Optimization of biometry for best refractive outcome in cataract surgery

Abstract

High-precision biometry and accurate intraocular lens (IOL) power calculation have become essential components of cataract surgery. In clinical practice, IOL power calculation involves measuring parameters such as corneal power and axial length and then applying a power calculation formula. The importance of posterior corneal curvature in determining the true power of the cornea is increasingly being recognized, and newer investigative modalities that can estimate both the anterior and posterior corneal power are becoming the standard of care. Optical biometry, especially using swept-source biometers, with an accuracy of 0.01–0.02 mm, has become the state-of-the-art method in biometry. With the evolution of IOL formulas, the ultimate goal of achieving a given target refraction has also moved closer to accuracy. However, despite these technological efforts to standardize and calibrate methods of IOL power calculation, achieving a mean absolute error of zero for every patient undergoing cataract surgery may not be possible. This is due to inherent consistent bias and systematic errors in the measurement devices, IOL formulas, and the individual bias of the surgeon. Optimization and personalization of lens constants allow for the incorporation of these systematic errors as well as individual bias, thereby further improving IOL power prediction accuracy. Our review provides a comprehensive overview of parameters for accurate biometry, along with considerations to enhance IOL power prediction accuracy through optimization and personalization. We conducted a detailed search in PubMed and Google Scholar by using a combination of MeSH terms and specific keywords such as “ocular biometry,” “IOL power calculations,” “prediction accuracy of refractive outcome in cataract surgery,” “effective lens position,” “intraocular lens calculation formulas,” and “optimization of A-constants” to find relevant literature. We identified and analyzed 121 relevant articles, and their findings were included.

Introduction

Patients’ expectations regarding refractive outcomes after cataract surgery have continuously increased over the past few years. With the improvement of surgical techniques and equipment, cataract surgery is no longer a simple cataract extraction but has gained popularity as a refractive procedure aimed at achieving optimal unaided vision. Hence, high-precision biometry and accurate intraocular lens (IOL) power calculation have become essential components of cataract surgery.

In clinical practice, power calculation is accomplished by obtaining measurements such as corneal power and axial length (AL) by using ultrasound amplitude-scan or optical biometry and then applying a power calculation formula.[1] In addition to AL and corneal power, numerous other parameters such as anterior chamber depth (ACD), lens thickness (LT), corneal white-to-white (WTW), effective lens position (ELP), and even the age of the patient influence the refractive outcome.[1] If biometry and IOL power prediction are perfect, then the prediction error (PE) for every eye would be zero, and the mean error (ME) and the mean absolute error (MAE) for all patients having cataract surgery would also be zero.[2] However, despite efforts in physics and technology to standardize and calibrate methods of IOL power calculation, the concern that remains is: What is the best approach for determining the optimal IOL power for refractive cataract surgery with the minimum possible PE?

This review provides a comprehensive overview of parameters for accurate biometry along with consideration to enhance IOL power prediction accuracy by optimization and personalization. It also highlights the limitations of the present-day IOL formulas and provides optimization methods to achieve precise visual acuity for patients undergoing cataract surgery. To gather pertinent literature, an extensive literature search and desk review was conducted in databases such as PubMed, and Google Scholar by using specific keywords such as “ocular biometry,” “IOL power calculations,” “prediction accuracy of refractive outcome in cataract surgery,” “effective lens position,” “intraocular lens calculation formulas,” and “optimization of A-constants.” This comprehensive search yielded 121 articles that were considered relevant to the topic and were included in our analysis of findings. Furthermore, we expanded our research by examining the references cited within these articles. For frequently cited references, the earliest ones were chosen for citation. Only studies with full-text availability in English from the years 1967–2023 were included in our review.

Parameters affecting the prediction accuracy

Precision of clinical measurements and accuracy of IOL calculations are significant factors for attaining the targeted postoperative refractive outcome and minimizing the PE.[2] Norrby[1] evaluated the effect of the variability of the input parameters on refractive outcomes and estimated their relative contribution to the final PE [Table 1]. The biggest source of error is the inaccuracy in the IOL formulas’ predictions of the postoperative IOL position (35.5%). Inaccuracies in AL and keratometry measurements, whether arising from the measurements themselves or associated underlying assumptions, account for 17% and 10.1% of the error, respectively. It might seem surprising that postoperative refraction determination is a significant contributor to the PE (27%); however, the standard deviation for subjective refraction is 0.39 diopter (D), indicating wide variability in the measurement of refractive outcomes.[1]

Table 1.

Relative contribution of input parameters to prediction error[1]

Input parameter	Percentage responsible for inaccuracy
Prediction of the effective lens position	35.5%
Axial length measurement	17%
Keratometry measurement	10.1%
Postoperative refraction	27%

Although the goal of achieving zero postoperative refractive PE for all eyes remains distant, it is a realistic objective to improve outcomes by reducing the spread of PEs.[2] The evolution of IOL formulas has moved closer to achieving accuracy, with the ultimate goal of successfully achieving a given target refraction in 100% of cases. However, achieving this outcome is challenging due to the limitations of the IOL formulas in accommodating the variety of eyes that exist. Tables 2a and b highlight these limitations. Ideal case selection is a prerequisite before surgery for cataract to ensure patient satisfaction as well as optimal outcomes. On average, only 73.4% of eyes are considered within the range of normal based on anterior segment size and AL.[3,4] Eyes deviating from normal geometry may not only yield imprecise measurements but also require individualized adjustments that go beyond the range of validity provided by the standard IOL formulas. One such adjustment is the Wang–Koch adjustment to finetune IOL power calculations in eyes with AL greater than 25.0 mm.[5,6,7] It can be considered as an adjunct for some third- and fourth-generation IOL formulas such as Holladay 1, Hoffer Q, SRK/T, Haigis, and Holladay 2 [Table 2a]. Here, the measured AL is adjusted to correct for systemic inaccuracies in these long eyes. Overall, the effect of the Wang–Koch adjustment is to shift refractive outcomes in long eyes from hyperopic to myopic.

Table 2a.

