Diffractive intraocular lenses reshape 577-nm retinal laser photocoagulation spots in a model eye

Abstract

To evaluate whether diffractive intraocular lenses (IOLs) reshape 577-nm retinal laser photocoagulation spots in a model eye. A standardized model eye received 577-nm continuous-wave laser photocoagulation through a retinal contact lens at fixed settings: 200-µm spot, 100 mW, and 0.2 s. One monofocal IOL (ZCB00) and two diffractive IOLs (ZMB00, ZXR00) were tested in 15 sessions; a no-IOL condition provided power normalization and quality control. Delivered power was measured calorimetrically and normalized to no-IOL controls. Thermal-paper responses were threshold-segmented to quantify core fraction and disk area. Session-blocked models compared pooled r = 3 mm values; C-to-r = 3 mm changes were used to assess field-dependent effects. At pooled r = 3 mm, core fraction was 0.406 (95% CI, 0.382–0.430) for ZCB00, 0.179 (0.155–0.203) for ZMB00, and 0.201 (0.177–0.225) for ZXR00. Mean differences versus ZCB00 were − 0.227 (95% CI, − 0.261 to − 0.193) for ZMB00 and − 0.205 (− 0.239 to − 0.171) for ZXR00 (both Holm-adjusted P < 0.001). Disk area ratios versus ZCB00 were 0.754 (95% CI, 0.704 to 0.808) for ZMB00 and 0.682 (0.637 to 0.731) for ZXR00 (both Holm-adjusted P < 0.001). Normalized delivered power remained close to unity (0.993–0.997). C-to-r = 3 mm comparisons showed larger morphology changes for diffractive IOLs, whereas normalized delivered power changed little with radius. Under the specified 577-nm model-eye conditions, the tested diffractive bifocal and EDOF IOLs altered threshold-defined photothermal spot morphology relative to the monofocal control, while normalized aperture-integrated power remained near unity, suggesting field-dependent redistribution of target-plane irradiance.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s10103-026-05042-9.

Introduction

Retinal photocoagulation remains an important medical laser treatment for several sight-threatening retinal disorders [1–3]. Its effect depends on absorption of laser light and conversion of optical energy into heat at the target tissue [4]. For fixed laser parameters, the target response may also depend on the optical pathway that delivers the focused beam to the target plane.

In pseudophakic eyes, retinal photocoagulation is delivered through an intraocular lens (IOL). Diffractive IOLs redistribute transmitted light across multiple foci or an extended focal range for presbyopia correction [5–9], thereby altering through-focus optical performance [10–12]. Their characteristic optical feature is an annular diffractive profile, typically arranged as concentric rings on the IOL optic. These structures may also affect the focused treatment laser beam as it passes through the IOL.

Previous optical assessments of diffractive IOLs have largely used idealized coaxial, collimated conditions and focused on image-forming performance [13, 14]. In contrast, retinal photocoagulation delivers a focused beam through a contact treatment lens and the pseudophakic optical pathway to a target plane. Whether diffractive IOLs alter target-plane spot morphology during focused photocoagulation delivery under treatment-like optical conditions, and whether any such field-dependent morphological changes are accompanied by corresponding changes in aperture-integrated delivered power, remain unclear. We therefore developed a standardized model-eye platform to compare monofocal and diffractive IOLs under fixed 577-nm laser settings, combining threshold-defined spot morphology with normalized delivered-power measurements.

Materials and methods

Study design and measurement strategy

This bench model-eye study evaluated target-plane spot morphology and delivered power during 577-nm continuous-wave photocoagulation through monofocal and diffractive IOLs under fixed optical geometry and laser settings. Two complementary readouts were used: aperture-integrated delivered power at the detection plane, measured calorimetrically and normalized to within-session no-IOL controls, and threshold-defined photothermal spot morphology recorded on thermal paper at the target plane, quantified by core fraction and disk area. The four r = 3 mm target-plane positions were averaged into one pooled off-axis value for each IOL and session, with session as the experimental unit. The central position, C (r = 0), served as the coaxial reference for drift assessment and C-to-r = 3 mm comparisons.

Model eye

A custom model eye was constructed based on the Gullstrand model eye and prior optical-bench literature on model-eye IOL evaluation [13, 15–18]. The platform was designed to provide a fixed pseudophakic optical pathway for comparative laser-delivery measurements rather than to reproduce biologic retinal tissue responses.

