Spectral domain optical coherence tomography (SD-OCT) is a useful way of imaging the surface of the optic disc and peripapillary retina in a variety of optic neuropathies. For example, the average thickness of the retinal nerve fiber can be used to assess the degree of optic disc edema (1–5). The SD-OCT can also image subsurface structures such as the peripapillary retinal pigment epithelium–Bruch's membrane (ppRPE) layer and choroid. Histomorphometric studies and in vivo imaging in animals and humans have shown that a chronic increase in the intraocular pressure can posteriorly displace the peripapillary sclera and lamina cribosa (6–13). Conversely, increased intracranial pressure can deform the sclera and ppRPE-layer anteriorly toward the vitreous (14,15). In extending this line of study, we analyzed the shape of the ppRPE-layer in patients with presumed meningiomas of the optical nerve sheath (pONSM) to characterize the shape deformation and identify tumor-related factors that might affect this deformation.
We used geometric morphometrics (GM) to quantitatively analyze the shape of the ppRPE-layer imaged on the SD-OCT raster. GM defines shape as the geometric property of a form that remains after filtering out variations in position, scale, and orientation. It is a well-established technique developed to quantify and statistically analyze variation in the shape of biological forms and their covariation with other variables using conventional multivariate statistical methods (16–19). It has been extensively used in the fields of biology, anthropology, and paleontology.
The analytic techniques used in GM are complex but accessible using a variety of software applications that are widely available. For those readers who wish to learn more, Sanfilippo et al (18,20) provide an overview of shape analysis that includes a discussion of GM and its application in assessing the shape of the optic cup in glaucoma. More detailed information can be found in the monograph by Zelditch et al (16) and a Web site published elsewhere (15) and described below. This study was approved by the SUNY Stony Brook Committee on Research Involving Human Subjects.
We analyzed 11 eyes from 10 patients with a diagnosis of pONSM. The presumptive diagnosis was based on both clinical and magnetic resonance imaging (MRI) findings, which consisted of unilateral (9 cases) or bilateral (1 case) slowly progressive vision loss, signs of optic neuropathy (with optic disc edema or optic atrophy), and enlargement and enhancement of the prechiasmal optic nerve sheath on magnetic resonance imaging. None of the cases were confirmed pathologically. The demographics and key clinical findings are summarized in Figure 1. We compared these patients to 30 normal eyes. Among normals, we excluded any patients with abnormal acuity, color vision, a relative afferent pupillary defect, elevated intraocular pressure; abnormal visual fields, ophthalmoscopic findings or SD-OCT evidence of an optic neuropathy, optic atrophy, glaucoma, or congenital disc anomalies (e.g., drusen, hypoplasia, oblique insertion, tilting, high myopia, staphylomas, or otherwise dysplastic).
SD-OCT were acquired with a Cirrus SD-OCT (Carl Zeiss Meditec, Inc, Dublin, CA). Sharply focused uniformly illuminated images centered over the optic nerve head were obtained using 2 standard protocols: (1) optic disc cube 200 × 200 and (2) a 5-line horizontal high definition raster (9 mm long, 0.25 mm intervals). The raster scan was positioned through the central portion of the optic disc with signal strength of ≥7. Images were saved in the highest quality .jpeg format. To more accurately assess the image (shape), we converted the display aspect ratio from 3:2 (750 × 500 pixels) (Fig. 2A, C) to a true aspect ratio of 9:2 (750 × 167 pixels) (Fig. 2B, D), which has a uniform spatial scale along both vertical and horizontal dimensions.
Digitizing Structural Semi-landmarks
Imaging software (Photoshop; Adobe Systems, San Jose, CA) was used to superimpose a transparent line grid spanning 2500 μm on either side of the neural canal opening (NCO). The grid was positioned parallel to the flattest portion of the ppRPE-layer on both sides of the NCO, with a starting reference point positioned at the innermost termination of the ppRPE-layer. The grid was used to position 10 semi-landmarks (slightly less than 278 μm apart) along the posterior border of the ppRPE-layer on the temporal and nasal side of the NCO. Points 1 to 10 were placed temporally and 11 to 20 nasally. Left eye images were flipped horizontally so that all shape figures depict the nasal RPE on the right side of the image; the temporal RPE is located on the left side (Fig. 2B).
