Preoperative screening algorithm for LASIK, PRK, SMILE & phakic ICL — topography, pachymetry, and ectasia risk stratification
Every refractive surgery evaluation begins with a battery of objective measurements designed to answer one question before any other: is this cornea structurally and optically stable enough to have tissue removed or a lens implanted? Modern Scheimpflug tomography has made subclinical ectatic disease detectable years before it would show up on refraction alone.
Placido-disk topography projects concentric illuminated rings onto the tear film and reconstructs anterior corneal curvature from the reflected ring pattern — fast, well validated, but limited to the anterior surface only. Modern screening has largely moved to Scheimpflug tomography (Oculus Pentacam, and comparable systems such as Galilei dual-Scheimpflug-Placido and slit-scanning Orbscan), which rotates a camera through the optical axis to reconstruct both anterior AND posterior corneal elevation, full-thickness pachymetry mapping, and anterior chamber geometry in a single scan.
The key output for candidacy screening is the Belin-Ambrósio Enhanced Ectasia Display (BAD-D), a composite z-score that compares a patient's anterior elevation, posterior elevation, and pachymetric progression indices against a normative database. A BAD-D final value (D) above roughly 1.6–2.6 is considered suspicious for subclinical ectasia; values below that range are reassuring but not, by themselves, sufficient to clear a patient with other red flags present.
Corvis ST adds a biomechanical dimension: a rapid air-puff deforms the cornea while a high-speed Scheimpflug camera records the deformation response, generating parameters correlated with corneal stiffness. The older Ocular Response Analyzer (ORA) pioneered this with corneal hysteresis (CH) and corneal resistance factor (CRF) — corneas with reduced biomechanical stability can appear geometrically normal yet still be at elevated ectasia risk.
Forme fruste keratoconus — a topographically mild, often asymmetric pattern with no slit-lamp signs — is the single most dangerous miss in refractive screening: eyes that appear to pass by manifest refraction and slit-lamp exam alone have historically accounted for a disproportionate share of post-LASIK ectasia cases.
Central corneal thickness (CCT) is measured by ultrasound pachymetry (the clinical reference standard) or optical pachymetry embedded in the tomographer; both the central value and the thinnest pachymetric point (which is often slightly inferotemporal to center, not literally central) matter, since ectatic corneas thin asymmetrically before they steepen. Population-average CCT is roughly 540–560 µm; corneas at or below 480–500 µm markedly narrow the safety margin for any tissue-removing procedure.
Scotopic (dark-adapted) pupil diameter is measured because larger scotopic pupils increase the chance that the peripheral, untreated cornea contributes stray light to the retinal image after a small optical zone ablation — historically linked to night-time glare and halos, though modern large-optical-zone and wavefront-guided ablations have substantially reduced this risk relative to older-generation lasers.
Refractive stability is a hard gate: manifest refraction must be stable to within about 0.5 D over the prior 12 months, documented by at least two refractions. Unstable refraction — common in patients under 21, in early keratoconus, and in uncontrolled diabetes with fluctuating lens hydration — is an automatic deferral regardless of how favorable the tomography looks.
Dry eye is both a contraindication in severe form and the most common patient complaint after any corneal refractive procedure, because the flap or PRK epithelial defect transects corneal nerves and transiently reduces blink-triggered reflex tearing. Standard workup includes tear breakup time (TBUT, normal ≥10 seconds), Schirmer testing (basal or reflex, normal ≥10 mm/5 min), the OSDI questionnaire (Ocular Surface Disease Index, scored 0–100), and meibomian gland evaluation by meibography or expression, since meibomian gland dysfunction is the leading cause of evaporative dry eye in this population.
Surgeons typically treat significant dry eye and meibomian gland disease preoperatively — with lubrication, punctal plugs, topical anti-inflammatories, or lid hygiene — before finalizing candidacy, rather than treating it as an outright exclusion in mild-to-moderate cases.
Once the screening data is in hand, procedure selection becomes a tissue-budget problem: how much stroma can be safely removed or displaced, and how much refractive error needs correcting, within the thickness available. Percent Tissue Altered formalizes that budget into a single safety threshold.
Santhiago and colleagues formalized Percent Tissue Altered (PTA) as: PTA = (flap or cap thickness + ablation depth) ÷ preoperative central corneal thickness × 100. Ablation depth scales roughly 12–16 µm per diopter of spherical equivalent correction for standard myopic treatments (larger for high astigmatism or hyperopic treatments that ablate the periphery).