Limitations of existing IOL power formulas

IOL Formula	Generation	Principle	Variables used	Limitations
SRK/T	3rd Generation	Vergence	AL, K, A-constant, and postoperative refractive target	Less accurate in long eyes* than modern vergence-based formulas [i.e., Barrett Universal-II, Emmetropia Verifying Optical Formula (EVO), Hill-RBF, and Kan] Assumes normal ACD
Holladay 1	3rd Generation	Vergence	AL, K, SF, and postoperative refractive target	Less accurate in long eyes*, resulting in hyperopic outcomes
Hoffer Q	3rd Generation	Vergence	AL, K, pACD, and postoperative refractive target	No use of anatomic ACD; so, theoretically less reliable in anatomically abnormal anterior segments* Recommended to be replaced by Hoffer QST
Holladay 2	4th Generation	Vergence	AL, K, ACD, LT, WTW, CCT, and age	Less accurate in long eyes* Requires IOL consultant software
Haigis	4th Generation	Vergence	AL, K, ACD, 3 constants: a0, a1, a2,	Less accurate in long eyes* Less accurate in eyes with extreme LT values
OKULIX	4th Generation	Ray Tracing	AL, K, ACD, CCT, IOL design parameters, and postoperative refractive target	Calculating sum-of-segments AL is difficult in most clinical settings because the vitreous thickness is not available to calculate in many optical biometers Demands that the precise optical profile of the IOL be known, which is proprietary and may not be shared by the IOL manufacturer, thus available for use for a limited number of IOLs
Barrett Universal-I & II	5th Generation	I-Vergence II-Ray tracing	AL, K, ACD, LT (optional), WTW (optional), LF/DF or A-constant, and postoperative refractive target	Less accurate in short eyes (i.e., AL ≤22.0 mm] or requiring implantation of a lens power ≥30 D Barrett Universal - II may work for axial lengths between 20.8 and 22.0 mm
Olsen- C	5th Generation	Ray tracing	ACD, LT, and postoperative refractive target	Requires purchase of PhacoOptics® – IOL Power Calculation Software

*Wang/Koch adjustment formula:

1. For optimizing AL to reduce hyperopic outcomes in patients with long eyes[5]

Holladay1, 2-center optimized AL=0.8814 x IOLMaster AL +2.8701

Haigis, 2-center optimized AL=0.9621 x IOLMaster AL +0.6763

SRK/T, 2-center optimized AL=0.8981 x IOLMaster AL +2.5637

Hoffer Q, 2-center optimized AL=0.8776 x IOLMaster AL +2.9269

2. Modified equations for optimizing the AL: Using User Group for Laser Interference Biometry lens constants and manifest refraction converted to 6 m.[6]

Modified Holladay 1 optimized AL=0.817 x (measured AL) + 4.7013

Modified SRK/T optimized AL=0.8453 x (measured AL) + 4.0773

3. Modified equations for optimizing the AL: For the Holladay 2 formula, adding AL adjustment can improve the accuracy of the lens calculation.[7]

Linear Holladay 1 optimized-AL=0.8048 x (OLCR-AL) + 4.9195

Linear Holladay 2 optimized-AL=0.8332 x (OLCR-AL) + 4.2134

Table 2b.

Limitations of existing IOL power formulas

IOL Formula	Generation	Principle	Variables used	Limitations
Kane	New Generation	Blended (Vergence, Regression and Artificial Intelligence-based)	AL, K, ACD, LT (optional), CCT (optional), Gender, A-constant (developed to be similar to SRK/T A-constant), and postoperative refractive target	Less accurate than Barrett Universal-II
EVO	New Generation	Vergence	AL, K, ACD, LT (optional), CCT (optional), Corneal refractive LVC status, and postoperative refractive target	Solely intended for use with a pseudophakic IOL and not for use with a refractive phakic IOL Does not come pre-installed on any optical biometer Paucity of published evidence currently available
Ladas-Super Formula	New Generation	Artificial Intelligence	Depending upon variables, constructs a super surface using Hoffer Q, Holladay 1 with Wang-Koch axial length adjustment* and Haigis with deep learning, and postoperative refractive target	Overall, inferior to the Barrett Universal II, especially for short eyes
Hill- RBF	New Generation	Artificial Intelligence	AL, K, ACD, LT (optional), WTW (optional), CCT (optional), A-constant, and postoperative refractive target	Significantly influenced by LT, independently of the ACD Myopic shift with thin lenses and a hyperopic shift with thick lenses
Hoffer QST	New Generation	Artificial Intelligence	AL, K, pACD, and postoperative refractive target	Paucity of published evidence currently available
Intraoperative Aberrometry	New Generation	Vergence	Postoperative refractive target, K, corneal diameter, and AL	Not significantly different from the best preoperative biometry-based methods available for IOL power selection in short eyes No clear consensus regarding utility in routine cataract surgery (i.e. outside of post-LVC and astigmatic eyes) Requires additional equipment: ORA (Alcon)
Pearl-DGS	New Generation	Artificial Intelligence	AL, K, ACD, LT, WTW, CCT, A-constant, Machine learning models to predict posterior corneal radius and theoretical Internal lens position, and postoperative refractive target	Paucity of published evidence currently available

AL - Axial Length, K - Keratometry, SF - Surgeon Factor, pACD - Personalized Anterior Chamber Depth, LT - Lens Thickness, WTW - White to White, CCT - Central Corneal Thickness, IOL - Intraocular Lens, LF - Lens Factor, DF -Design Factor

Keratometry

The determination of corneal power is of prime importance in IOL calculations. However, there is no instrument available that can directly measure the true corneal power. Instead, it is assessed by measuring the anterior corneal curvature (K-reading) using a keratometer. Nevertheless, the K-reading alone does not accurately represent the true power of the cornea as it does not account for the curvature of the posterior surface. Until recently, it was commonly assumed that the radius of the posterior surface was 82% of the front surface.[4] However, with advancements in technology, we can now determine the true power, and all IOL formulas are designed to convert the K-reading into a net power.[8]

Measurement of anterior corneal curvature is probably as precise as it can be using keratometers. A standard deviation of 0.02 mm, corresponding to a refractive error of 0.11 D, can be achieved under ideal conditions in normal eyes during keratometric measurements.[1] In eyes that deviate from the normal, such as highly oblate aspheric post-LASIK eyes, keratometry can result in a hyperopic outcome of 3.00 D or higher.[9] Topography and adequate consideration of asphericity are needed in such situations to reduce errors.

K-readings obtained using the standard manual keratometer are measurements between two points 3.2 mm apart on a 44-D cornea using four light spots.[8,10,11,12,13,14,15] The IOLMaster automated keratometer (IOLMaster 500) employs telecentric keratometry using six light spots within a 2.4-mm diameter around the corneal center, resulting in a measurement of approximately 0.37 D higher for the same eye. The Haag–Streit Lenstar utilizes a width of 1.8 and 2.2 mm, measuring even closer to the center of the cornea; therefore, its measurement of the same cornea results in an even steeper measurement.[10,12] IOLMaster 700 uses three zone telecentric keratometry at 1.5, 2.5, and 3.2 mm diameters with 18 spots, yielding the highest mean keratometry values.[16] Pentacam uses a rotating Scheimpflug camera to measure corneal radii of curvature by using a central 3-mm zone and yields the highest values of mean posterior keratometry; however, the mean total keratometry (TK) is lower than that obtained using IOLMaster 700.[8,11,12,16,17,18] Galilei G4 has a dual Scheimpflug analyzer and a rotating Scheimpflug system in addition to the Placido disk to measure both anterior and posterior corneal surfaces and yields flatter true K values as compared to IOLMaster 700.[15,17,19] Table 3 illustrates the various modalities available for preoperative assessment of corneal power.

Table 3.