Geometric Morphometric Analysis
We used GM tps software (19) to analyze the digitized semi-landmarks for each subject including tpsUtil, tpsDig2, tpsRegr, and tpsPLS. The software performs a generalized least squares Procrustes superimposition which normalizes the shapes by filtering out differences in location, size, and rotation; a thin-plate spline that uses an algorithm to depict shape differences as a smooth deformation of one shape into another and generates shape variables that can be analyzed with conventional multivariate statistical methods; and principal component analysis (PCA) to identify and display the main sources of shape variance between subjects.
Generalized Least Squares Procrustes Superimposition
Generalized least squares Procrustes superimposition is the iterative process of estimating a mean shape and then superimposing all of the objects onto this mean shape. This is performed in 3 steps (16). First, the set of points for each subject is adjusted so that their centroid (mean of all of the x coordinates, mean of all the y coordinates) is translated to the origin. This is done by subtracting centroid coordinates from the coordinates of each landmark. Second, each configuration is scaled by dividing by the centroid size (the square root of the summed squared distances of each landmark from the centroid of its landmark configuration). Third, rotational differences are removed by iteratively minimizing the summed squared distances between corresponding landmarks.
The thin-plate spline has 2 important functions. First, it is used to depict shape differences as a smooth deformation of one shape into another using an algorithm that interpolates potential changes between landmarks of a reference shape (the mean shape) and the shape it is being compared with. These same shape differences can also be visualized using vectors at each landmark showing the magnitude and direction of the differences at each landmark.
The thin-plate spline is also used to define a set of shape variables, partial warps, that capture the shape differences among the objects being compared. The partial warp scores provide data matrices that can be analyzed with conventional multivariate statistical methods.
Principal Component Analysis
Because shape variation is multidimensional, PCA was used to express as much of the variation as possible in just a few dimensions that are linear combinations of the partial warps. This allows one to identify and display most of the variation in shape between subjects. The relative contribution of each dimension is proportional to its variance. The tpsRelW software was used for these computations.
A test statistic adapted to assessing shape differences was proposed by Goodall (16,21). It compares sums of squared Procrustes differences between and within the samples being compared and expresses it as an F-ratio. Valid statistical tests can be made by comparing the observed F value to an empirical distribution based on a large number (10,000 in this study) of random permutations of the assignments of individuals to the groups being compared. The proportion of Goodall's F statistics from permuted data sets that are equal to or larger than the observed Goodall's statistic is interpreted as the probability value for the test. Goodall's F test only considers the total amount of shape variation and does not display the nature of the differences.
For traditional morphometric measurements, for example, disc elevation, average RNFL, and NCO diameters, we used Student t test and analysis of variance where appropriate. We analyzed how shape variables might covary with a variety of factors including age, refractive error, RNFL thickness, elevation of the optic disc, overall length of the tumor (size), and its proximity to the globe. Correlation between shape variables and RNFL, disc height, age, and tumor size and proximity to the globe were assessed using a Partial least squares analysis (using J Rohlf's tpsPLS software). MRI was used to measure the (1) length of the tumor and (2) proximity to the globe (measured as the distal leading edge of tumor to the posterior wall of the globe in millimeters). We also used an algorithm that combined the length of the tumor with a weighted score that was based on its proximity to the globe. For example, tumors <2 mm from the globe received 20 points in addition to the length of the tumor, <5 mm received 10 additional points, <10 mm received 5 additional points, and so on.
Ten women with presumed ONSM between the ages of 37–66 years (mean, 48.7 years) were followed over an average of 58 months (range, 32–141 months). Nine of the patients had a unilateral lesion, 1 patient had bilateral involvement, with 7 right eyes and 3 left eyes (Fig. 1). At presentation, 10 of the eyes had optic disc edema, 1 had optic atrophy. Four of the eyes went on to develop optic atrophy. Four went on to develop retinochoroidal venous collaterals (optociliary shunt vessels). The pONSM of 2 patients showed calcification on computed tomography (CT). The MRI in all cases showed thickening or enlargement of the optic nerve with enhancement, usually along the optic nerve sheath. The greatest deformation was associated with meningiomas abutting the globe, although shape deformation was also observed among patients with tumors remote from the globe. There was one exception with a normal shape in a patient where the tumor was distant from the globe located at the orbital apex and optic canal.