The model eye consisted of a sandblasted custom-machined aluminum-alloy anterior-segment mounting plate and posterior optical chamber (Online Resource 1), assembled with a model cornea, graphite nylon ring, silicone fixation ring, detection window, and, for IOL conditions, a test IOL (Fig. 1). Sandblasted surfaces reduced stray reflection and scatter. The cornea was fluorosilicone acrylate copolymer (refractive index, 1.415), and the detection window was high-transmittance optical glass. The internal model space from the cornea to the detection window was filled with 0.9% sodium chloride solution (refractive index, approximately 1.335).

Fig. 1.

Optical eye model used for bench photocoagulation experiments. (a) Exploded view of the model showing the model cornea, anterior segment mounting plate, graphite nylon ring, test IOL, silicone fixation ring, posterior optical chamber, and detection window. (b) Assembled model. (c) Simplified axial schematic showing the key optical elements and prespecified axial distances. The distance from the corneal vertex to the IOL plane was 5.0 mm, and the distance from the IOL plane to the detection plane was 18.0 mm

The model cornea had an outer diameter of 11.0 mm, a thickness of 0.50 mm, an anterior radius of 7.58 mm, and a posterior radius of 6.61 mm, corresponding to an equivalent refractive power of approximately + 43 D. The detection window measured 18.0 mm in diameter and 0.15 mm in thickness. The graphite nylon ring measured 12.0 mm in outer diameter, 6.0 mm in aperture, and 0.30 mm in thickness. The 6.0-mm aperture simulated a dilated treatment condition and matched the 6.0-mm optic diameter of all study IOLs. The silicone ring used to secure the IOL measured 12.0 mm in outer diameter with a 1.0-mm cross-sectional diameter. The distance from the corneal vertex to the IOL plane was 5.0 mm, and the distance from the IOL plane to the detection plane was 18.0 mm (Fig. 1).

.

IOL mounting and model assembly were performed by the same cataract surgeon under an ophthalmic surgical microscope (OPMI Lumera T; Carl Zeiss Meditec AG, Jena, Germany) for visual control of centration and tilt. After each IOL reload, assemblies showing decentration or tilt were reassembled before measurement.

IOL conditions

The study included one no-IOL reference condition and three TECNIS-platform IOLs (Johnson & Johnson Vision, Irvine, CA, USA): ZCB00, ZMB00, and ZXR00 (Table 1). All IOLs had a nominal base power of + 20.0 D. The no-IOL condition was used only for within-session power normalization and quality control and was not treated as a test IOL group.

Table 1.

Characteristics of the evaluated intraocular lenses (IOLs)

IOL	Design	Overall/optic diameter (mm)	Power, base/add (D)	Diffractive zone/central ring (mm)	Diffractive rings, n
ZCB00	Monofocal	13.0/6.0	+ 20.0/—	—	—
ZMB00	Diffractive bifocal	13.0/6.0	+ 20.0/+4.0	6.0/1.0	22
ZXR00	Diffractive EDOF	13.0/6.0	+ 20.0/—	6.0/1.60	9

IOL intraocular lens, EDOF extended depth of focus

Diffractive zone/central ring indicates diffractive zone diameter/central ring diameter

n indicates the number of diffractive rings

Dashes indicate not applicable

Laser system, treatment lens, and coupling

A 577-nm ophthalmic laser system (Iridex Corp., Mountain View, CA, USA) was used in continuous-wave mode. A BM 900 slit-lamp microscope (Haag-Streit AG, Köniz, Switzerland) was used with the magnification changer set to 1.6× for observation, focusing, and alignment. Laser settings were fixed at a nominal target-plane spot diameter of 200 μm, a power of 100 mW, and an exposure duration of 0.2 s. A TransEquator contact retinal laser lens (Volk Optical, Mentor, OH, USA) was used to deliver the treatment beam to the model-eye target plane. The setup was mounted on an optical breadboard to maintain alignment. Sodium hyaluronate ophthalmic gel (iviz; Bausch & Lomb Freda Pharmaceutical Co., Ltd., Jinan, China) served as the coupling medium between the treatment lens and model cornea. After each exposure, the detection window was inspected under the slit-lamp microscope for visible contamination and cleaned when necessary.