Investigative modalities to assess preoperative keratometry

Instrument	Technology	Principle	Comparative Outcomes	Posterior Corneal Curvature Assessment
Manual Keratometry	Variable object size or variable image size using biprisms	Doubling Principle	Comparable clinical outcomes with manual keratometry and Lenstar LS 900	No
Automated Keratometry	First Purkinje image used to calculate corneal curvature	Reflectometry	More repeatable readings than manual keratometers	No
Lenstar LS 900	Pro version contains “T-cone,” double-ring Placido disk topographer, integrated EyeSuite toric software with Hill-RBF, Barrett, and Olsen formula	Optical lowcoherence reflectometry	Lenstar LS 900 and ORA suggest comparable visual outcomes in toric IOLs	No
IOLMaster 700	Integrated swept-source technology with retinal OCT scan	Partial coherence interferometry	Good agreement between Pentacam and IOLMaster for corneal power calculation but less so for cylinder and axis of astigmatism	Yes
Orbscan IIz	Digitally recreates posterior corneal curvature using triangulation of previously generated elevation and anterior topography	Placido disc and slit scanning topography	Measures higher keratometry value than iTrace	Yes
Pentacam	Compensates for ocular movements	Scheimpflug imaging	Less repeatability than Orbscan	Yes
Galilei	Uses revolving dual channel Scheimpflug camera with Placido disc technology	Ray tracing technology	Comparable readings with Galilei and Pentacam	Yes
iTrace	Zernike polynomial mapping usingPlacido corneal topography	Ray tracing aberrometry	Comparable with Pentacam for astigmatism >2 D	Yes
ORA	Measures distortion of fringe pattern to calculate refractive values	Wavefront interferometry	Lesser mean postoperative residual astigmatism than Lenstar	Yes
Cassini	Uses color diode lights	LED ray tracing technology	Measured keratometry value higher than Pentacam	Yes

Although the average differences in keratometry measurements between the devices have been compared, it is important to consider the range of variation to gauge the interchangeability of each device.[20,21,22] The difference in mean keratometry between the IOLMaster and an automated refractor at the 95% limit-of-agreement (LoA) range was reported to be 1.20 D by Elbaz et al.[23] and 1.83 D by Whang et al.[12] When compared with manual keratometry at the 95% LoA range, the difference was reported to be 0.45 D by Shirayama et al.[15] and 1.80 D by Whang et al.[12] When compared with Galilei analyzer, the difference was reported to be 0.27 D.[15] Asena et al.[24] found this 95% LoA range between 0.80 D and 1.12 D for the difference between IOLMaster 700 and IOLMaster 500 and between 1.37 D and 1.96 D for the difference between IOLMaster 700 and Pentacam HR. The K-readings are hence very repeatable when only one instrument is used; however, they are not consistent between instruments.[4] Therefore, it is important that if one is using the IOLMaster’s biometer for AL but a manual keratometer for K-readings, then do not switch to using the IOLMaster manufacturer’s A-constant. If you want to switch, then use K-readings of the IOLMaster’s keratometer.

With new lenses such as toric, multifocal, accommodative, and aspheric IOLs on board and the increasing demand for precision, addressing posterior corneal astigmatism has become an important consideration in cataract surgery. In addition, corneal refractive surgeries change the ratio of the anterior-to-posterior corneal radius, thereby further necessitating the measurement of posterior corneal curvature in all eyes post refractive surgery for IOL power calculations.[1] Traditionally, posterior corneal radii are not directly measured but instead assumed by simply combining the anterior and posterior surfaces into one by using an assumed overall corneal refractive index. Different IOL calculation formulas incorporate different values for these parameters, and any resulting systematic differences are compensated for by adjusting various parameters or formula constants.[25]

Modern optical methods such as dynamic Scheimpflug and slit scanning techniques, optical coherence tomography (OCT), the iTrace system, and other imaging devices for the anterior segment [Table 3] enable reproducible measurement of the posterior corneal radius. Incorporating these data into suitable formulas by using Gaussian optics allows for the calculation of total corneal power without resorting to keratometric indices. The Pentacam AXL, a widely used device for keratometry assessment in clinical practice, uses a 360° rotating Scheimpflug camera to measure both the anterior and posterior corneal surfaces, and then the total corneal refractive power is calculated using ray tracing and Snell’s law of refraction. The IOLMaster 700, one of the most widely used optical biometers for IOL power calculation, measures the anterior corneal radius by using telecentric keratometry and simultaneously measures the posterior corneal radius and central corneal thickness (CCT) by using swept-source optical coherence tomography (SS-OCT). Its telecentric keratometry technology, a distance-independent method, measures the anterior surface at 1.5, 2.5, and 3.2 mm diameters by using 18 projected reference points (equivalent to nine meridians). The posterior surface is mapped using six meridional SS-OCT scans.[16] SS-OCT scans also provide central topography, which provides a decisive advantage to aid in IOL selection for premium IOLs such as a toric, multifocal, or extended depth of focus. In contrast, the CASIA2, as a newly introduced three-dimensional anterior segment OCT (AS-OCT), measures keratometry with a faster scanning speed (50,000 axial scans/s), wider scanning range (16 mm × 16 mm × 13 mm), and higher imaging resolution. i-Trace is a ray tracing aberrometer that employs both wavefront aberrometry and Placido corneal topography. It can calculate corneal, lens, and total aberration separately along with topographic screening.[26] The Cassini diagnostic device utilizes specular reflection of 679 colored light-emitting diodes (LEDs) to measure anterior corneal curvature and evaluates the posterior corneal surface by using seven additional infrared LEDs based on the second Purkinje.[27] Although there have been numerous studies comparing different devices, the definitive superiority of a single method has not been established.[11,12,15,16,17,18,19,20,21,22,23,24,26]

TK represents a new measure of total corneal power. TK is designed to be compatible with standard keratometry for normal eyes and can be used in classic IOL power calculation formulas.[28] The use of either posterior corneal curvature (PK) or TK may represent an additional benefit in post-keratorefractive surgery cases by replacing generalized extrapolations with precise measurements.[29] Recently, Wang et al.[30] found that the standard Haigis IOL formula in conjunction with TK was comparable to alternate post-refractive IOL formulas, including the Barrett True-K formula. Alsetri et al.,[31] in a retrospective comparative case series of 31 post-myopic LASIK patients, found that the use of the IOLMaster 700 TK and calculated posterior corneal curvature (PK1 and PK2) measurements resulted in improved accuracy of IOL calculations. Lawless et al.[32] found similar improvements in the prediction accuracy of IOL calculation methods by the addition of posterior corneal values and TK as measured by the IOLMaster 700 in their 50 consecutive patients (72 eyes) with prior laser refractive procedures.

