Morphometric analysis of corneal endothelium following radial keratotomy

articles Morphometric analysis of corneal endothelium following radial keratotomy Keizo ehiba, M.D. Kazuo Tsubota, M.D. Setsuko S. Oak Ronald A. Laing, Ph.D. Boston, Massachusetts Joseph Goldstein R.Ph., eOA Sanford Hecht, M.D. Waltham, Massachusetts ABSTRACT We performed morphometric analysis of the central corneal endothelium on 24 eyes of 19 patients who had had anterior radial keratotomy. The endothelium was analyzed for a variety of parameters , including cell area, perimeter, side lengths, cell shape, and number of sides. Mean, standard deviation, and coefficient of variation were calculated for each parameter. The mean cell density decreased from 2,835 to 2,677 cells/mm 2 , mean cell perimetel" incl'eased from 71.4 J.Lm to 74.3 J.Lm, and mean side length increased from 11.8 J.Lm to 12.3 J.Lm following surgery. The changes in these three parameters were statistically sign incant (P < 0.05). The coefficient of variation of cell area (polymegathism) changed from 0.319 to 0.307, the hexagonality changed from 62.5% to 59.6%, and the cell shape changed from 0.872 to 0.867. The changes in these parameters were not statistically different before and after surgery. The group of patients that had no reported microperforations showed only a small decl'ease of cell density (1.6%), while the gmup of patients that had microperforations showed a large decrease of cell density (14.3%). The cell perimeter and side lengths showed a similar pattern. The group of corneas with the optical zone diameter less than 3.5 mm showed a decrease in mean cell density fl"om 2,994 to 2,725 cells/mm 2 , and the cell shape changed from 0.874 to 0.866 following surgery. The changes in these parameters were statistically signincant (P < 0.05) before and after surgery. Among all factors associated with radial keratotomy, microperforation and a small diameter of central optical zone appear to be the two greatest risk factors. Key Words: corneal e ndothelium , radial keratotom y, specular microscopy morphome tric analysis Although anterior radial keratotomy does not involve direct incision to the corneal endothelium, minimal or signincant endothelial cell loss has been reported.1- 8 However, few clinical factors that allow the surgeon to predict reasonably which factors or

surgical procedures will have an unfavorable response on the corneal endothelium have yet been identified. Morphometric analysis of the corneal endothelium 9 can provide considerable quantitative information concerning endothelial changes resulting from ocular

From Boston Universit!! School of Medicine, Boston, and Boston Eye Research Institute, Waltham, Massachusetts. Presented at the Symposium on Cataract, IOL and Refractive Surgery, Los Angeles, April 1986. Supported by NEI grants EY 01227 and EY .5924. Reprint requests to Ronald A. Laing, Ph.D. , Department of Ophthalmology, Boston University School of Medicine, 80 East Concord Street, Boston, Massachusetts 02118.