Retrospective analysis of post-LASIK ectasia cases found PTA above 40% was strongly associated with ectasia even in eyes that had a numerically "safe" residual stromal bed (RSB) by older RSB-only criteria — meaning PTA and RSB are complementary, not interchangeable, checks. A cornea can pass an RSB-only rule of ≥250 µm and still fail on PTA if the flap itself is unusually thick relative to a thin starting cornea.
Residual stromal bed is simply CCT minus flap/cap thickness minus ablation depth; the field consensus safety floor is 250 µm, with many surgeons preferring a 300 µm buffer for higher-risk profiles (young age, high myopia, borderline topography).
A patient with 500 µm CCT correcting −8.00 D with a 120 µm femtosecond flap: ablation depth ≈ 8 × 14 = 112 µm. RSB = 500 − 120 − 112 = 268 µm (passes the 250 µm floor but is below the 300 µm comfort margin), while PTA = (120+112)/500 × 100 ≈ 46% — over the 40% threshold. This is exactly the scenario PTA was designed to catch: PRK or SMILE (no flap, or a thinner cap) would substantially lower that number.
LASIK (Laser-Assisted in Situ Keratomileusis): a femtosecond laser (or, less commonly today, a mechanical microkeratome) cuts a hinged corneal flap (femto: ~100–120 µm; microkeratome: ~130–180 µm and less predictable), the flap is lifted, an excimer laser reshapes the underlying stroma, and the flap is repositioned. Fast visual recovery (often 20/20 by next day) but the flap interface is a lifelong structural weak point and the primary reason PTA/RSB math matters most here.
PRK (Photorefractive Keratectomy) and its variants (LASEK, epi-LASIK): the epithelium is removed (mechanically, with dilute alcohol, or with a laser) and the excimer ablation is applied directly to Bowman's layer and anterior stroma — no flap. Slower visual recovery (days, with transient discomfort) but no flap-related RSB penalty, making it the default fallback for thinner corneas or lower PTA headroom.
SMILE (Small Incision Lenticule Extraction): a femtosecond laser cuts an intrastromal lenticule (with a cap of about 120–130 µm) that is removed through a small 2–4 mm incision — no flap is lifted. Comparable tissue removal to LASIK but a more intact anterior cornea/nerve plexus, associated with somewhat less induced dry eye; currently approved chiefly for myopia and myopic astigmatism.
ICL (Implantable Collamer Lens / phakic IOL): a foldable lens is placed in the posterior chamber, in front of the natural crystalline lens, without removing any corneal tissue at all. Because it sidesteps the PTA/RSB problem entirely, ICL becomes the procedure of choice for high myopia, thin corneas, or any patient who fails the tissue-budget checks — provided anterior chamber depth and endothelial cell count are adequate and vault (the space between the ICL and the native lens) can be targeted to roughly 250–750 µm.
Corneal curvature (K readings) sets its own boundaries: post-treatment corneas that end up flatter than roughly 36 D or steeper than about 49–50 D can produce poor optical quality regardless of tissue thickness, which constrains how much myopic or hyperopic correction any surface-based (LASIK/PRK/SMILE) procedure can safely deliver — very high corrections often push toward ICL instead.
Occupation and lifestyle round out the decision: contact-sport athletes and military candidates are frequently steered away from LASIK's flap (dislocation risk on blunt trauma) toward PRK or SMILE; patients with significant night-driving demands and large scotopic pupils benefit from larger optical zones and wavefront-guided profiles; and patients wanting the fastest possible visual recovery for occupational reasons often prefer LASIK's next-day results over PRK's multi-day healing window.
A subset of findings on topography, tomography, and systemic history function as hard or near-hard stops. Recognizing them early avoids irreversible harm: unlike glasses or contact lenses, a laser ablation cannot be undone, and post-refractive ectasia is one of the most feared complications in elective ophthalmic surgery.
The classic topographic signature of ectatic disease is inferior steepening — an asymmetric bowtie pattern with the inferior lobe steeper than the superior lobe — often quantified by indices such as the I-S (inferior-superior) value. A skewed radial axis (SRAX), where the steepest meridians above and below the horizontal are not mirror-symmetric, is a subtler and often earlier sign than inferior steepening itself.
On Scheimpflug tomography, high posterior elevation (posterior float exceeding the normative range at the thinnest point) frequently precedes any visible anterior curvature change, which is precisely why posterior-surface analysis is considered more sensitive for detecting forme fruste keratoconus than anterior topography alone. Kmax — the steepest simulated keratometry value on the map, not the same as the average central K — climbing to 47.2 D or higher, especially paired with focal inferior or paracentral steepening, is a widely used numeric red flag.