To support the geometric interpretation of upstream beam-footprint expansion, a separate simplified aiming-beam setup illustrated footprint enlargement at a plane 18 mm anterior to focus. Direct footprint measurement within the complete coupled treatment pathway was not feasible in the present setup. This supportive demonstration was not included in the inferential analysis (Online Resource 2).

Point definitions, target alignment, and primary analysis radius

A planar coordinate system was defined with the detection-window center as the origin. Five locations were prespecified: C (0, 0), XP (+ 3, 0), XM (− 3, 0), YP (0, + 3), and YM (0, − 3) (Fig. 2). Alignment used a printed target behind the detection window. At each location, the aiming beam was precisely focused on the target point (Fig. 2), and the slit-lamp position was locked. The target was then replaced with either the power-meter probe or thermal paper without changing the locked position (Online Resource 3). All procedures were performed by the same ophthalmologist experienced in retinal photocoagulation.

Fig. 2.

Printed alignment target, IOL mounting, and target-based aiming. (a) Coordinate schematic showing the detection window, 3-mm reference circle, and five predefined target locations: C (0,0), XP (+ 3,0), XM (− 3,0), YP (0, + 3), and YM (0, − 3). IOL mounting and corresponding target-based aiming views using the aiming beam are shown for ZCB00 (b, e), ZMB00 (c, f), and ZXR00 (d, g)

C served as the coordinate origin and central coaxial reference for drift assessment and radius-comparison analyses. It was not selected as the primary analysis location because clinically relevant photocoagulation is often extracentral. The r = 3 mm radius was prespecified as a representative off-axis location, enabling assessment of field-dependent effects while limiting vignetting, illumination nonuniformity, alignment error, and geometric mismatch between the planar detection surface and the curved retina at greater eccentricities. Four symmetric r = 3 mm points were averaged to reduce directional bias.

Power measurement and normalization

Delivered power at the detection plane was measured with a calorimetric laser power meter, VLP-2000-Z-200 mW (Beijing Ranbond Technology Co., Ltd., Beijing, China), with the probe coaxial behind the detection window. The meter had a 0–200 mW range, 22-mm detector aperture, 0.01-mW display resolution, and stated uncertainty < 3%. Because the detector aperture exceeded the 18-mm detection-window diameter, this measurement represented aperture-integrated delivered power rather than local irradiance. After the calorimetric reading stabilized, three readings were averaged at each point in each session. Normalized delivered power was defined as:

where the numerator and denominator were the corresponding IOL and no-IOL mean powers at the same point within the same session.

Sampling strategy and quality control

To minimize temporal drift and environmental variation, data acquisition was distributed across 15 independent sessions. At session start, the no-IOL reference was measured, and each test IOL was removed and remounted before its measurements. The standardized sequence was: initial no-IOL quality-control measurement at C; no-IOL power measurement at the five prespecified locations (C, XP, XM, YP, and YM); after each test IOL was mounted, triplicate power-meter readings and one thermal-paper exposure at each location; and final no-IOL quality-control measurement at C. Across sessions, IOL order was balanced so each IOL appeared equally often in each order position.

Thermal-paper image quality was controlled by excluding spots with visible contamination or scanning artifacts and reacquiring them within the same session. Power-measurement repeatability and system stability were assessed by the coefficient of variation (CV) of the three repeated power readings at each point and the C-point no-IOL drift ratio. The CV and drift ratio were calculated as:

where the numerator and denominator of the drift ratio were the mean no-IOL powers measured at C at session end and start, respectively. Sessions were accepted when CV < 5% at all points and the drift ratio was 0.95 to 1.05; otherwise, the session was repeated after troubleshooting.

Thermal-paper recording, scanning, and image quantification

Thermal paper (Yinghua; Shanghai Kaidun Office Equipment Co., Ltd., Shanghai, China) was placed behind the detection window to record spot morphology. One spot was recorded per sheet without a back heat sink. Within 10 ± 2 min after exposure, images were scanned at 4800 dpi with an EPSON Perfection V19 II (Seiko Epson Corporation, Suwa, Japan) flatbed scanner, with all automatic enhancement/correction disabled, and saved as TIFF files.

