The diamond knife

What made radial keratotomy possible was not the blade. It was the footplate.

A diamond edge crosses the corneal stroma effortlessly; the difficulty was never cutting. It was always stopping. Too superficial, and the incision achieves nothing. Too deep, and it opens the anterior chamber. What turned an idea into a reproducible operation was a mechanical stop: a diamond mounted so that only a set length protrudes below a footplate resting on the epithelium. Depth belongs to the instrument, not to the surgeon’s hand — incision after incision, eye after eye.

Everything else about radial keratotomy is arguable. That design principle is not.

A first attempt, and its price

Tsutomu Sato began incising corneas in Japan in the late 1930s, from a simple observation: keratoconic eyes flatten after a break in Descemet’s membrane. His technique combined anterior incisions with posterior ones, the latter made from within the anterior chamber, directly through the endothelium.

The refractive result was real. The bill arrived twenty years later, when a large proportion of those eyes decompensated into bullous keratopathy. The endothelium does not regenerate, and at the time nobody was counting cells.

The lesson is structural rather than technical: the cornea accepts an intervention immediately and settles the account decades later. Every refractive procedure has been judged on that timescale ever since, rightly or otherwise.

Fyodorov and the industrial phase

Sviatoslav Fyodorov revived the technique in the Soviet Union in the 1970s, stripped of its posterior incisions: anterior only, the endothelium spared. Mid-peripheral radial incisions deliberately weaken the paracentral cornea; under the effect of intraocular pressure, the periphery bulges and the centre flattens. Myopia falls.

The scale of its spread in the 1980s is hard to picture today: hundreds of thousands of eyes operated on in the United States, and in Moscow an assembly-line arrangement in which the patient moved from one surgeon to the next, each performing one step. Refractive surgery became a high-volume speciality before it had ten years of follow-up.

What the instrument had to solve

The nomogram carried the refractive intention: number of incisions, diameter of the central optical zone, age, intraocular pressure, target correction. The knife carried the risk.

Depth. In the PERK protocol, the blade was set by micrometer against intraoperative ultrasound pachymetry — to the full value of the thinnest paracentral thickness, on the grounds that a blade advancing through compressible tissue never reaches its nominal depth. Deep incisions give more effect and more instability: the whole risk-benefit balance of the operation is played out in the last fifty micrometres.

Direction. Cutting centrifugally, from the edge of the optical zone towards the periphery, protects the visual axis but tends to produce shallower incisions where the effect is greatest. Cutting centripetally gives more effect but risks encroaching on the optical zone. Russian and American schools never agreed, and hybrid techniques cut in both directions in a single pass.

The geometry of the edge. Single bevel, double bevel, vertical facet: each profile behaves differently in tissue, and each drifts as the diamond wears. The instrument was a consumable that presented itself as an investment.

What the data showed

Radial keratotomy was studied with more rigour than most procedures of its era. The PERK study followed 793 eyes across nine centres, using a standardised eight-incision technique.

Ten years after surgery, 53 % of operated eyes saw 20/20 uncorrected and 85 % reached 20/40 or better. Among patients operated on in both eyes, 70 % wore no distance correction. Losses of best corrected acuity were rare — two lines or more in 3 % of eyes.

And 43 % of eyes had drifted towards hyperopia by at least one dioptre. That drift was rapid at first — about 0.21 D per year between six months and two years — then continued indefinitely at roughly 0.06 D per year, with no observable endpoint, and no preoperative variable allowed it to be predicted. To this were added a diurnal fluctuation, the incisions opening and closing with corneal hydration, halos and glare related to the scars, and increased spherical aberration and coma.

The operation worked. It simply never stopped working.

These eyes are now in the cataract clinic

Radial incisions do not disappear. They remain permanent lines of weakness, and they break the relationship between anterior curvature and total corneal power on which the usual biometric formulas rest. Both the surgical manoeuvre and the lens calculation have to be adapted — this is now one of the cataract surgeries with the most specific requirements.

What matters here is simpler: the excimer laser supplanted radial keratotomy in the 1990s not because it cut more precisely, but because it removed tissue instead of dividing it. A photoablated cornea is thinner but stable; an incised cornea is intact but lastingly weaker.

What survived

The diamond knife did not disappear with the technique that made its name.

The same instrument, the same footplate logic, still opens limbal relaxing incisions: paired arcs at or near the limbus, on the steep meridian, for the mild corneal astigmatism met at the time of cataract surgery. Depth is set the same way, against pachymetry — around 90 % of local thickness in current practice — and the nomograms descend in a direct line from those written for radial keratotomy.

It has become a minority technique, squeezed from both sides. Toric IOLs correct the same astigmatism inside the eye, with better predictability and no corneal wound. Femtosecond platforms trace the same arcs at programmed depth, without manual variability. A narrow but real indication remains for the blade: modest astigmatism, an eye already open, no toric IOL available or justified, and a decision taken at the table rather than in the planning software.

The legacy

Take away radial keratotomy’s refractive claims: one idea remains intact. Control of depth belongs to the geometry of the instrument, not to the skill of the hand holding it.

That principle reads directly in the mechanical stop of the microkeratome, in the programmed depth of a femtosecond lamellar cut, in the pachymetry-before-anything reflex of every corneal surgeon trained since. We inherited the discipline and abandoned the operation.

Techniques pass. The discipline that grounds the indication endures.

References

1. Sato T, Akiyama K, Shibata H. A new surgical approach to myopia. Am J Ophthalmol. 1953;36(6 Pt 1):823-829.
2. Fyodorov SN, Durnev VV. Operation of dosaged dissection of corneal circular ligament in cases of myopia of mild degree. Ann Ophthalmol. 1979;11(12):1885-1890.
3. Waring GO 3rd, Lynn MJ, Nizam A, et al. Results of the Prospective Evaluation of Radial Keratotomy (PERK) Study five years after surgery. Ophthalmology. 1991;98(8):1164-1176.
4. Waring GO 3rd, Lynn MJ, McDonnell PJ. Results of the Prospective Evaluation of Radial Keratotomy (PERK) Study 10 years after surgery. Arch Ophthalmol. 1994;112(10):1298-1308.
5. Budak K, Friedman NJ, Koch DD. Limbal relaxing incisions with cataract surgery. J Cataract Refract Surg. 1998;24(4):503-508.
6. Nichamin LD. Nomogram for limbal relaxing incisions. J Cataract Refract Surg. 2006;32(9):1408.