J CATARACT

REFRACT SURG-VOL 13, MAY 1987

263

surgery or diseases and can also reveal subtle changes not apparent by other methods. Although recent studies of radial keratotomy3,8 used computer-assisted morphometric analysis, most previous studies 1,2,4-7 have measured only changes in endothelial cell density following surgery. In this article, we report the use of computerassisted morphometric analysis to analyze the photographs of corneas subjected to radial keratotomy to identify several risk factors introduced by radial keratotomy to the corneal endothelium. MATERIALS AND METHODS Twenty-four eyes of 19 patients who had had anterior radial keratotomy performed at the Boston Eye Research Institute were studied. All patients in this study were 20 years of age or older (mean age 33.7 years) and had a refractive range before surgery of - 1.75 to - 9.00 diopters of myopia (spherical equivalent, and mean diopter - 4.90). Patients who had previous ocular surgery or corneal disease were excluded from this study. Both prior to and following surgery, all patients had an ophthalmologic examination, including refraction, applanation tonometry, corneal pachymetry, slitlamp and ophthalmoscopic examinations, and specular microscopy. The time duration between surgery and the postoperative examination ranged from 6 to 30 months. Twelve specular microscopic photographs of the central portion of each cornea were taken using the Bio-Optics LSM 2000C at a magnification on the negative of 100x. All eyes had eight radial incisions using a diamond or metal blade set at a depth of 100% of central corneal thickness. Diameters of the optical zone were 3 mm to 4.5 mm. Two hundred cells were digitized and analyzed for each cornea using the Bio-Optics Endothelial Image Analysis System by digitizing the cell corners using the "corner" program. U sing the "corner" and "histo" programs, the endothelium was analyzed for a variety of parameters, including cell area, cell perimeter, cell shape, number of sides, cell side lengths. The mean, standard deviation, coefficient of variation, and skewness of each parameter were calculated automatically. Cell density (cells/mm 2 ) was calculated by dividing 106 by the mean cell area. RESULTS The precision with which the digitization and subsequent calculation could be performed was measured to be less than 5%, which was comparable to previous reports. 10 The data obtained are summarized in Tables 1 and 2. Although there were significant (P < 0.05) differences among all the corneas in mean endothelial cell density 264

(2,835 :±: 432 to 2,677 :±: 528 cells/mm 2 ), mean perimeter (71.4 :±: 5.4 to 74.3 :±: 7.9 J.Lm), and mean side lengths (ll.8 :±: 0.9 to 12.3 :±: 1.4 J.Lm) before and after radial keratotomy, the coefficient of variation of cell area (polymegathism) showed no significant difference. There were also no significant differences in the hexagonality (62.5 :±: 9.7 to 59.6 :±: 6.8) and cell shape (0.872 :±: 0.007 to 0.867 :±: 0.016). For purposes of analysis, the corneas were divided into several groups. In the group of 14 corneas that had no reported microperforations, there was no significant change in endothelial cell density (2,924 ± 461 to 2,876 + 364 cells/mm 2 ) (1.6% decrease), perimeter (70.6 ± 5.5 to 71.4 ± 4.6 J.Lm) (1.1% increase), and side lengths (ll.8 ± 0.8 to ll.9 ± 0.8 J.Lm) (1.0% increase) before and after surgery. However, the group of corneas with microperforations showed significant changes in cell density (2,825 ± 471 to 2,421 ± 603 cells/mm 2 ), mean perimeter (72.0 ± 5.3 to 78.2 ± 10.0 j..l.m), and mean side lengths (ll.4 ± 0.9 to 13.0 ± 1. 9 J.Lm). These differences were statistically significant (P < 0.05, Table 3). The group of corneas with an optical zone diameter less than 3.5 mm showed a decrease in mean cell density from 2,994 ± 420 to 2,725 ± 594, an increase in mean cell perimeter from 71.0 to 73.5, and an increase in mean side lengths from 11.6 to 12.2. The cell shape also showed changes from 0.874 ± 0.005 preoperatively to 0.866 ± 0.017 postoperatively. Thes"e differences were statistically significant (P < 0.05). However, the group of corneas that had an optical zone diameter greater than 3.5 mm showed no significant differences before and after radial keratotomy (Table 4). There were no significant differences between the group of patients that had only one operation and the group that had more than one. Nor was there any significant difference between the group of contact lens wearers and the group of nonwearers. For the degree of myopia, no significant differences were found in the group of eyes that had a myopic range of less than - 4.75 D (average of this study group) before surgery and eyes with over - 4.75 D myopic range. No significant changes in endothelial polymegathism and change of hexagonality in any of the groups were seen. Surprisingly, there was an increase in cell density (from 0.3% to 15.3%, mean 4.7%) in five eyes (Table 1). DISCUSSION Although some published studies of human and animal models of radial keratotomy show minimal or no endothelial cell loss after radial keratotomy, 1,2,8 other studies report a significant cell loss of up to 16%.3-7 Many investigators have reported that endothelial cell

J CATARACT REFRACT SURG-VOL

13, MAY 1987

Table 1. Patient data.