TIFF images were batch processed in ImageJ version 1.54p (National Institutes of Health, Bethesda, MD, USA) using a custom rule-based macro. After 8-bit grayscale conversion and inversion, background signal was estimated from corner regions, and two pilot-derived fixed relative thresholds were applied uniformly to all images. The low and high thresholds were defined as:

The low-threshold mask identified the total affected footprint and constrained segmentation. Within this footprint, the high-threshold mask identified the darker color-forming response. DISK was defined by hole filling of this high-threshold response region, and CORE was defined as DISK minus the high-threshold response region. Core fraction was calculated as:

Disk area was measured as the area of DISK. The same macro was applied to all images without group-specific adjustment. A representative spot-segmentation schematic is provided in the supplementary material (Online Resource 4).

Optical control experiments

Selective central-occlusion control

Selective central occlusion of ZXR00 assessed relative spatial sampling. Under microscopic guidance, a black, visually opaque, water-resistant coating covered the central 1.60-mm-diameter region within the innermost diffractive ring; surrounding annuli remained uncoated, and the 6.0-mm graphite aperture restricted transmission outside the optic. Six independent repetitions included model-eye reassembly. Focus was established at r = 3 mm, where the aiming spot remained visible, and held axially during lateral translation among C and four off-axis positions without refocusing. Residual 577-nm power was measured calorimetrically in triplicate per position; four off-axis means were pooled per repetition. Paired C and pooled r = 3 mm values were analyzed in milliwatts without no-IOL normalization using a two-sided exact Wilcoxon signed-rank test (Online Resource 5).

Axial-defocus control

An exploratory axial-defocus control used centered, fixed ZCB00. In three cycles, thermal-paper spots were recorded at C and the four r = 3 mm positions at best focus and after nominal 1.0-mm forward and backward axial displacement of the slit-lamp/laser-delivery base without refocusing (45 spots); other procedures matched the primary experiment. C was analyzed separately, with the four off-axis values pooled per cycle. With cycle as the unit, outcomes were summarized by mean ± SD, within-cycle percentage change, and direction consistency (Online Resource 6).

Statistical analysis

Session was treated as the experimental unit. For each session and IOL, XP, XM, YP, and YM were averaged to generate one pooled r = 3 mm value. Primary inference compared ZCB00, ZMB00, and ZXR00 at pooled r = 3 mm; the no-IOL condition was used only for within-session power normalization and quality control. Core fraction was the primary morphology endpoint, disk area was the gated key secondary endpoint, and normalized delivered power (η) was a supportive readout.

Between-IOL comparisons at pooled r = 3 mm were performed using session-blocked linear models, with IOL condition as the factor of interest and session as the blocking factor. Core fraction was analyzed on the original scale, whereas disk area and η were analyzed on the log scale and reported as back-transformed geometric means or ratios. Disk area was tested only if the primary endpoint was significant. Holm adjustment was applied to the two prespecified contrasts versus ZCB00 within each endpoint.

For the radius-comparison analysis, within-session changes from C to pooled r = 3 mm were calculated for each IOL. Core fraction was analyzed as an absolute change, expressed as pooled r = 3 mm minus C. Disk area and η were analyzed as relative changes on the log scale, expressed as within-session R3/C ratios, where R3 denotes pooled r = 3 mm. Within-IOL C-to-r = 3 mm comparisons were performed separately for each IOL using paired tests and were treated as descriptive. Between-IOL comparisons of these radius-dependent changes were performed using session-blocked linear models, with Holm adjustment applied to the two prespecified contrasts versus ZCB00.

Exact paired sign-flip tests based on session-level paired differences on the corresponding analysis scale were used as sensitivity analyses for the prespecified between-IOL contrasts. Model diagnostics were assessed by visual inspection of residual-versus-fitted and normal Q-Q plots. All tests were two-sided. Holm-adjusted P < 0.05 was considered statistically significant for prespecified between-IOL inferential contrasts. Unadjusted P values from within-IOL paired tests and exact sign-flip sensitivity analyses were reported descriptively and were not used for primary inference. All analyses were performed in R version 4.6.0 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Quality control and analysis set

All repeated power-meter readings met the CV criterion, with a maximum within-point CV of 2.94%. The C-point no-IOL drift ratio was stable across sessions (mean, 0.999; range, 0.987–1.009). Six thermal-paper measurements failed prespecified image-quality control and were repeated before final analysis. The final analysis included 225 thermal-paper observations (15 sessions × 3 IOLs × 5 locations). Primary inference used one pooled r = 3 mm session-level value per IOL. Representative raw scans and corresponding DISK and CORE masks are shown for center and r = 3 mm conditions (Fig. 3).