OCT data at presentation were not available in 4 of the cases. However, after a mean follow-up of 58 months, data from the most recent visit showed that the mean RNFL thickness was distributed bimodally. Those with chronic disc edema averaged 156 μm (±44 μm), and those with optic atrophy averaged 67.4 μm (±6.5 μm). The unaffected eye averaged 92.6 μm (±8.9 μm) compared with normal subjects with a mean of 92.2 μm (±9.6 μm).
A generalized Procrustes superimposition of 20 semi-landmarks for 41 patients (30 normal optic discs and 11 pONSM eyes) is shown in the scatter plot (Fig. 3A). The consensus or mean shapes of the Procrustes transformed semi-landmarks for each group are shown in Figure 3B demonstrating the differences in the shape of the ppRPE-layer in normals vs pONSM. Note that these shapes were obtained from images with an aspect ratio 9:2. However, the horizontal raster images shown in Figure 1 are displayed as they are generated by the Cirrus SD-OCT with a 3:2 aspect ratio.
Our results demonstrate that the mean ppRPE-layer in normals has a V-shaped configuration sloped outwardly (away from the vitreous) as it approaches its central margin at the NCO. In contrast, the consensus (mean) shape for subjects with pONSM have an inverted-U shaped ppRPE-layer that is anteriorly displaced toward the vitreous. There is slight nasal skew in the inverted-U shape compared with the relatively symmetrical V-shape in normals. The difference between the 2 groups was highly significant at P = 0.001 (permutation statistics). Two examples of a normal V-shaped RPE layer and an inverted U-shape RPE layer are illustrated both in the uncorrected (3:2 aspect ratio) and corrected (9:2 aspect ratio) forms in Figure 2. Figure 1 displays each of the uncorrected raster scans (3:2 aspect ratio) from both the affected and the contralateral unaffected eye for comparison. The deformations vary from an obvious inverted U-shape indentation of the globe (e.g., 1–5) to mild relative flattening of the globe compared with the normal side (e.g., 6–10). There is no apparent deformation in patient 11 who had the smallest tumor furthest from the globe.
A variance—covariance matrix of the shape variables derived from both groups of patients was used to perform a PCA (using tpsRelW software (19)). The first 2 principal components together account for 81% of the variance; 62% from PC1 alone. Figure 4 shows the distribution of principal component scores from each subject along the first 2 PC axes. The shape implied along the PC1 axis depicts a continuum of shapes that ranges from an inverted-U (on the negative side of the abscissa) to a V-shape (on the positive side of the abscissa). The PC2 along the ordinate describes a shape change that goes from NCO contraction (up-in) on the negative side to NCO expansion (down-out) on the positive side (Fig. 4). The intersection of PC1 and PC2 represents the consensus shape of the ppRPE-layer semi-landmarks for all specimens. The PC plot shows 2 distinct clusters (pONSM and normals) that segregate along PC 1. With one notable exception (patient 11), nearly all subjects with pONSM exhibit some degree of the inverted-U shape deformation to relative flattening compared with the normals. There was no obvious segregation of the 2 groups along PC2. The probability that this difference in shape between pONSM and normal subjects could have arisen by chance was 1 in 10,000 permuted data or P = 0.0001.
There was no statistically significant correlation between shape variables and the age of the subject (r = 0.67), disc height (r = 0.67), mean RNFL (0.51), length of the tumor (r = 0.61), proximity of the tumor to the globe (0.66), or length + proximity (r = 0.71). An algorithm that assigned points based on proximity to the globe plus tumor length (in millimeters) did show a moderate correlation (r = 0.75, P = 0.037) with shape.
This study shows that shape of the ppRPE-layer in patients with pONSM differs from normals, and this difference was statistically significant. It is characterized by an inward deformation of the ppRPE-layer toward the vitreous that presumably reflects the shape of the underlying load bearing peripapillary sclera (15) and correlates with both the size and proximity of the tumor to the globe. Larger tumors contiguous to the sclera are associated with greater amounts of deformation. The shape of the ppRPE-layer in 1 patient with a small tumor, most distant from the globe, showed no deformation and was otherwise normal.
The peripapillary shape deformation demonstrated on the SD-OCT can, at times, be seen on an MRI, CT, or B-scan echography as flattening of the globe in some cases associated with acquired hyperopia and choroidal folds (22–28). The SD-OCT however is more sensitive and provides greater detail than the MRI and CT. The application of GM to these images can statistically distinguish shape differences from normals even when differences appear normal by gross inspection (15).
