis described by Pardini et al. [26] and OCT-based vector analysis approaches reported by Kummelil et al. [27].

The choice of 26-gauge needles in the present series, in contrast to Yamane’s original 30-gauge description [1], deserves further comment. Beyond the practical advantage of accommodating PMMA haptics without trimming, the 26-gauge bevel is more readily identifiable under standard operating-microscope illumination than the finer 30-gauge bevel, enhancing the reliability of the bevel-disappearance endpoint. The principal trade-off is a wider scleral entry, which produces greater haptic–needle clearance during docking and may contribute to the observed baseline decentration (discussed further in the Limitations). The 26-gauge approach is widely used in routine clinical practice [23], which supports the external generalisability of the present findings.

The choice of IOL haptic material substantially influences flange behaviour and longitudinal stability. PMMA-based haptic IOLs, as used in this series (Sensar AR40e), provide a defined cauterisation profile and reproducible flange morphology, but their relatively rigid structure may be more sensitive to scleral tunnel asymmetry, and intraoperative haptic complications such as breakage or distortion during externalisation have been reported even in experienced hands. Icoz and Erkan [21] recently compared two three-piece PMMA-haptic IOLs (Alcon MA60AC and Sensar AR40e) under the modified Yamane technique and reported broadly comparable safety and visual outcomes, supporting the use of PMMA-haptic IOLs in this setting while noting that PVDF haptics are often preferred in clinical practice for their greater resistance to manipulation. Polyvinylidene fluoride (PVDF)-haptic IOLs (e.g., FIL-SSF Carlevale; Soleko, Rome, Italy) have been associated with different decentration profiles under intraoperative OCT evaluation [25], and modified double-flanged platforms have been introduced for managing dislocated posterior IOLs [22]. Comparative data on haptic material under standardised tunnel geometry remain scarce, and head-to-head studies pairing the bevel-guided tunnel endpoint with different haptic platforms would clarify whether haptic material modulates the stability benefit observed here.

Effective lens position (ELP) prediction in sutureless scleral fixation differs systematically from intracapsular IOL implantation because the fixated optic plane lies anterior to the expected capsular bag position; conventional IOL constants and biometric formulas may therefore yield a residual hyperopic or myopic offset [10, 28]. Reproducible scleral tunnel geometry, by stabilising haptic angulation and anteroposterior optic position, would be expected to reduce ELP variance and consequently improve refractive predictability; however, formal IOL constant optimisation for scleral-fixated lenses remains an emerging area, and was beyond the scope of the present geometric analysis. The mean spherical equivalent of − 1.00 ± 0.75 D at 6 months observed here is consistent with a slightly more anterior ELP than the intracapsular position, in agreement with previous reports [10]. Future studies coupling standardised tunnel geometry to refractive outcome modelling are needed to determine whether bevel-guided tunnel construction translates into clinically meaningful gains in refractive accuracy.

A critical insight from the geometric analysis was the dissociation between longitudinal positional stability and absolute initial centration. The mean change in IOL decentration over 6 months was clinically negligible (− 0.004 mm; p = 0.457), yet the mean absolute decentration at the intraoperative baseline was 0.401 mm—a magnitude at which Eppig et al. [8] reported measurable image-quality degradation in a model-eye analysis of aspheric IOLs. Initial centration is influenced by needle entry angle, chord length symmetry, and scleral elasticity—factors that the bevel cue does not address. Because longitudinal positional stability and initial centration are governed by different biomechanical determinants, standardising tunnel length resolves positional instability but cannot prevent planar misalignment. Supporting this dissociation, Truong Khah Van et al. [29] demonstrated that postoperative residual astigmatism correlates specifically with horizontal IOL decentration rather than tilt, underscoring the optical significance of entry angle variability independent of tunnel length. The geometric framework described here, by mapping limbus, optic, and flange centroids as independent vectors, provides an auditing methodology for identifying directional biases. Yin et al. [18] reported a pooled decentration rate of 7.3% for the flanged technique; the bevel-guided method addresses tunnel length variability, while optimising entry angle will require complementary strategies such as preoperative limbal marking.

The bevel-guided approach provides a simple, teachable endpoint requiring no additional equipment. The open-source Fiji/ImageJ geometric framework offers a quantitative alternative to subjective clinical grading [8], although integration into routine practice would require standardised imaging protocols and validation against anterior segment OCT.

This study has several limitations that should be considered when interpreting the results. First, as a single-centre, single-surgeon retrospective case series of 24 eyes, the generalisability of the bevel-guided technique across different levels of surgical experience, scleral biomechanics, and patient populations remains unevaluated. In particular, the inter-surgeon reproducibility of the bevel-disappearance endpoint, and its applicability to surgeons earlier in their learning curve, remain unproven; a multi-surgeon validation study stratified by experience level is therefore planned. The absence of a randomised control arm utilising standard proprioceptive tunnel estimation precludes direct comparison. Second, the geometric analysis relied exclusively on two-dimensional anterior segment imaging. While this approach was highly sensitive for measuring planar X–Y decentration and Euclidean displacement, it cannot quantify anteroposterior IOL tilt along the Z-axis, which requires three-dimensional evaluation via anterior segment optical coherence tomography [2, 24]. Importantly, the present analysis was deliberately designed to evaluate longitudinal positional stability arising from standardised tunnel length, and not to evaluate angular tilt; tilt assessment lies outside the scope of this geometric audit but remains an essential complementary analysis and is being incorporated into a prospective AS-OCT extension of this protocol. Similarly, lenticular (IOL-induced) astigmatism was not separately quantified. Although preoperative corneal astigmatism was available from biometric measurements, complete postoperative manifest refraction with separated cylinder magnitude and axis was not uniformly recorded in this retrospective dataset; without these postoperative components, lenticular astigmatism cannot be reliably partitioned from total refractive astigmatism by vector decomposition. This represents an additional limitation that prospective protocols should address by including standardised postoperative manifest refraction.