Fig. 3.

Representative raw spot images and segmentation masks. Representative thermal-paper spot images from the off-axis (r = 3 mm) and center (r = 0 mm) locations are shown for each IOL condition (ZCB00, ZMB00, and ZXR00). For each example, the raw scan is shown together with the corresponding disk mask and core mask. Raw images are presented without image enhancement. Mask panels are contrast-inverted for display, with white indicating segmented pixels. All panels are shown at the same magnification. Scale bar = 200 μm

Primary endpoint: core fraction

At pooled r = 3 mm, the session-blocked adjusted mean core fraction was 0.406 (95% CI, 0.382–0.430) for ZCB00, 0.179 (95% CI, 0.155–0.203) for ZMB00, and 0.201 (95% CI, 0.177–0.225) for ZXR00. Compared with ZCB00, core fraction was lower for both diffractive IOLs, with adjusted mean differences of − 0.227 (95% CI, − 0.261 to − 0.193; Holm-adjusted P < 0.001) for ZMB00 and − 0.205 (95% CI, − 0.239 to − 0.171; Holm-adjusted P < 0.001) for ZXR00 (Table 2; Fig. 4).

Table 2.

Prespecified primary analysis at pooled r = 3 mm

End point	Group or contrast	Estimate (95% CI)	P adj
Core fraction	ZCB00	0.406 (0.382 to 0.430)
	ZMB00	0.179 (0.155 to 0.203)
	ZXR00	0.201 (0.177 to 0.225)
	ZMB00 vs. ZCB00	-0.227 (-0.261 to -0.193)	< 0.001
	ZXR00 vs. ZCB00	-0.205 (-0.239 to -0.171)	< 0.001
Disk area	ZCB00	123,254 (117,373 to 129,430)
	ZMB00	92,921 (88,487 to 97,577)
	ZXR00	84,094 (80,081 to 88,308)
	ZMB00 vs. ZCB00	0.754 (0.704 to 0.808)	< 0.001
	ZXR00 vs. ZCB00	0.682 (0.637 to 0.731)	< 0.001
η	ZCB00	0.995 (0.991 to 0.998)
	ZMB00	0.997 (0.994 to 1.000)
	ZXR00	0.993 (0.990 to 0.996)
	ZMB00 vs. ZCB00	1.003 (0.998 to 1.007)	0.498
	ZXR00 vs. ZCB00	0.999 (0.994 to 1.003)	0.523

CI  confidence interval, IOL intraocular lens, η normalized delivered power, Padj  Holm-adjusted P value 

Group estimates for disk area are geometric means in µm²

For disk area and η, between-IOL contrasts are expressed as geometric mean ratios

Fig. 4.

Primary between-IOL comparison at pooled r = 3 mm. (a) Core fraction. (b) Disk area. (c) Normalized delivered power (η). Dots indicate session-level values. Boxes show the median and interquartile range, and whiskers indicate the 5th and 95th percentiles. Adjusted P values are Holm-adjusted for the two prespecified contrasts versus ZCB00

Key secondary endpoint: disk area

Because the primary endpoint was significant, analysis proceeded to the gated key secondary endpoint, disk area. At pooled r = 3 mm, session-blocked geometric mean disk area was 123,254 μm² (95% CI, 117,373–129,430) for ZCB00, 92,921 μm² (95% CI, 88,487–97,577) for ZMB00, and 84,094 μm² (95% CI, 80,081–88,308) for ZXR00. Relative to ZCB00, the geometric mean ratios were 0.754 (95% CI, 0.704–0.808; Holm-adjusted P < 0.001) for ZMB00 and 0.682 (95% CI, 0.637–0.731; Holm-adjusted P < 0.001) for ZXR00 (Table 2; Fig. 4).