Axial length

Accurate and reproducible AL measurement is the stepping stone for IOL power calculation. According to Olsen T,[33] AL plays a key role in determining postoperative refraction and is responsible for 54% of the actual refractive error. An AL measurement error of 100 µm translates to a postoperative refractive error of 0.28 D.[33]

The AL of the eye can be measured using ultrasound (either contact or immersion) or optical means (IOLMaster, Lenstar, and others). A-scan ultrasonography is routinely used in ophthalmic biometry, enabling a typical longitudinal resolution of 150–200 µm and clinical accuracy of AL measurement of approximately 100–120 µm.[33,34] Depending on the measured intraocular distance, precision values of 0.3–10.0 µm have been reported.[35]

Because of its high precision, better refractive outcome, ease and rapidity of use, and reduced risk of infection, optical biometry [partial coherence interferometry (PCI), optical low coherence reflectometry (OLCR), optical low coherence interferometry (OLCI), and SS-OCT] has essentially replaced ultrasonic biometry. With an accuracy of 0.01–0.02 mm, which is almost five times that of ultrasonic (US) biometry, it has become a state-of-the-art method in biometry.[36,37]

Partial coherence interferometry measures axial eye length parallel to the vision axis because the patient fixates the measurement beam or a coaxial fixation beam in ultrasound biometry, which can result in deviations from the visual axis during measurements. In addition, eye lengths determined by partial coherence interferometry are approximately 460µ-µm longer than those measured using ultrasound biometry.[38] One possible reason for the larger partial coherence interferometry values is that ultrasound is reflected mainly from the internal limiting membrane, whereas light is reflected from the retinal pigment epithelium, resulting in a difference corresponding to the retinal thickness in the fovea, which is approximately 130 µm.[39,40]

Findle demonstrated that PCI enables better prediction of the refractive outcome than US biometry for all tested IOL power formulas.[41] Olsen described a mean PE of 0.65 D by using ultrasonic measurement and 0.43 D by using the optical method.[42] However, Montes-Mico et al.[43] reported that measurements between IOLMaster, Lenstar, and immersion biometry were highly correlated for AL (R = 0.99) in patients with cataract.

Measurements using an optical biometer are not greatly affected by the extremes of AL, making it more useful for measuring eyes that are too long or too short. In addition, the technique of optical biometry is not greatly affected by the state of accommodation and pupil size. Packer reported it as user-friendly and less dependent on technician expertise than the ultrasound method.[44] It is also superior in eyes with posterior staphyloma[44,45] because of the more precise localization of the fovea. Unlike ultrasound, the optical biometer has settings to measure AL in silicon oil-filled eyes.[46,47] It also makes adjustments for AL measurements in aphakic and pseudophakic eyes. The precision and resolution of PCI in pseudophakic eyes are found to be more than 20 times better than those with conventional ultrasound. In addition, biometry using the scanning version of PCI can be performed in pseudophakic eyes, allowing for more precise determination of IOL-dependent A-constants required for most IOL power calculation formulas.[40]

Despite the improvement in their software, PCI optical biometers fail to achieve measurements in approximately 6%–13%[47,48,49,50] of eyes because of dense cataract,[48,49] posterior subcapsular cataract,[48,49] difficult fixation,[49] macular pathology,[49] or other less common reason, where ultrasonic measurement with immersion technique is required. In these cases, PCI cannot measure AL with an accuracy greater than 10%–20%.[50,51]

A second limitation of PCI optical biometers is the lack of a LT measurement, which is a required variable in the Holladay II IOL power calculation software. This is in contrast to immersion ultrasound, which is applicable to all types of cataracts and can generate phakic LT, especially for surgeons using the Holladay II formula.[52] The IOLMaster 500, a partial coherence interferometry-based biometer, uses PCI to measure only AL. Corneal curvature, horizontal iris width (WTW), and ACD are assessed using imaging techniques, and there is no assessment of corneal, crystalline lens, or retinal thickness. In contrast, the Lenstar LS900 uses optical low-coherence reflectometry to measure corneal thickness, ACD, crystalline or IOL thickness, and AL.[52,53,54] The Lenstar also assesses the central corneal curvature, horizontal iris width (WTW), pupil size, and decentration of the pupil and visual axis by image analysis, without the need for realignment.

SS-OCT-based biometers provide clear advantages over other technologies used in ocular biometry due to their ability to measure AL along six different axes, offering better penetration ability and extremely rapid data acquisition.[55,56,57] The deeper light penetration or long-range OCT imaging of posterior segment structures not only results in a high success rate of AL measurement but also reduces refractive PEs, making it useful for accurate refractive correction. Furthermore, the high speed of SS-OCT allows for the collection of two- or three-dimensional data in hundredths of milliseconds with high lateral and axial resolution (5 µm) while minimizing the unfavorable influence of patient eye movements on scan quality. In addition, SS-OCT utilizes an invisible light source, which is less disturbing for patients and provides an improved signal-to-noise ratio.

The IOLMaster 700 SS-OCT biometer, which is the first SS-OCT-based biometric device, obtains multiple measurements in a single capturing process. It presents measurements of AL, ACD, LT, and CCT from a single OCT image aligned with the eye’s visual axis. Moreover, it enables the identification of the foveal configuration (foveal pit), ensuring correct fixation during the measurements.[58]

The addition of foveal SS-OCT scans to routine fundus biomicroscopy examinations greatly improves the preoperative diagnostic ability to identify macular abnormalities. However, this combination is inferior to an assessment of the macula by macular spectral-domain OCT. Because macular alterations can affect the measurement of AL and the final refractive result, scanning the macula by using SS-OCT biometers during biometry is a useful screening process for detecting macular pathology in settings where retinal OCT is not routinely performed.[58,59,60]

In a study by Montés-Micó R. comparing six optical biometers, statistically significant differences were found when measuring most ocular parameters. SS-OCT-based biometers demonstrated the best agreement for measuring axial distances, including ACD, LT, and AL.[61] However, the author noted that depending on the specific parameter and its use, these biometers may still be interchangeable because the clinical effect of the differences may be negligible. Oh et al.[62] compared three SS-OCT devices (ANTERION, CASIA2, and IOLMaster 700) and observed good agreement in parameters such as anterior corneal curvature, CCT, ACD, and LT. The AL measurements of ANTERION and IOLMaster 700 showed excellent agreement and can be considered interchangeable. However, it should be noted that the TK value of each device was different and should not be used interchangeably.

Non-contact optical biometry devices can also reliably measure AL, CCT, and corneal radii in children aged 4 years or more.[63,64] Huang et al,[65] in their assessment of the reliability and reproducibility of SS-OCT biometry in healthy children aged 6–12 years, found that the reproducibility of averaged measurements from three consecutive readings was higher than that of a single measurement. Tomita et al.[66] evaluated the success rate of biometry variables essential for myopia studies and calculations of lens power in elementary school students by using an OA-2000 SS-OCT-based biometer. They found that AL, CCT, and corneal curvature can be accurately measured in all cases. However, ACD and LT were not accurately measured in some cases.