The Ambrósio Relational Thickness (ART) index relates pachymetric progression (how rapidly the cornea thickens from its thinnest point outward) to Kmax; abnormally low ART values indicate a thin point that is both anatomically thin and progressing abnormally fast — a combination far more specific for ectatic risk than thickness or curvature considered separately.
Forme fruste keratoconus can be present in a cornea with normal Placido topography and 20/20 corrected vision — the posterior elevation and pachymetric progression indices on Scheimpflug tomography are frequently the only clues, which is exactly why tomography (not topography alone) is now the screening standard of care.
Central corneal thickness below roughly 480–500 µm sharply narrows the margin for any tissue-removing procedure and, combined with any topographic irregularity, is generally an outright LASIK/PRK/SMILE exclusion (ICL remains an option since no stroma is removed). Progressive or unstable refraction — more than about 0.5 D of change within the prior 12 months — must be resolved before proceeding, since operating on a cornea that is still actively changing shape risks compounding an undiagnosed ectatic process with an iatrogenic one.
Systemic contraindications include active or poorly controlled autoimmune/connective-tissue disease (e.g., active rheumatoid arthritis, systemic lupus erythematosus, Sjögren syndrome), which impairs corneal wound healing and raises the risk of sterile keratolysis; uncontrolled diabetes mellitus, which both destabilizes refraction via lenticular hydration shifts and impairs epithelial healing; active herpetic (HSV/VZV) keratitis, since excimer ablation can reactivate latent virus and trigger vision-threatening stromal keratitis; and pregnancy or nursing, during which hormonally driven corneal hydration changes make refraction unreliable until several months postpartum.
Severe dry eye that fails preoperative treatment is a relative-to-absolute contraindication, since any corneal refractive procedure will transiently worsen it further via nerve transection.
Before Randleman's score, ectasia risk assessment was largely gestalt — an experienced surgeon's judgment across several independent variables. The 2008 system converted that judgment into a reproducible point total, explicitly weighting topographic pattern above all other inputs because it was found to be the single strongest predictor.
Topographic pattern (0–4 points) — the heaviest-weighted variable: 0 for normal, 1 for asymmetric bowtie, up to 3–4 for skewed radial axis or a pattern suspicious for (or diagnostic of) keratoconus/pellucid marginal degeneration. This component alone can push a patient into the high-risk band regardless of every other input.
Residual stromal bed (0–4 points): scored on a sliding scale where RSB ≥350 µm scores 0, and RSB progressively below 300 µm, 250 µm, and 200 µm adds points — directly operationalizing the same tissue-budget logic behind PTA.
Age (0–2 points): younger patients score higher (2 points for under 30, less for older patients) because younger corneas — particularly in the 18–25 range — have a longer time horizon over which subclinical ectatic changes can progressively manifest, and because keratoconus itself typically has its onset and most rapid progression in the second and third decades of life.
Preoperative corneal thickness (0–1 point) and manifest refraction spherical equivalent (0–1 point) contribute smaller, single-point weightings — thin CCT and high myopia both independently associate with ectasia risk, but far less strongly than topographic pattern or RSB.
In the original validation, eyes scoring ≥3 points had a relative risk of ectasia roughly 20-fold higher than eyes scoring 0 — but the single highest-weighted input, abnormal topography, could by itself account for most of that risk even when every other component scored zero.
Total scores are conventionally banded as low risk (0–2 points, LASIK generally considered safe from an ectasia standpoint pending the rest of the workup), moderate risk (3–4 points, proceed with caution — consider PRK instead of LASIK to remove the flap-related contribution, or defer), and high risk (≥5 points, LASIK is discouraged; surface ablation at reduced correction, SMILE, or ICL are the safer alternatives, or refractive surgery is deferred entirely).
The score was developed and validated in an era before routine posterior-elevation and biomechanical analysis (Corvis ST, ORA) were widely available, so contemporary practice generally uses it as one input layered on top of — not a replacement for — full Scheimpflug tomography, BAD-D, and ART index review. A patient can score low on the Randleman system yet still show a suspicious posterior elevation or abnormal biomechanics that independently argues against LASIK.
The score also does not directly incorporate family history of keratoconus or a history of eye rubbing (a mechanical risk factor increasingly implicated in ectatic progression), both of which experienced surgeons still weigh qualitatively alongside the numeric total.
The last step before surgery translates every prior measurement into an actual laser ablation (or lens power) plan, and forces an explicit, documented go/no-go decision — the point where screening data either clears a patient for a specific procedure or redirects them to a safer alternative, or to no surgery at all.