Third, the 6-month follow-up period, although sufficient to capture early positional stabilisation, may not detect late complications including haptic erosion, which is governed by tunnel depth rather than length and was not directly assessed in this study, and which Pakravan et al. [30] reported at a mean of 278 days after surgery. Twelve-month and longer safety data, including formal assessment of conjunctival integrity over the haptic tuck sites, are being collected in an ongoing prospective extension of this cohort and will be reported separately. Fourth, bevel length was verified by microscopic measurement against a metric ruler on the specific needle used (Fig. 1A); whether this measurement is consistent across needle batches and manufacturers has not been systematically evaluated. In addition, bevel visibility under the operating microscope may be reduced in eyes with subconjunctival haemorrhage, heavily pigmented sclera, or shallow microscope illumination angles; in these settings the bevel cue should be supplemented by tactile spatula-based feedback (discussed in the Background as a recommended adjunct, although not part of the surgical protocol applied in the present retrospective cohort) and by microscope-illumination adjustment. Fifth, flange-dependent geometric parameters were measurable only on intraoperative images where flanges were directly visible; these landmarks were not detectable on 6-month slit-lamp photographs, limiting longitudinal analysis of fixation geometry.

Sixth, the internal diameter of standard 26-gauge needles is substantially larger than the manufacturer-specified haptic cross-section of three-piece IOLs with modified-C PMMA haptics such as the Sensar AR40e (Johnson & Johnson Vision product information; commercial 26-gauge hypodermic needles conform to ISO 7864:2016 dimensional requirements [31]), producing a “loose fit” during haptic docking; although necessary to allow atraumatic haptic exchange, this clearance may contribute to the observed mean intraoperative decentration of 0.401 mm and therefore couples baseline centration to needle gauge selection. A direct comparison with tighter-lumen 27- or 30-gauge needles (which require haptic trimming or shrinkage but offer reduced clearance) would clarify this trade-off and is the subject of an ongoing protocol. Seventh, inter- and intra-observer reliability of the Fiji/ImageJ landmark extraction protocol was not formally assessed in this initial audit; duplicate independent grading with calculation of intraclass correlation coefficients should be incorporated into future iterations. Finally, it is acknowledged that a complete formal demographic and clinical comparison between the included (n = 24) and excluded (n = 31) eyes was not feasible in this retrospective dataset. Importantly, exclusions were driven by image-acquisition factors (specular reflection, motion blur, partial corneal oedema, inadequate pupillary dilation, oblique microscope illumination, and conjunctival fold burial) that are mechanistically independent of underlying patient or disease characteristics; systematic demographic bias from the exclusion process is therefore unlikely a priori. Nonetheless, because complete standardised preoperative variables (BCVA, full biometry profile, and detailed comorbidity coding) were not uniformly available for the excluded group, a formal Table 0 with confidence intervals and statistical comparison could not be reliably constructed. The relatively high image-quality exclusion rate (31 of 55 eyes) limits absolute generalisability of the centration values to the entire operated cohort, and a complete included-versus-excluded comparison is therefore deferred to the prospective extension protocol, which incorporates a standardised imaging workflow specifically designed to minimise image-quality exclusions and to enable a full demographic and clinical comparison across the entire treated cohort.

A multicentre randomised trial comparing bevel-guided standardisation with conventional proprioceptive estimation, stratified by surgeon experience, would establish whether this method reduces IOL decentration and tilt across heterogeneous populations. Incorporating anterior segment OCT alongside the 2D vector framework would enable three-dimensional positional modelling and validation against volumetric imaging. The quantitative auditing framework could be adapted into a structured postoperative review tool, enabling surgeons to track directional biases in their fixation technique across sequential cases and identify systematic patterns requiring correction. Finally, follow-up beyond 12 months is needed to determine whether positional stability persists and whether late haptic erosion rates are reduced compared with historical controls [18, 30].

Conclusions

The needle bevel provides a simple, reproducible, and cost-free visual cue for standardising intrascleral tunnel length during double-needle ISHF. In this cohort, excellent longitudinal positional stability was observed, with all eyes demonstrating displacement below 0.06 mm over 6 months. The quantitative geometric landmark framework described here offers an objective methodology for postoperative auditing of IOL centration. Initial centration, governed by entry angle and chord length symmetry rather than tunnel length, remains a separate biomechanical challenge that will require complementary standardisation strategies.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

BCVA

Best-corrected visual acuity

CI

Confidence interval

D

Diopters

iOCT

Intraoperative optical coherence tomography

IOL

Intraocular lens

IOP

Intraocular pressure

IQR

Interquartile range

ISHF

Intrascleral haptic fixation

NSAIDs

Non-steroidal anti-inflammatory drugs

OCT

Optical coherence tomography

PCR

Posterior capsule rupture

PXF

Pseudoexfoliation

SD

Standard deviation

Author contributions

HA: study conception, surgical technique development, intraoperative image acquisition, geometric image analysis, statistical analysis, and manuscript writing.EA: surgical assistance, critical revision of the manuscript for important intellectual content. Both authors read and approved the final manuscript.

Funding

This study received no external funding.

Data availability

The datasets generated and analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Institutional Review Board of Izmir Tepecik Training and Research Hospital (IRB no. 2025/08–017) and adhered to the tenets of the Declaration of Helsinki. Written informed consent was obtained from all participants.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.