Leighton et al.[67] demonstrated that the IOLMaster 700, under cycloplegia, is highly repeatable and reproducible for measuring ocular biometry, including LT and ACD, in children. Their study results showed that the SS-OCT IOLMaster 700 has superior repeatability for AL compared to the PCI IOLMaster and A-Scan ultrasound. Higashiyama et al.[68] also confirmed that the Argos SS-OCT combination is useful for accurately detecting changes in the anterior segment of the eye after cycloplegia in pediatric patients aged 4–14 years. The biometer showed an increased ACD and a decreased LT after cycloplegia. Hussaindeen et al.,[69] in their study comparing the AL measurements using ARGOS-SS-OCT and IOLMaster-PCI-based biometers in the pediatric population, found that both biometers provided measurements that were within clinically acceptable limits. However, they concluded that the ARGOS-SS-OCT-based biometer can be recommended for use in the pediatric age group due to its faster acquisition speed and improved resolution rates.

Table 4 provides a list of biometers and their features along with some insights into their differences. In our experience, SS-OCT-based optical biometers are the gold standard for ocular biometry. The properties of the device in terms of repeatability and reproducibility, ease of use, and integrated diagnostic tools must be contemplated when choosing among the various devices based on this technology. In cases where the clinic lacks corneal tomography, an SS-OCT biometer such as ANTERION can be a great option, providing high-quality corneal maps and anterior segment metrics.

Table 4.

List of optical biometers and their features

Biometer	Principle	Technology
AL-Scan	PCI	Uses an 830-nm superluminescent diode to measure the AL Uses a 970-nm LED for assessing K and pupil size (PS), and a 525-nm LED for determining WTW Scheimpflug principle is used for CCT and ACD measured with an 470-nm LED WTW and PS are ascertained by fitting the best circle with the lowest error square to the detected edge K1 and K2 are obtained from an analysis of the images of two mires of spots over two areas (2.4 and 3.3 mm in diameter) No LT values are obtained with this biometer Offers an optional built-in ultrasound biometer to avoid moving patients to another machine if an A-scan is needed
Aladdin	OLCI	Uses 820-mm superluminescent diode, for measuring AL, ACD, CCT, and LT K readings are calculated using 24-ring Placido disk-ring reflection at 1024 reference points in approximately 3.0-mm optical zones (3, 5, and 7-mm zones) WTW distance is calculated from the corneal topography PS is determined via infrared and a white LED Dynamic pupillometry allows users to see lens centration and the contraction and dilatation of the pupil under photopic and mesopic conditions to assist in premium IOL selection Comes with a corneal topographer and wavefront analyzer, which helps to decide which patients would be suitable for premium IOL implantation
Lenstar LS 900	OLCR	Uses an 820-nm superluminescent diode to measure the CCT, ACD, LT, and AL K readings are calculated by analyzing the anterior corneal curvature at 32 reference points oriented as T-Cone, a double-ring (1.65- and 2.3-mm diameter) Placido disk topographer that measures a 6-mm optical zone WTW and PS measurement is based on high-resolution color photography of the eye Comes with the option of the EyeSuite IOL toric planner software
Argos	SS-OCT	Provides a 2-dimensional OCT using 3000 A-scans/s and a wavelength of 1060 nm (20-nm bandwidth) K is measured from the OCT image in conjunction with a 2.2 mm diameter ring made up of 16 LEDs Corneal diameter (CD) is measured using the OCT image and Alcon image guidance uses this image to compute WTW CCT, ACD, LT, and AL are measured using the OCT, considering different refractive indices (cornea: 1.376, aqueous and vitreous humors: 1.336, lens: 1.410)
IOLMASTER700	SS-OCT	Uses a two-dimensional OCT((44 mm scan depth and 22 mm resolution in tissue) with a scanning rate of 2000 scans/s (a wavelength of 1055 nm, from 1035 to 1095 nm) OCT scan informs about the CCT, ACD, LT, and AL values A 950 nm light source is used to obtain telecentric K in 3 zones (1.5, 2.5, and 3.2 mm on an average cornea of 7.9-mm radius, using a refractive index of 1.3375) Measures PS and the horizontal WTW distance with an LED source of 800 nm May help pick up macular pathologies, particularly intraretinal fluid and macular holes, in patients seen for pre-cataract surgery biometry For toric lens implantation, allows the surgeon to compare a reference image of the eye that uses blood vessels as landmarks to delineate the axis of the intraoperative eye
OA-2000	SS-OCT	Uses OCT at a scanning rate of 1250 scans/s and a wavelength of 1060 nm to measure AL, ACD, CCT, and LT To obtain K values at different diameters (2.0, 2.5, and 3 mm, using a refractive index of 1.3375), projects a Placido disc, considering 9 rings (each having 256 points) onto the cornea (5.5 mm) Pupil diameter and WTW are obtained from video analysis
Anterion	SS-OCT	Uses 1300-nm wavelength of light with a scanning rate of 50,000 scans per second Uses OCT-based structural images to generate all ocular biometric measurements including anterior segment Acts as an anterior segment tomographer, providing topography of the corneal front and back surface at thousands of locations Both anterior and posterior corneal curvature is measured using 65 B-scans on 8 mm (using a refractive index of 1.3375 for the anterior cornea and 1.376 for the posterior cornea) Captures a wider scan depth (14.5 mm) and scan width (16.5 mm), measuring the axial length in the range of 14–32 mm. with an in-tissue axial resolution of <10 μm

Nevertheless, a problem that exists with both ultrasonic and optical methods is the lack of a gold standard to verify the absolute values of AL. Because no international organization for standardization has been defined for AL measuring instruments, different devices are calibrated relative to each other based on clinical data. Different instruments may produce different results, thereby requiring a final adjustment to clinical results. Ideally, the same eye should be measured using these different instruments before the instrument can be used for reliable IOL power calculations.

Anterior chamber depth

Studies based on preoperative and postoperative ultrasound biometry have demonstrated that 38% of the errors in predicted refraction after the implantation of an IOL can be attributed to errors in estimating the postoperative ACD.[33]

Formulas developed by Fyodorov et al.[70,71] Colenbrander,[72] Hoffer,[73] and Binkhorst[74] use a constant value for the postoperative ACD. This constant indicates the IOL position based on the style of the IOL and its placement in the eye. Modern IOL power formulas have incorporated ACD measurements to improve the accuracy of the IOL power prediction curve. Third-generation formulas adjust the postoperative ACD based on AL and corneal curvature or apply principles of Gaussian optics.[75,76,77,78,79,80] In fourth-generation formulas, ACD is crucial for predicting the ELP. In the Haigis formula,[81] it is critically important because this formula uses only AL and ACD to predict the ELP. ACD is also essential for the Olsen,[82] Holladay II (unpublished), Hoffer H,[83] and Hoffer H-5 formulas[84] as well as for ray-tracing software.