Pathologically, ONSM may infiltrate the optic nerve head and the adjacent choroid (29–31). Shape deformation in such instances probably is due to direct compression of the optic nerve with axonal flow blockade and proximal swelling, in much the same way that an intraconal orbital tumor will compress the globe and induce choroidal folds and hyperopia. However, the mass effect on the globe with most pONSM, even those contiguous to the globe, is mild by comparison. Moreover, in 5 of 11 eyes (patients 4, 7, 8, 9, and 10) the deformation was evident even when the tumor did not appear (based on MRI) to directly compress the globe. Taban et al (28) have described a patient with remote parasellar meningioma involving the optic canal and orbital apex, who exhibited choroidal folds and flattening of the globe by B-scan ultrasound.
The mechanism by which an ONSM can deform the globe is unknown. It is noteworthy that the shape deformation in pONSM is indistinguishable from the shape of the ppRPE-layer in patients with intracranial hypertension and papilledema. Eight eyes of our patient cohort had optic disc edema at presentation. However, disc edema alone does not explain these shape deformations because we have previously shown that there is no shape deformation in patients with disc edema due to anterior ischemic optic neuropathy (15). In an animal model, Morgan et al (7) have shown that displacements of the optic disc surface is probably the result of translaminar pressure differences between the subarachnoid and intraocular fluid compartments. This is the likely explanation of the subsurface deformations at the level of the ppRPE-layer in patients with papilledema (15,32).
There are a number of similarities in the clinical signs of both chronic papilledema and pONSM. Patients with papilledema initially present with bilateral or asymmetric disc edema, normal acuity and color vision, and enlargement of the blind spot on visual fields. Likewise, patients with meningiomas of the optic nerve sheath present early on with isolated unilateral optic disc edema, normal acuity, and enlargement of the blind spot, and then slowly progress with loss of peripheral visual field, and eventual optic atrophy and shunt vessels on the disc surface (33–36). In the absence of a large intracranial tumor component, patients with isolated intraorbital pONSM do not have intracranial hypertension. The clinical similarities in their effects on the nerve head between chronic papilledema and pONSM may be a consequence of the underlying biomechanics common to both of these entities.
Some investigators (37–43) have suggested that the cerebrospinal fluid (CSF) in the perioptic subarachnoid space (poSAS) may not freely communicate with the intracranial subarachnoid space based on measurable concentration gradients of CSF proteins and contrast along the optic nerve sheath. This concentration gradient has similarly been demonstrated in a patient with an ONSM (38). It is possible that the shape differences in pONSM are the result of localized increased pressure within the poSAS because the tumor effectively functions as a one-way hydraulic valve.
The biomechanical effects of the meningioma on the optic nerve sheath may also play a role. Strain is a measure of local deformation of a material or tissue induced by stress (applied force per cross-sectional area). The amount of strain is also influenced by the material properties of the tissue; that is, its ability to resist deformation (stiffness or compliance). Normally, the intraorbital segment of the optic nerve is slack and compliant; the ends fixed within the neural canal and the annulus. Eye movements stress and strain the nerve and the peripapillary eye wall. Thickening, calcification, and increased pressure in the poSAS in patients with meningiomas presumably stiffen the optic nerve sheath. This decrease in compliance may also contribute, if not cause, the deformation observed on OCT.
The functional significance of this deformation is unknown. Burgoyne et al (13) have suggested that distortions of the laminar pores and surrounding sclera, in glaucoma at least, may impede axoplasmic transport and blood flow that ultimately affects axonal and glial function within the optic nerve head. This may play an important role in papilledema and pONSM with disc edema because the magnitudes of shape differences are substantially greater than they are in glaucoma (15).
Regardless of the underlying mechanism, recognition of peripapillary deformation of the ppRPE-layer on the SD-OCT may be helpful in selected cases. Examples include distinguishing pseudopapilledema from true disc edema, detection of infiltration optic nerve sheath tumors in patients with disc edema or retinal vein occlusions, and assisting to identify shunt failure particularly in patients with optic atrophy.