Follow-up Sex (months)

Contact Lens (years)

Myopia (diopters)

Number

Age (years)

32

F

24

3

-4.75

2

31

M

30

o

-2.50

Optical Perforation (1-3) Reoperation Zone (mm)

Preoperative Postoperative Shape Shape

3 2

34.0

0.877

0.851

3.25

25.0

0.878

0.804

3.50

+

Cell Loss (%)

3

59

M

7

35

-9.00

3.00

2

22.8

0.875

0.859

4

30

F

20

3

-.5.00

3.00

2

14.0

0.870

0.873

5

32

F

30

3

-5.75

3.50

0.870

0.874

20

F

6

-4.75

3.00

7.8

0.872

0.872

7

29

M

6

5 o

o o

11.5

6

-4.00

3.00

o

5.3

0.878

0.872

8

30

M

7

1

-5.75

3.00

3

4.9

0.874

0.873

3.00

4.7

0.872

0.873

4.1

0.872

0.879

3.6

0.870

0.874

9

30

M

8

o

-7.75

10

30

M

24

7

-4.75

3.20

11

5

-4.25

3.00 3.50

o o o o

3.6

0.870

0.874

3.00

1

3.3

0.878

0.863

20

F

12

29

M

8

12

-4.25

13

34

F

10

10

-3.25

14

39

M

11

1

-3.25

3.75

2

3.2

0.850

0.863

15

55

F

19

o

-8.50

3.00

1

3.2

0.870

0.874

16

28

M

19

2

-3.00

3.00

0.881

0.877

39

F

9

10

-7.00

3.00

o o

2.5

17

2.0

0.875

0.866

18

31

M

8

7

-1.75

4.50

2

1.5

0.867

0.822

19

42

F

8

o

-4.00

3.50

0.7

0.874

0.876

20

29

M

6

12

-3.25

3.50

o o

+

0.3

0.869

0.869

21

22

M

6

5

-7.00

3.50

1

+ 1.8

0.869

0.863

22

42

M

6

4.00

42

M

14

o o

-5.00

23

+ +

24

21

F

9

5

-6.25

o o o

11

+

+ +

-3.75

3.00

loss was nonprogressive after surgery.l,3-.5,8 Even though cell loss due to radial keratotomy may not be progressive, cell loss due to normal aging alone (normal rate 0.03S% to 0.71 % loss per year)ll may contribute to the effects on the endothelium, and possibly make Table 2. Summary of changes. Preoperative Mean ± SO

4.00

Postoperative % P< Mean ± SO Increase 0.05

2.3

0.864

0.863

3.5

0.869

0.869

+ 15.3

0.872

0.873

corneas that had radial keratotomy more susceptible to later bullous keratopathy. Minimizing the cell loss due to radial keratotomy as well as identifying the risk factors that contribute to cell loss are therefore important. Our results show that the group of corneas with microperforations and the group with an optical zone diameter less than 3.S mm had significant cell loss (P < O.OS), while those groups with no reported microperforation and with an optical zone diameter greater than 3.S mm had no significant cell loss.

2,835 ± 432

2,677 ± 528

- 5.6

Yes

Perimeter (JLm)

71.4 ± 5.4

74.3 ± 7.9

+3.9

Yes

Side lengths (JLm)

U.8 ± 0.9

12.3 ± 1.4

+ 4.0

Yes

CN of mean

0.319 ± 0.090

0.307 ± 0.055

- 4.0

No

Density (cells/mm 2 )

( -) 2,924 ± 461 ( +) 2,825 ± 471

62.5 ± 9.7

59.6 ± 6.8

-4.6

No

Perimeter (JLm)

0.872 ± 0.007

0.867 ± 0.016

- 0.6

No

Side lengths (JLm)

Densitv (cells/mm 2 )

area

% hexa-

gonality Shape (circle = 1)

Table 3. Microperforations.