Supportive endpoint: normalized delivered power

At pooled r = 3 mm, the session-blocked geometric mean η was 0.995 (95% CI, 0.991–0.998) for ZCB00, 0.997 (95% CI, 0.994–1.000) for ZMB00, and 0.993 (95% CI, 0.990–0.996) for ZXR00. Relative to ZCB00, the geometric mean ratios were 1.003 (95% CI, 0.998–1.007; Holm-adjusted P = 0.498) for ZMB00 and 0.999 (95% CI, 0.994–1.003; Holm-adjusted P = 0.523) for ZXR00 (Table 2; Fig. 4). Descriptive absolute pooled r = 3 mm power was similar across IOL conditions, ranging from 88.3 to 88.7 mW, compared with 88.9 mW in the no-IOL reference condition.

Radius comparison

In the prespecified radius-comparison analysis, core fraction and disk area showed larger C-to-r = 3 mm changes in the diffractive IOLs than in ZCB00. For core fraction, the mean change was + 0.006 in ZCB00, compared with − 0.196 in ZMB00 and − 0.173 in ZXR00. Relative to ZCB00, the additional off-axis reductions were − 0.202 (95% CI, − 0.247 to − 0.156; Holm-adjusted P < 0.001) for ZMB00 and − 0.179 (95% CI, − 0.224 to − 0.133; Holm-adjusted P < 0.001) for ZXR00.

Disk area showed a similar pattern. The R3/C ratios were 0.969 for ZCB00, 0.756 for ZMB00, and 0.678 for ZXR00. Relative to ZCB00, the between-IOL ratios of change were 0.781 (95% CI, 0.695–0.877; Holm-adjusted P < 0.001) for ZMB00 and 0.700 (95% CI, 0.623–0.787; Holm-adjusted P < 0.001) for ZXR00.

By contrast, η changed little with radius. The R3/C ratios for η were 1.005 for ZCB00, 1.006 for ZMB00, and 1.004 for ZXR00. Relative to ZCB00, the between-IOL ratios of change were 1.001 (95% CI, 0.995–1.006; Holm-adjusted P = 1.000) for ZMB00 and 0.999 (95% CI, 0.994–1.005; Holm-adjusted P = 1.000) for ZXR00 (Table 3; Fig. 5).

Table 3.

Prespecified within-session change from C to pooled r = 3 mm

End point	Group or contrast	Estimate (95% CI)	P / Padj
Core fraction	ZCB00	Δ(R3 − C) 0.006 (-0.002 to 0.014)	0.122
	ZMB00	Δ(R3 − C) -0.196 (-0.228 to -0.163)	< 0.001
	ZXR00	Δ(R3 − C) -0.173 (-0.213 to -0.133)	< 0.001
	ZMB00 vs. ZCB00	-0.202 (-0.247 to -0.156)	< 0.001
	ZXR00 vs. ZCB00	-0.179 (-0.224 to -0.133)	< 0.001
Disk area	ZCB00	R3/C 0.969 (0.915 to 1.025)	0.250
	ZMB00	R3/C 0.756 (0.679 to 0.843)	< 0.001
	ZXR00	R3/C 0.678 (0.605 to 0.760)	< 0.001
	ZMB00 vs. ZCB00	0.781 (0.695 to 0.877)	< 0.001
	ZXR00 vs. ZCB00	0.700 (0.623 to 0.787)	< 0.001
η	ZCB00	R3/C 1.005 (0.997 to 1.014)	0.217
	ZMB00	R3/C 1.006 (0.999 to 1.013)	0.101
	ZXR00	R3/C 1.004 (1.000 to 1.009)	0.072
	ZMB00 vs. ZCB00	1.001 (0.995 to 1.006)	1.000
	ZXR00 vs. ZCB00	0.999 (0.994 to 1.005)	1.000

CI confidence interval, IOL intraocular lens, η normalized delivered power, Padj  Holm-adjusted P value

Δ(R3 − C) indicates the within-session change from C to pooled r = 3 mm

R3/C indicates the within-session ratio of pooled r = 3 mm to C

For disk area and η, between-IOL contrasts are expressed as ratios of change

Fig. 5.