Wavefront-guided ablation uses an aberrometer to measure the eye's entire optical wavefront — not just sphere and cylinder, but higher-order aberrations (HOAs) like coma, trefoil, and spherical aberration expressed as Zernike polynomial coefficients — and programs the excimer laser to correct that full wavefront rather than a manifest-refraction approximation of it. This generally induces fewer new HOAs than conventional ablation, particularly relevant for larger scotopic pupils and night-vision symptoms.
Topography-guided ablation (e.g., contoured ablation profiles derived directly from the anterior corneal shape map) instead targets corneal surface irregularity itself, and is particularly useful for eyes with irregular astigmatism, decentered prior ablations, or mild topographic asymmetry that falls short of an outright red flag — smoothing the corneal surface as part of the refractive correction.
Both approaches still operate within the same PTA/RSB tissue budget established earlier — customization changes where and how the ablation is distributed across the optical zone, not the total safety ceiling on how much tissue may be removed.
By this stage every prior data stream — tomography (BAD-D, ART, posterior elevation), pachymetry, pupil size, refractive stability, dry-eye status, PTA/RSB arithmetic, and the Randleman score — is reviewed together. A patient who is clean across all of these is typically offered a choice between LASIK, PRK, and SMILE based on lifestyle preference and healing-time tolerance; a patient who fails any single hard gate (CCT, topographic red flag, instability, systemic disease) is redirected to PRK at reduced correction, to ICL, or to non-surgical management, or their evaluation is deferred and repeated.
For ICL candidates specifically, planning shifts to anterior chamber depth (generally required ≥3.0 mm), endothelial cell density, and white-to-white/sulcus-to-sulcus measurements to size the lens; postoperative vault — the gap between the implanted lens and the native crystalline lens — is targeted to roughly 250–750 µm, since too little vault risks anterior subcapsular cataract from lens touch, and too much vault risks angle-closure glaucoma from pupillary block.
A well-run candidacy screen is deliberately conservative: across large modern series, on the order of 10–20% of self-referred LASIK candidates are declined or redirected to an alternative procedure after full tomographic and biomechanical workup — the algorithm is designed to say no often enough that the "yes" cases are genuinely safe.
Candidacy screening does not end at the surgical decision — it extends into structured postoperative follow-up, because a small fraction of eyes that pass every preoperative check still develop late ectatic change, and outcomes tracking is how the screening algorithm itself gets validated and refined over time.
In contemporary series using femtosecond flap creation and wavefront- or topography-guided ablation, roughly 90–95% of LASIK eyes achieve uncorrected distance visual acuity of 20/20 or better, with the large majority within one line of their best spectacle-corrected vision. PRK and SMILE achieve comparable final acuity, differing chiefly in the time course to get there — PRK eyes typically take one to several weeks to reach their eventual endpoint, versus next-day stabilization for LASIK.
A residual refractive error outside the target range prompts consideration of an enhancement procedure — most often a flap lift-and-retreat for LASIK (feasible for years after the original surgery given adequate residual tissue) or a fresh surface ablation for PRK/SMILE eyes. Enhancement rates run roughly 1–5% depending on the magnitude of the original correction (higher corrections regress or under/overshoot more often) and are themselves subject to the same PTA/RSB tissue-budget check as the index procedure — an eye with minimal remaining stromal headroom may not be a safe enhancement candidate even if its distance vision is imperfect.
Historical estimates place post-LASIK ectasia incidence at roughly 0.04–0.6%, with the wide range reflecting both the era of surgery (pre- versus post-modern-tomography screening) and how aggressively risk factors were screened in the source population — more recent series with routine Scheimpflug tomography and PTA/Randleman-informed selection report figures toward the low end of that range. Ectasia can present anywhere from months to more than a decade after an uneventful procedure, which is why isolated case reports of very late presentation continue to appear in the literature even in well-screened cohorts.
Because of that latency, many surgeons recommend periodic tomography — not just refraction and visual acuity checks — for at least several years postoperatively, watching specifically for the same posterior elevation, pachymetric progression, and Kmax trends used in preoperative screening. When late ectasia does occur, management options include corneal collagen cross-linking (CXL) to halt progression, rigid or scleral contact lenses to restore vision, and in advanced cases, intracorneal ring segments or corneal transplantation — underscoring why the preoperative red-flag and Randleman screening steps earlier in this workflow exist in the first place: prevention is far more effective than any available treatment for established ectasia.
Corneal collagen cross-linking (riboflavin/UV-A, the Dresden protocol and its accelerated variants) was developed specifically to halt ectatic progression by strengthening stromal collagen cross-links — it has become both a treatment for post-refractive ectasia and, increasingly, a consideration for prophylactic combination with LASIK in borderline-risk eyes.