Hence, accurate measurements of ACD are crucial to minimize the potential for undesirable refractive outcomes. ACD can be measured clinically using various methods, including optical pachymetry, ultrasonic pachymetry, optical low coherence reflectometry biometry, PCI biometry, scanning slit video keratography (Orbscan), slit-scanning Scheimpflug imaging (Pentacam), dual Scheimpflug imaging, anterior segment optical coherence tomography (AS-OCT), and magnetic resonance imaging. Both central ACD (between the anterior surface of the lens and the anterior surface of the cornea) and anatomical ACD (between the anterior surface of the lens and the posterior surface of the cornea) can be evaluated using anterior segment images.[85] Several studies have compared ACD measurements obtained using different devices.[86,87,88,89,90]

IOLMaster provides slightly smaller ACD measurements compared to other devices, whereas Lenstar LS900 provides slightly higher ACD readings.[45] This difference arises because Lenstar measures ACD from the corneal endothelium to the anterior lens surface, whereas IOLMaster measures ACD from the corneal epithelium to the anterior lens surface. It is important to note that the ACD measurement obtained using IOLMaster is not applicable for determining ACD in pseudophakic eyes as the evaluation software is specifically designed for phakic eyes. Orbscan consistently yields higher measurements for ACD compared to Galilei and Pentacam. For instance, Salouti et al.[91] found mean differences of 0.32 and 0.30 mm between Orbscan–Galilei (Dual Scheimpflug system) and Orbscan–Pentacam, respectively, suggesting that these modules are not interchangeable in every clinical situation.

Hoffer’s initial study on ACD,[92] published in 1980, analyzed 6950 phakic cataractous eyes by using immersion A-scan US and reported an average ACD of 3.24 mm ± 0.44 mm (SD). A more recent study by Hoffer,[87] using two optical laser biometers on 100 phakic eyes, showed an average ACD of 2.98 mm ± 0.49 mm with the optical method of PCI and 3.11 mm ± 0.47 mm with the laser method of the Lenstar. In general, studies[50,53,93,94,95,96] comparing ultrasonic ACD with PCI measurements demonstrated a deeper ACD with PCI; however, two studies[43,45] showed shallower readings, and two others[23,97] showed no difference. Overall, PCI measurements tend to indicate an ACD that is, on average, 0.1 mm deeper than ultrasound measurements.[85,98]

Even when dedicated AS-OCT, such as Visante-OCT, which achieves an axial resolution of approximately 18 µm and provides highly detailed, in-depth images of the anterior chamber, is used, differences exist. When comparing ACD measurements, Visante-OCT yields the highest values, followed by Lenstar LS900 and IOLMaster.[85] Discrepancies also exist when comparing ACD values obtained using rotating Scheimpflug cameras and PCI devices, with some authors finding no difference[99] while others reporting a statistical difference.[23,100]

However, measurement of ACD by using different SS-OCT devices, namely IOLMaster 700, CASIA2, and ANTERION, showed excellent agreement, with the 95% LoAs being narrow and clinically insignificant.[62] Comparison of ACD measurements between SS-OCT devices (IOLMaster 700) and dual Scheimpflug imaging system (Galilei G4) also revealed mean ACD measurements to be closely correlated, with an interclass correlation of 0.965.[18]

A literature review by Domínguez-Vicent et al.[101] assessing device interchangeability on ACD revealed that the following device pairs can be used interchangeably to calculate IOL power: These are the names of device pairs which can be used interchangeably for ACD assessment for IOL power calculation. This has been mentioned in this review article by Dominguez- Vicent et al. which was carried out to clarify which clinical devices can or cannot be used interchangeably in clinical practice to measure ACD and /or WTW distance.

White-to-white measurement

WTW corneal diameter is one of the ocular biometric components that has had applications in selecting anterior chamber lenses for years. It also plays a role in calculating the power of the IOL with the Holladay 2 formula in cataract surgery. WTW corneal diameter is integral when implanting sulcus-fixated IOLs, including implantable collamer lenses. As the sulcus diameter (SD) is not directly accessible (except through the use of ultrasound microscopy), the formula SD = WTW + 1 mm is used to find the suitable outer diameter of the IOL.

WTW measurements can be obtained through simple manual measurement with a ruler or by using more advanced imaging devices. Optical biometry allows for automatic measurement of the horizontal iris width, which corresponds to WTW. Other devices capable of measuring WTW include ultrasound A-scan, ultrasound biomicroscope, Orbscan, Pentacam, and Galilie systems. Various studies comparing WTW measurements obtained using different devices have reported statistically significant differences.[101,102,103]

From a clinical point of view, we believe that Anterion and IOLMaster 700 can be considered interchangeable, and so can Anterion and Pentacam HR.

Effective lens position

ELP is the term used to denote the position of the IOL in the eye. This position does not coincide with the image principle position of the IOL itself. Although it is related to the true IOL position in the eye, it is neither that nor the position of the principal plane; instead, it is the fictitious position that gives the desired result.

When evaluating the parameters contributing to the PE in IOL power calculations, Norrby stated that the biggest source of error is the inaccuracy in predicting the postoperative IOL position (35%).[1] A standard deviation of 0.31 mm for the postoperative IOL position corresponding to a 0.45-D error in a mean-sized eye has been reported. This effect of IOL position errors on refraction increases significantly toward shorter ALs and decreases for longer eyes. It becomes zero for 0.00-D IOL power and changes signs in even larger eyes.[9] Therefore, the basic task when estimating the IOL power for achieving the desired postoperative refraction is to predict the postoperative position of the IOL based on preoperative data. If the IOL position is not accurately predicted, even the use of the most sophisticated optical calculations will not result in a good prediction of postoperative refraction. Conversely, if the position is well predicted, an exact optical treatment will result in a better prediction of postoperative refraction.

According to Haigis, it is not possible to characterize the ELP with a single number.[104] Prior to the 1980s, ELP was a constant value of 4 mm for every patient and every lens (mainly for anterior chamber IOLs). In the early 1980s, Binkhorst introduced the concept of using AL as a scaling factor for ELP, improving its accuracy as a single variable predictor.[105] In the late 1980s, the use of two variable predictors, namely K (corneal curvature) and AL, further enhanced the scaling accuracy of ELP.[106] The position and size of the capsular bag should determine the position of the IOL, which is located in the capsular bag. Therefore, the position and thickness of the crystalline lens should also be considered in the IOL position prediction algorithm to improve accuracy.[9] In the mid-1990s, Olsen et al. improved ELP accuracy by incorporating two additional variables: preoperative ACD and LT.[107] In 1997, Norrby described another approach known as the lens haptic plane (LHP) approach to determine the IOL position for normal looped lenses. The LHP approach makes a clear separation between the prediction of postoperative IOL position and the optical power calculation. The position of the LHP is related to the anatomy of the eye and can be visualized as the equator of the capsule. It is independent of the specific IOL model, making the preoperative determination of LHP a universal task separate from any formula constants associated with IOL models.[108]

The third-generation formulas, namely Holladay,[75] SRK/T,[76] and Hoffer Q,[80] use thin lens theory for predicting postoperative ELP, employing different prediction algorithms and adjustment factors for short and long eyes. Although these formulas work well for IOL power calculations, they do not directly provide information about the position of the actual thick IOL within the eye.[1]