1. Karam EZ, Hedges TR. Optical coherence tomography of the retinal nerve fibre layer in mild papilloedema and pseudopapilloedema. Br J Ophthalmol. 2005;89:294–298.
2. Menke MN, Feke GT, Trempe CL. OCT measurements in patients with optic disc edema. Invest Ophthalmol Vis Sci. 2005;46:3807–3811.
3. Ophir A, Karatas M, Ramirez JA, Inzelberg R. OCT and chronic papilledema. Ophthalmology. 2005;112:2238.
4. Martinez MR, Ophir A. Optical coherence tomography as an adjunctive tool for diagnosing papilledema in young patients. J Pediatr Ophthalmol Strabismus. 2011;48:174–181.
5. Scott CJ, Kardon RH, Lee AG, Frisen L, Wall M. Diagnosis and grading of papilledema in patients with raised intracranial pressure using optical coherence tomography vs clinical expert assessment using a clinical staging scale. Arch Ophthalmol. 2010;128:705–711.
6. Morgan WH, Chauhan BC, Yu DY, Cringle SJ, Alder VA, House PH. Optic disc movement with variations in intraocular and cerebrospinal fluid pressure. Invest Ophthalmol Vis Sci. 2002;43:3236–3242.
7. Morgan WH, Yu DY, Cooper RL, Alder VA, Cringle SJ, Constable IJ. The influence of cerebrospinal fluid pressure on the lamina cribrosa tissue pressure gradient. Invest Ophthalmol Vis Sci. 1995;36:1163–1172.
8. Yan DB, Coloma FM, Metheetrairut A, Trope GE, Heathcote JG, Ethier CR. Deformation of the lamina cribrosa by elevated intraocular pressure. Br J Ophthalmol. 1994;78:643–648.
9. Levy NS, Crapps EE. Displacement of optic nerve head in response to short-term intraocular pressure elevation in human eyes. Arch Ophthalmol. 1984;102:782–786.
10. Levy NS, Crapps EE, Bonney RC. Displacement of the optic nerve head. Response to acute intraocular pressure elevation in primate eyes. Arch Ophthalmol. 1981;99:2166–2174.
11. Bellezza AJ, Rintalan CJ, Thompson HW, Downs JC, Hart RT, Burgoyne CF. Deformation of the lamina cribrosa and anterior scleral canal wall in early experimental glaucoma. Invest Ophthalmol Vis Sci. 2003;44:623–637.
12. Coleman AL, Quigley HA, Vitale S, Dunkelberger G. Displacement of the optic nerve head by acute changes in intraocular pressure in monkey eyes. Ophthalmology. 1991;98:35–40.
13. Burgoyne CF, Downs JC, Bellezza AJ, Suh JK, Hart RT. The optic nerve head as a biomechanical structure: a new paradigm for understanding the role of IOP-related stress and strain in the pathophysiology of glaucomatous optic nerve head damage. Prog Retin Eye Res. 2005;24:39–73.
14. Kupersmith MJ, Sibony P, Mandel G, Durbin M, Kardon RH. Optical coherence tomography of the swollen optic nerve head: deformation of the peripapillary retinal pigment epithelium layer in papilledema. Invest Ophthalmol Vis Sci. 2011;52:6558–6564.
15. Sibony P, Kupersmith MJ, Rohlf FJ. Shape analysis of the peripapillary RPE layer in papilledema and ischemic optic neuropathy. Invest Ophthalmol Vis Sci. 2011;52:7987–7995.
16. Zelditch ML, Swiderski DL, Sheets HD, Fink WL. Geometric Morphometrics for Biologists: A Primer. London, United Kingdom: Academic Press; 2004.
17. Adams DCR, Rohlf FJ, Slice DE. Geometric Morphometrics: ten years of progress following the “revolution”. Ital J Zool. 2004;71:5–16.
18. Sanfilippo PG, Cardini A, Hewitt AW, Crowston JG, Mackey DA. Optic disc morphology—rethinking shape. Prog Retin Eye Res. 2009;28:227–248.
20. Sanfilippo PG, Cardini A, Sigal IA, Ruddle JB, Chua BE, Hewitt AW, et al.. A geometric morphometric assessment of the optic cup in glaucoma. Exp Eye Res. 2010;91:405–414.