% Increase

P< 0.05

364 603

- 1.6 -14.3

No Yes

(-) 70.6 ± 5.5 ( +) 72.0 ± 5.3

71.4 ± 4.6 78.2 ± 10.0

+ 1.1 + 7.9

No Yes

( -) 11.8 ± 0.8 (+) 11.4 ± 0.9

11.9 ± 13.0 ±

+ 1.0 + 7.7

No Yes

Preoperative Mean ± SD

Postoperative Mean ± SD

2,876 ± 2,421 ±

0.8 1.9

( - ) no microperforation; ( + ) microperforation

J CATARACT REFRACT SURG-VOL

13, MAY 1987

265

Table 4. Opticafzone diameter. Preoperative Mean ± SD

Postoperative Mean ± SD

p< % Increase 0.0.5

1* 2,994 ± 420 Densitv (cells;mm2) 2 2,421 ± 314

2,725 ± 594 2,421 ± 326

-9.0 0

Yes

1 2

70.1 ± 4.2 77.4 ± 4.2

73.5 ± 8.3 77.7 ± 4.7

+4.6 0

Yes No

1 2

11.6 ± 0.8 12.8 ± 0.7

12.2 ± 1..5 12.8 ± 0.9

+4.9 0

Yes No

0.866 ± 0.017 0.868 ± 0.010

-0.9 +0.6

Yes No

Perimeter

(pm) Side lengths (/-Lm)

Shape 1 0.874 ± 0.005 (circle = 1) 2 0.863 ± 0.019 *1 = 3.0 to 3.5 mm,

1\'0

2 = 3.75 to 4.5 mm

Microperforation, a direct incision of the Descemet"s membrane/endothelium complex, appears to cause an adverse effect on the corneal endothelium. Other studies report that histological observation has shown that clinically unnoticeable microperforations 3 and ridging of the endothelium 12 ,13 occurred during the surgery in human and animal eyes and that the incision was of varying depth even when the blade was set to the same depth each time. 3 ,12,13 The surgeon should therefore be aware of the possible occurrence of microperforations and their effect on the corneal endothelium. In addition to cell loss, a significant increase in pleomorphism (P < 0.05) was observed in the patient group having an optical zone diameter less than 3.5 mm. Pleomorphism is believed to be indicative of a compromised endothelium with a reduced functional reserve. 14 - 19 Most of the endothelial cells had an elongated shape, possibly due to microperforation near the central cornea that resulted in cell death and migration of the neighboring endothelial cells. Since the central cornea is thinner than the periphery, greater care must be taken to avoid microperforations in the central part of the cornea, especially when surgery is performed with a small optical zone diameter. Although previous studies 6 ,1l,14 have shown that many morphometric abnormalities seen soon after surgery disappear within six months, it is important to have long-term follow-up of these corneas, which had significant pleomorphism, since many changes may be persistent and may indicate a compromised endothelium. In this study, there was an increase in cell density in five eyes. No clinical explanation for this phenomenon is apparent. However, we believe it is possible that mitosis might have been triggered in the central cornea by the radial keratotomy. This explanation for the increase in cell density is still speculative and no evidence to support this theory exists. Our results indicate that among all the factors associated with radial keratotomy we studied, microperforation and a small central optical zone diameter are 266

J CATARACT

the greatest risk factors. Therefore, the surgeon's choice for the depth of incision and diameter of the optical zone to be used are important factors for safe, successful radial keratotomy. Although we did not include it, studying the endothelium between and beneath the incision where the most insult was done to see the direct effect on paracentral and peripheral areas of cornea is important. Unfortunately, such a study is difficult with present instrumentation. Computer-assisted morphometric analysis reveals more subtle changes in several parameters that might affect the corneal endothelium. Surgeons can also get more objective information from the data through statistical analysis done by computer. The length of time between surgery and postoperative analysis are also important. There may be a time-related factor to explain some statistically significant changes found in this study. For example, changes in perimeter andlor side length may be the early manisfestation of polymegathism. Therefore, long-term, detailed morphometric studies should be continued to give more thorough data to determine the safety of radial keratotomy and to determine other possible risk factors. REFERENCES

1. Hoffer KJ, Darin JJ, Pettit TH, Hofhauer JD, et al: Three years

2.