The fourth-generation formulas, including Holladay II[109] and Haigis,[104] as well as the fifth-generation formulas, such as Barrett Universal II[110] and Olsen,[111] incorporate additional parameters in the calculation of the ELP. These additional parameters differ depending on the specific formula used. The Haigis formula considers the preoperative ACD as one of the variables in predicting the ELP. In contrast, the Holladay II, Olsen, and Barrett Universal II formulas include LT, ACD, and corneal WTW measurement as additional parameters in the ELP calculation. The Haigis formula differs from other formulas in its approach to ELP prediction. It uses three constants that are specifically designed to approximate the nonlinear relationship between AL and IOL position. This approach allows for a more accurate estimation of the ELP without the need for other adjustment factors. In contrast, Olsen’s formula is based on thick-lens theory. When dealing with plus lenses, it is essential to maintain the thick lens anterior to the thin lens equivalent. Understanding the relationship between the actual thick lens and the thin lens equivalent can provide the necessary tools to improve existing formulas, particularly for powers greater than +34 D.[112] An advantage of this thick-lens model is that it accurately positions the ocular components at their true positions. In its latest version, Olsen’s formula incorporates five preoperative parameters to predict the position of the IOL.[113] The Hill–Radial Basis Function (Hill-RBF) formula[114] works without the need for a distinct calculation of the ELP. It uses desired refraction, AL, central corneal power, and ACD to predict the power of the IOL.

Due to the variability in predicting the ELP, slight myopic refractive PEs are observed in the case of the Hoffer Q and Holladay 2 formulas, whereas the Olsen formula produces hyperopic refractive PEs. However, when the mean numerical refractive PE is adjusted to zero, no statistically significant differences are found in the median absolute error among the seven formulas: Barrett Universal II, Haigis, Hill RBF, Hoffer Q, Holladay 1, Holladay 2, and Olsen.[115]

In practice, the ELP is not calculated separately but is determined by the regression relationships (formula constants) that are inherent to the formulas. These constants are model-specific. In essence, the IOL constants, including the A-constants, represent the average effective position of the IOL in the eye. However, these A-constants do not directly provide information about the position of the lens. Nevertheless, they can be converted to an ELP by using the following regression equation: ELP = [(A-constant × 0.5663) – 62.005]/0.9704.[14]

Optimization and personalization of lens constants

Lens constants are the constants provided by the IOL manufacturers. These lens constants, including the A-constant, should be able to make allowances for different powers and changes in lens form, as well as for the fact that different powers tend to be used in eyes with different physical characteristics.

These lens constants are the averages that the IOL manufacturers arrived at many years ago. When new lenses are introduced in the market, there is no recalculation of the constants for these new lenses; the constants are simply extrapolated from similar existing IOLs. Furthermore, constants that work for contact ultrasound do not produce optimum results when using immersion or optical measurements.[116] Hill observed that very-low-power lenses may require a different lens constant than normal power lenses of the same model. Using an existing lens constant optimized for schematic eye parameters may result in a refractive surprise, often a hyperopic one for very low plus-power IOLs and a myopic surprise for minus-power IOLs.[117]

With the exception of the Haigis formula, all other currently popular IOL formulas utilize a single “lens constant” for completion of the calculation; the rest of the terms of the formula are derived from measurable data. These constants are named differently for different formulas. The SRK group of formulas uses the term “А-constant,” Holladay formulas name it “Surgeon’s Factor” [SF], and the Hoffer-Q formula characterizes them as pACD or the “personalized anterior chamber depth.” Because each of these formulas is constructed differently, the constants are also different and cannot be used interchangeably between formulas. However, they do share a commonality. Apart from the lens constant, all other terms of a formula are either fixed, such as corneal refractive index and vertex distance, or directly measured, such as keratometry and AL. This means that lens-specific parameters that might influence the IOL power, such as design, haptic angulation, and material, do not have any bearing on the rest of the formula. Therefore, all these factors can only be incorporated into IOL calculations by including them in the IOL constant. Haigis[104] suggested the use of different constants if classical one-constant IOL power formulas are used to avoid variability in IOL power calculation (Fig. 1) due to formula errors.

Figure 1.

IOL formula variability

Optimization is the process of finding the specific value of a lens constant that, when used for that particular IOL type, will result in the most accurate IOL power calculations. For a group of patients for whom a particular type of IOL has been implanted, optimization is performed by calculating the lens constant in such a manner that the formula produces the exact refractive error that was encountered in that eye.[118] Such a value of the lens constant would make that formula “perfect” for that specific case.

The proper method for optimizing the A-constant is to solve the IOL formula in reverse to find the constant by using input variables such as preoperative corneal and AL measurements, the implanted IOL power, and the stabilized postoperative refraction. An example of the calculation of the A-constant for the SRK formula is provided in Fig. 2. To determine the ideal IOL power, the final, stable, postoperative refraction is required. It is important to remember that the refractive error at the spectacle plane is not the same as that at the IOL plane. To calculate the ideal IOL power by using spectacle plane refraction, a refractive factor (RF) is needed. The RF is a value that needs to be multiplied by the refractive error to give us an estimate of the error at the IOL plane. Retzlaff et al.[119] used an RF value of 1.25 when the implanted IOL power was more than 16.0 D and a value of 1.0 otherwise. Modern estimates of this value typically range from 1.3 to 1.8, with variations depending on eye anatomy. Now that ideal power is available, the А-constant can be calculated. Of course, the problem is that the ideal IOL power may not be accurately calculated using a presumed RF. This is a limitation of this form of optimization.

Figure 2.

Sample back-calculation of A-constant for optimization

The_ptimizs of back-calculation of the А-constant is then repeated for several cases involving the same IOL model and surgery type. In each case, the A-constant is likely to be different. All these values are averaged, and the resultant value is the optimized А-constant. This optimized A-constant can be used for prospective calculations instead of the A-constant provided by the manufacturer. It must be remembered that the А-constant calculated in this manner is specific to the SRK II formula and cannot be used with the SRK/T.[119] Modern theoretic formulas such as the SRK/T, Holladay,[75,112] Hoffer Q,[80,83] and Haigis[25,81] offer an extended scope of calculations, including calculation for ametropia. Therefore, it is possible to predict the refractive error that is expected when a certain power of IOL is implanted. This is because the calculation for RF is internalized in these formulas. This changes the strategy for optimization. There is no attempt to determine the lens constant that would have produced the ideal IOL power, which remains a nebulous entity. Instead, we seek to back-calculate the ametropic equation so that the derived lens constant would predict a refractive error equal to the actual observed refractive error. This approach sidesteps the ideal IOL problem and involves only real, measurable values. In contrast to the relatively simpler calculation for the SRK II formula, the calculations for these other formulas are complex and require the resolution of quadratic roots, making them difficult to understand and perform manually. Dr. Hill stated that a reasonable job of optimizing the lens constant can be done using the software that comes with the biometers.[118] Alternatively, there are free lens constant optimization spreadsheets available, such as those developed by Dr. Saurabh* and Dr. Hill**, which can be used for this purpose.