21. Goodall C. Procrustes methods in the statistical analysis of shape. J R Stat Soc. 1991;53:285–339.
22. Brodsky MC, Glasier CM. Magnetic resonance visualization of the swollen optic disc in papilledema. J Neuroophthalmol. 1995;15:122–124.
23. Gass A, Barker GJ, Riordan-Eva P, MacManus D, Sanders M, Tofts PS, McDonald WI, Moseley IF, Miller DH. MRI of the optic nerve in benign intracranial hypertension. Neuroradiology. 1996;38:769–773.
24. Sharma M, Volpe NJ, Patel T, Kimmel A. Intracranial hypertension associated with acquired hyperopia and choroidal folds. Retina. 1999;19:260–262.
25. Dailey RA, Mills RP, Stimac GK, Shults WT, Kalina RE. The natural history and CT appearance of acquired hyperopia with choroidal folds. Ophthalmology. 1986;93:1336–1342.
26. Stimac GK, Mills RP, Dailey RA, Shults WT, Kalina RE. CT of acquired hyperopia with choroidal folds. AJNR Am J Neuroradiol. 1987;8:1107–1111.
27. Jacobson DM. Intracranial hypertension and the syndrome of acquired hyperopia with choroidal folds. J Neuroophthalmol. 1995;15:178–185.
28. Taban M, Kosmorsky GS, Singh AD, Sears JE. Choroidal folds secondary to parasellar meningioma. Eye (Lond). 2007;21:147–150.
29. Henderson JW. Orbital Tumors. New York, NY: Brian C Decker; 1980:472–496.
30. Henderson JW, Campbell RJ. Primary intraorbital meningioma with intraocular extension. Mayo Clin Proc. 1977;52:504–508.
31. Martin VA, Schofield PB. Meningioma invading the optic nerve. Br J Ophthalmol. 1957;41:161–166.
32. Kupersmith MJ, Sibony P, Mandel G, Durbin M, Kardon R. Optical coherence tomography of the swollen optic nerve head: deformation of the peripapillary RPE layer in papilledema. Invest Ophthalmol Vis Sci. 2011;52:6558–6564.
33. Sibony PA, Krauss HR, Kennerdell JS, Maroon JC, Slamovits TL. Optic nerve sheath meningiomas. Clinical manifestations. Ophthalmology. 1984;91:1313–1326.
34. Saeed P, Rootman J, Nugent RA, White VA, Mackenzie IR, Koornneef L. Optic nerve sheath meningiomas. Ophthalmology. 2003;110:2019–2030.
35. Turbin RE, Pokorny K. Diagnosis and treatment of orbital optic nerve sheath meningioma. Cancer Control. 2004;11:334–341.
36. Wilson WB. Meningiomas of the anterior visual system. Surv Ophthalmol. 1981;26:109–127.
37. Jaggi GP, Harlev M, Ziegler U, Dotan S, Miller NR, Killer HE. Cerebrospinal fluid segregation optic neuropathy: an experimental model and a hypothesis. Br J Ophthalmol. 2010;94:1088–1093.
38. Jaggi GP, Mironov A, Huber AR, Killer HE. Optic nerve compartment syndrome in a patient with optic nerve sheath meningioma. Eur J Ophthalmol. 2007;17:454–458.
39. Killer HE, Flammer J. Unilateral papilledema caused by a fronto-temporo-parietal arachnoid cyst. Am J Ophthalmol. 2001;132:589–591.
40. Killer HE, Jaggi GP, Flammer J, Miller NR, Huber AR, Mironov A. Cerebrospinal fluid dynamics between the intracranial and the subarachnoid space of the optic nerve. Is it always bidirectional? Brain. 2007;130:514–520.
41. Killer HE, Jaggi GP, Miller NR. Optic nerve compartment syndrome. Acta Ophthalmol. 2011;89:e472; author reply e72–e73.
42. Killer HE, Jaggi GP, Miller NR, Huber AR, Landolt H, Mironov A, Meyer P, Remonda L. Cerebrospinal fluid dynamics between the basal cisterns and the subarachnoid space of the optic nerve in patients with papilloedema. Br J Ophthalmol. 2011;95:822–827.
43. Killer HE, Laeng HR, Flammer J, Groscurth P. Architecture of arachnoid trabeculae, pillars, and septa in the subarachnoid space of the human optic nerve: anatomy and clinical considerations. Br J Ophthalmol. 2003;87:777–781.