3. 4. 5. 6. 7. 8. 9. 10. 11.

12.

experience with radial keratotomy; the UCLA study. Ophthalmology 90:627-636, 1983 Nirankari VS, Katzen LE, Karesh JW, Richards RD, et al: Ongoing prospective clinical study of radial keratotomy. Ophthalmology 90:637-641, 1983 Dunn S, Jester JV, Arthur J, Smith RE: Endothelial cell loss following radial keratotomy in a primate model. Arch Ophthalmol 102:1666-1670, 1984 Jester JV, Steel D, Salz J, Miyashiro J, et al: Radial keratotomy in non-human primate eyes. Am] Ophthalmol 92:1.53-171, 1981 Rowsey JJ, Balyeat HD: Preliminary results and complications of radial keratotomy. Am] Ophtlwlmol 93:437-455, 1982 Salz JJ: Progressive endothelial cell loss f()llowing repeat radial keratotomy-a case report. Ophthalmic Sllrg 13:997-999, 1982 Yamaguchi T, Asbell PA, Ostrick M, SaRr A, et al: Endothelial damage in monkeys after radial keratotomy performed with a diamond blade. Arch Ophthalmol 102:765-769, 1984 MacRae SM, Matsuda M, Rich LF: The effect of radial keratotomy on the corneal endothelium. Am ] Ophthalmol 100:538-542, 1985 Laing RA, Sandstrom MM, Berrospi AR, Leibowitz HM: Changes in the corneal endothelium as a function of age. Exp Eye Res 22:587-594, 1976 Bourne WM: Morphologic and functional evaluation of the endothelium of transplanted human corneas. Trans Am Ophtlwlmol Soc 81:403-450, 1983 Murphy C, Alvarado J, Juster R, Maglio M: Prenatal and postnatal cellularity of the human corneal endothelium; a quantitative histologic study. Invest Ophthalmol Vis Sci 25:312-322, 1984 Yamaguchi T, Kaufman HE, Fukushima A, SaRr A, et al: Histologic and electron microscopic assessment of endothelial damage produced by anterior radial keratotomy in the monkey cornea. Am] Ophthalmol 92:313-327, 1981

REFRACT SURG-VOL 13, MAY 1987

13. Yamaguchi T, Polack FM, Valenti J, Kaufman HE: Endothelial damage after anterior radial keratotomy; an electron microscopic study of rabbit cornea. Arch OphthalmoI99:2151-2158, 1981 14. Matsuda M, Suda T, Manabe R: Quantitative analysis of endothelial mosaic pattern changes in anterior keratoconus. Am] Ophthalmol 98:43-49, 1984 15. Shultz RO, Matsuda M, Yee RW, Edelhauser HF, et al: Corneal endothelial changes in type I and type II diabetes mellitus. Am ] Ophthalmol 98:401-410, 1984 16. Matsuda M, Suda T, Manabe R: Serial alterations in endo-

J CATARACT REFRACT

thelial cell shape and pattern after intraocular surgery. Am] Ophthalmol 98:313-319, 1984 17. Bourne WM, Brubaker RF, O'Fallon WM: Use of air to decrease endothelial cell loss during intraocular le.lls-nnplantation. Arch OphthalmoI97:1473-1475, 1979 18. Shaw EL, Rao GN, Arthur EJ, Aquavella JV: The functional reserve of corneal endothelium. Ophthalmology 85:640-649, 1978 19. Rao GN, Shaw EL, Arthur EJ, Aquavella JV: Endothelial cell morphology and corneal deturgescence. Ann Ophthalmol 11:885-899, 1979

SURG- VOL 13, MAY 1987

267