The Lens Constants Optimizer v5.1 requires the user to fill in columns on a preprogrammed MS Excel^(TM) sheet [Fig. 3]. For each complete record, the software automatically back-calculates the lens constants. These constants are then evaluated for outliers, and any extreme individual values are truncated. This is important as it ensures that any cases that do not form a part of the general group statistically are eliminated. By eliminating these outliers from the analysis, the performance of the formula for “normal” cases can be improved. The program calculates the optimized lens constants for five popular formulas, namely the SRK II, The SRK/T, Hoffer-Q, Holladay 1, and Haigis. Optimizing constants in a two-variable formula such as Hoffer Q shifts the prediction curve to a more accurate location. In contrast, _ptimizing in a three-variable formula, such as the Haigis formula, also reshapes the curve to better account for the changes in the constant at different IOL powers.

Figure 3.

Screenshot of the lens constant optimizer showing the output for a set of sample data

Lens constants calculated for myopic eyes will be different from those calculated for hyperopic eyes. To be truly effective, optimization needs to be done separately for different types of eyes (short, normal, and long eyes). In addition, if different instruments are being used, different constants are needed to produce the best outcomes. One must also realize that optimization is a continuous process that needs to be repeated periodically for it to remain relevant.

A large study by Aristodemou et al.[120] shows that when using optical AL measurements, the benefits of optimization far exceed any differences that the choice of a third-generation formula can provide. This clearly underscores the need for all surgeons to optimize. Dr. Hill stated that “A surgeon who fails to optimize lens constants would have less accurate outcome despite state-of-the-art practice with the best of optical biometers and modern technique of surgery. He may be off 1 D on the myopic or hyperopic side, and only about 16% of his patients will be corrected to within 0.5 D of the intended refraction. However, if he optimizes his constants, he’ll increase this number fourfold and nearly double the number of outcomes within 1 D.”[118] What is, therefore, a minimum significant number of patients required for optimizing your own constants? Holladay noted that it may take anywhere from 20 to 50 cases to calculate your personal constant at the outset, depending on how meticulous one is as a surgeon. A month after collecting the postoperative data on one patient per day would result in 30 eyes to evaluate.[14,109]

Personalization

The process of optimization can be taken further by only considering data from a specific pool.[118] For instance, one such pool might consist of eyes that have been operated upon by surgeon Dr. X by using biometry device B and keratometry device C, with the measurements having been performed by Mr. D. Such narrow focusing is called personalization, and it refines optimization. Personalization allows the incorporation of systematic errors of the measurement devices as well as individual biases of the surgeon (especially surgically induced astigmatism) or technician and further enhances IOL power prediction accuracy. Table 5 reflects these systematic errors/inherent consistent bias. The only way to take all these sources of error is to measure your patients’ postoperative refractions and use that data to continually update your personalized lens constants.

Table 5.

Sources of systematic error/inherent bias

Type of error	Specific error
Measurement/instruments errors	A scan unit/biometer-related Keratometry Refraction
Formulas and IOL-related	Lens position Lens style IOL power accuracy
Surgical technique and lens geometry related	Index of refraction of the cornea Retinal thickness factor Wound closure and suture Postoperative steroids and wound healing

The lens constant can be personalized if optimization is done on a significant number of patients. In our experience, a surgeon who ensures that every instrument works perfectly, always uses the same instruments, and performs procedures in exactly the same manner will be able to calculate their personalized constant with just 20 cases, whereas a surgeon who uses different instruments and works with different technicians at multiple offices may need 50 cases to calculate a statistically significant personalized A-constant. One can use commercial software such as Dr. Holladay’s IOL Consultant or Dr. Saurabh’s Dynamic IOL Optimization (DIO) for more precise calculation and personalization of A-constants.

Dynamic IOL Optimization [Fig. 4] is a powerful personalized analysis system that bypasses the conventional optimization of the lens constants and instead focuses on the relative performance of the formulas as a whole. Conceptually, it can compare an infinite number of IOL power calculation algorithms. The software has a user-friendly interface that runs on the MS Excel^(TM) platform. The user is required to fill in the IOL model names, the surgeons’ names, etc., as a one-time exercise, with an option for later additions. Following this, the database is created, wherein the user enters case details including biometry, IOL details, and postoperative refraction. A minimum of 11 complete entries are required before the program can generate optimized IOL powers. When IOL power calculation is required, the user enters the case-specific biometry details and the chosen IOL model. The program then automatically scans the database and chooses a niche cohort. This cohort comprises eyes that have a structural configuration that is similar to the test eye. The parameters for this selection are AL and keratometry. This ensures that when optimizing the IOL power for an unusual eye, for example, a myope, only the matching portion of the database that contains similarly myopic eyes will be evaluated. Once the niche cohort is chosen, the program then automatically evaluates the relative performance of different IOL formulas in that cohort. Outliers are automatically excluded. This information is then prospectively applied to the test case, yielding a single, usable IOL power. Dynamic IOL optimization offers several advantages to the surgeon. First, it is easy to use. The entire process of DIO is facilitated by a very simple user interface. Second, there is no need for an additional equipment purchase as it works to make the most of the existing data. Third, the user need not choose a formula as per the ocular configuration. Instead, there is a smooth surface of prediction based on the surgeon’s own clinical outcomes. Fourth, because lens constants are bypassed, there is no need to consider separate values for the contact, immersion, or optical methods of measuring AL. Fifth, the program works continuously. As new data are added, the optimization protocol recalculates everything. New information is thus constantly incorporated into the system. This is better than optimizing the lens constants every now and then, doing it for different formulas, and for different AL ranges. Because cohort selection is continuous rather than discrete, there is zero data wastage. Hence,, by using this software, personalization of theoretic formulas is not overly cumbersome as it may seem.

Figure 4.

Screenshot of Lens Constant Optimizer showing the output for a set of Sample Data

Conclusion

These are exciting times for IOL power calculation, with everyday advancements in technology introducing new devices and IOLs to the market, all aiming to provide the best visual outcomes for cataract surgery patients. With new-generation formulas, IOL power calculation has become more accurate; however, there is a need to enhance the results to achieve zero PE in most cases. Optimizing and personalizing lens constants should be a fundamental part of any cataract surgery practice. Several strategies exist for the optimization and personalization of lens constants for a surgeon committed to better refractive outcomes of his cataract surgery. These range from software integrated into the biometry machines, through free tools and spreadsheets, to standalone, commercial systems such as Saurabh’s DIO and Holladay IOL Consultant. As many of these new approaches find acceptance, things can only get better for cataract surgery patients.

Online resources

https://rbfcalculator.com/online/index.html

https://calc.apacrs.org/barrett_universal2105

https://ascrs.org/tools/barrett-toric-calculator

https://iolcalculator.escrs.org

http://ocusoft.de/ulib/c1.htm

http://www.hicsoap.com

https://doctor-hill.com/iol-power-calculations/formulas

*Due to copyright concerns raised by Hoffer, the calculator is no longer available in the public domain. However, the authors may be contacted for conventional optimization of lens constants by individual surgeons, for personal, non-commercial uses.

**

https://doctor-hill.com/iol-power-calculations/resources-downloads

Financial support and sponsorship:

Nil.

Conflicts of interest:

There are no conflicts of interest.