Prespecified radius-comparison analysis shown as change-measure plots. (a) Change in core fraction from C to pooled r = 3 mm, expressed as Δ(R3 − C). (b) Disk area ratio, expressed as R3/C. (c) Normalized delivered power ratio (η), expressed as R3/C. R3 denotes pooled r = 3 mm. Dots indicate session-level values. Boxes indicate the median and interquartile range; whiskers indicate the 5th and 95th percentiles. The dashed horizontal reference line indicates 0 in panel a and 1 in panels b and c. Adjusted P values are Holm-adjusted for the two prespecified between-IOL contrasts versus ZCB00

Exact sign-flip sensitivity analyses were concordant with the session-blocked models for all prespecified between-IOL contrasts (Online Resource 7). Model diagnostic plots showed no major departures from model assumptions for the primary pooled r = 3 mm models or the radius-comparison models (Online Resources 8 and 9).

Selective central-occlusion control

With central occlusion, the 635-nm aiming spot was barely discernible at C but visible at r = 3 mm (Online Resource 5). Across all six repetitions, residual 577-nm delivered power was higher at pooled r = 3 mm than at C (32.91 ± 0.26 vs. 5.90 ± 0.09 mW; mean paired difference, 27.01 ± 0.27 mW; 95% CI, 26.72–27.30 mW; two-sided exact Wilcoxon signed-rank test, P = 0.031).

Axial-defocus control

Both displacement directions visibly blurred the aiming spot (Online Resource 6). At pooled r = 3 mm, forward and backward displacement increased core fraction by 7.58% and 11.56%, respectively, and decreased disk area by 13.57% and 12.37%. Across six direction-by-cycle comparisons, core fraction always increased and disk area decreased in five; none reproduced reductions in both endpoints. At C, core fraction increased and disk area decreased in all six comparisons.

Discussion

In this model-eye study, at pooled r = 3 mm, the tested diffractive IOLs produced lower core fractions and smaller disk areas than the monofocal control, while normalized aperture-integrated delivered power remained near unity. These findings suggest field-dependent redistribution of target-plane irradiance without reduced normalized aperture-integrated power.

It is important to recognize that retinal photocoagulation uses a treatment beam focused at the target plane rather than a collimated beam. Accordingly, in the present model, the beam footprint at upstream locations, including the IOL plane, was expected to be larger than the nominal target-plane spot. The simplified aiming-beam observation supported this geometric expectation. This geometry concentrates energy at the focus while keeping local irradiance lower along the preceding optical path, which may help limit unintended heating away from the treatment site. At the same time, the broader upstream footprint can illuminate a larger area of the IOL optic. In diffractive IOLs, illumination of the phase-relief structure can divide incident light among diffraction orders and reshape the point-spread function [19–21], providing a plausible optical basis for the altered target-plane spot morphology observed in the diffractive groups.

Thermal paper provided a practical, reproducible morphological readout that, like clinical photocoagulation, depended on an observable spot response. However, this readout represents a material-specific, threshold-dependent photothermal response rather than a direct irradiance map or biologic retinal response [22]. Spots showed a relatively pale center, a surrounding darker region, and a faint outer gradient. The pale center may reflect a distinct recording-layer response regime associated with greater local exposure, whereas the darker region may be influenced by local heat deposition and short-range lateral heat spread [23]. Because the tonal response was uncalibrated and potentially nonlinear, these regions were interpreted morphologically. The faint gradient had an unstable boundary, so inference was restricted to the more robustly segmentable DISK and CORE regions. Core fraction was prespecified as the primary endpoint because it captures internal morphology while being less affected by overall spot-size variation than absolute CORE area; disk area represented the threshold-defined main response region.

Interestingly, the thermal-paper images in this experiment did not show a regular, symmetric halo-like pattern often described with diffractive IOLs [24–28]. This observation is consistent with the present optical conditions. Prior optical-bench studies generally used coaxial, collimated, relatively large-aperture beams centered on the IOL optic, which favor symmetric illumination of the diffractive structure. In contrast, the present experiment used a focused projection pathway and a primary off-axis analysis radius of r = 3 mm. Under these conditions, incomplete and asymmetric sampling of the diffractive microstructure would be expected to produce local spot reshaping rather than a symmetric halo, especially in a thresholded thermal-paper readout.

Location-dependent morphology was assessed by comparing C with four pooled r = 3 mm positions under the nominal 200-µm target-plane spot condition. Both diffractive IOLs showed larger radial changes in core fraction and disk area than ZCB00, whereas normalized aperture-integrated delivered power changed little, consistent with a field-dependent effect. Selective central occlusion provided complementary evidence: after occluding the central 1.60-mm-diameter region of ZXR00, residual 577-nm delivered power was substantially greater at pooled r = 3 mm than at C in all six repetitions. Under this geometry, the off-axis path therefore appeared to extend farther beyond the occlusion and sample more surrounding annular diffractive structure. Quantitative 577-nm footprint measurements at the IOL plane or validated system-specific optical modelling are needed to confirm this interpretation. Axial defocus generally decreased disk area but increased core fraction, unlike the simultaneous reductions produced by diffractive IOLs relative to ZCB00. This opposite core-fraction response makes unrecognized axial defocus alone less likely to explain the findings.

To our knowledge, direct experimental evidence on diffractive IOLs and retinal laser delivery remains limited. Huang and Caspar (Invest Ophthalmol Vis Sci 2007;48 E-Abstract 3118) reported altered spot formation through an apodized diffractive multifocal IOL in a simplified 635-nm optical-bench setting, consistent with our threshold-defined morphological findings. The present study extends that observation to a standardized model-eye pathway incorporating a retinal contact lens and a clinical 577-nm continuous-wave photocoagulator, with quantitative comparison of photothermal morphology and aperture-integrated power at central and off-axis locations. Wakabayashi et al. [29] noted potential challenges during peripheral laser retinopexy through multifocal IOLs but found no apparent compromise in outcomes versus monofocal IOLs. Although addressing different evidence levels, their clinical findings and our bench data support attention to possible field-dependent effects.

Clinically, these findings support location-specific, response-guided photocoagulation through the tested diffractive IOLs. Because morphological changes were field dependent, settings titrated at one retinal location should not be assumed equivalent elsewhere; settings increased for an adequate peripheral response should be retitrated as treatment moves toward more central retinal locations. This is particularly relevant to macular micropulse treatment protocols in which continuous-wave extramacular titration is used to determine the macular treatment power; under the tested diffractive-IOL conditions, a weaker extramacular titration response could lead to an inappropriately high macular power setting. Normalized aperture-integrated delivered power remained near unity despite altered morphology, so a weaker local response does not necessarily indicate a comparable loss of total delivered power. Increasing power raises both total energy and any spatially redistributed component. Because the distribution and biological effects of this component were not measured, effects on retinal or other intraocular tissues remain uncertain. Adjustments should therefore be conservative and incremental, with attention to unintended tissue effects.

Several limitations should be considered. Target-plane 577-nm irradiance was not mapped: thermal paper provided an uncalibrated, potentially nonlinear, threshold-dependent photothermal readout interpreted morphologically, whereas calorimetry measured aperture-integrated power. Spatial redistribution remains an interpretation requiring calibrated beam profiling. The model eye did not reproduce retinal or choroidal responses, and IOL centration and tilt were not quantified. Generalizability is limited to this geometry, one 577-nm continuous-wave parameter set (200 μm, 100 mW, 0.2 s), and the tested + 20 D TECNIS-platform designs; other laser systems, treatment parameters, IOLs, and in vivo conditions require validation.

Conclusion

Under the tested model-eye configuration and 577-nm continuous-wave settings (200-µm nominal spot, 100 mW, 0.2 s), the + 20 D diffractive IOLs ZMB00 and ZXR00 produced lower core fractions and smaller disk areas than the monofocal ZCB00 control at pooled r = 3 mm, while normalized aperture-integrated delivered power remained near unity. Larger off-axis morphology changes with little radius-dependent change in η suggest field-dependent redistribution of target-plane irradiance without reduced aperture-integrated delivered power.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

K.W. designed the study, performed the experiments, collected and analyzed the data, prepared the figures and tables, and wrote the original draft. F.J. contributed to study design, experimental implementation, data interpretation, and critical revision of the manuscript. Y.L. and W.D. contributed to study design, experimental implementation, and resources. R.L. supervised the study, provided resources, contributed to data interpretation, and critically revised the manuscript. All authors reviewed and approved the final manuscript.

Data availability

The datasets generated and analyzed during the current study are not publicly available because they are stored as local experimental records, but are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.