Using unsupervised learning with variational bayesian mixture of factor analysis to identify patterns of glaucomatous visual field defects

PURPOSE: To determine whether an unsupervised machine learning classifier can identify patterns of visual field loss in standard visual fields consistent with typical patterns learned by decades of human experience. METHODS: Standard perimetry thresholds for 52 locations plus age from one eye of each of 156 patients with glaucomatous optic neuropathy (GON) and 189 eyes of healthy subjects were clustered with an unsupervised machine classifier, variational Bayesian mixture of factor analysis (vbMFA). RESULTS: The vbMFA formed five distinct clusters. Cluster 5 held 186 of 189 fields from normal eyes plus 46 from eyes with GON. These fields were then judged within normal limits by several traditional methods. Each of the other four clusters could be described by the pattern of loss found within it. Cluster 1 (71 GON + 3 normal optic discs) included early, localized defects. A purely diffuse component was rare. Cluster 2 (26 GON) exhibited primarily deep superior hemifield defects, and cluster 3 (10 GON) held deep inferior hemifield defects only or in combination with lesser superior field defects. Cluster 4 (6 GON) showed deep defects in both hemifields. In other words, visual fields within a given cluster had similar patterns of loss that differed from the predominant pattern found in other clusters. The classifier separated the data based solely on the patterns of loss within the fields, without being guided by the diagnosis, placing 98.4% of the healthy eyes within the same cluster and spreading 70.5% of the eyes with GON across the other four clusters, in good agreement with a glaucoma expert and pattern standard deviation. CONCLUSIONS: Without training-based diagnosis (unsupervised learning), the vbMFA identified four important patterns of field loss in eyes with GON in a manner consistent with years of clinical experience.     

一种评价青光眼视野损害的新方法——上下方Bjerrum区静态光阈值不对称

鉴于早期青光眼视野缺损具有上下方视野不对称特点,根据不对称视野计分可能发现早期青光眼视野损害。通过比较同一跟上下Bjerrum区相对应点群的平均光敏感度,分析54只已有早期或中期视野缺损的青光眼和105只正常眼的自动视野计资料。94.44%(51/54)的青光眼和92.38%(97/105)的正常眼被正确鉴定。此分析方法简单,对于检测青光眼视野缺损有较高的敏感性和特异性。     

Weighting according to location in computer-assisted glaucoma visual field analysis.

In recent years several aids for automated interpretation of visual field data have been suggested. We believed that incorporation of thorough knowledge of normal visual field variability would allow improvements in the performance of such aids since more attention would be paid to field results in areas with low physiological variability. Two visual field models for classification of fields in glaucoma based on comparisons of sensitivity values in the upper and lower hemifields and on analysis of test point clusters with diminished sensitivity were compared. Both models were constructed using logistic regression analysis in 101 normal eyes and 101 eyes with glaucoma. The first, more traditional model assumed Gaussian distributions of deviations from age-corrected normal thresholds and constant variability across the field (non-weighted model). The second model took into account empirically determined variability of pointwise threshold results and of cluster volumes in various visual field regions (weighted model). The two models were subsequently tested on an independent material of 163 normal eyes and 76 eyes with glaucoma. The weighted model gave significantly better classification of the fields in both materials. Accounting for physiological threshold variability can offer significant advantages in the construction of perimetric analysis aids for detection of glaucoma.     

Nerve fiber layer assessment with scanning laser polarimetry in glaucoma patients and glaucoma suspects.

PURPOSE: To investigate whether scanning laser polarimeter can differentiate glaucoma and suspected glaucoma patients from normals. METHODS: Polarimetric measurements were obtained using the nerve fiber analyzer (NFA)-I from 80 eyes of patients with glaucoma with mostly moderate glaucomatous optic nerve damage (37 eyes with primary open angle glaucoma, 21 with normal tension glaucoma, 17 with pseudoexfoliative glaucoma, 3 with angle closure glaucoma, and 2 with juvenile glaucoma), 53 eyes of patients suspected of glaucoma based on disc appearance, and from age-matched healthy volunteers as control groups. Ratios (superior/nasal, inferior/nasal, superior/inferior) were used for assessing nerve fiber layer (NFL) thickness. Student's t-test and linear regression analysis were used for statistical analysis. RESULTS: Both the glaucoma patients and glaucoma suspects had significantly lower NFL ratios (mean S/N 2.34 +/- 0.47, I/N 2.46 +/- 0.52, S/I 0.94 +/- 0.18) than the control groups (respectively 2.88 +/- 0.48, 2.88 +/- 0.48, 1.00 +/- 0.13) (p<0.05). There was an ample overlap between the patient groups and the normals. The superior and inferior NFL ratios in glaucoma patients gradually decreased as the mean defect in visual field increased (linear regression analysis, p<0.05). CONCLUSIONS: The NFL of glaucomatous eyes and eyes suspected of glaucoma based on disc appearance was significantly less thick than normals. NFA-I detects pathological abnormalities in some patients with glaucomatous optic nerve damage and normal visual fields as measured by conventional achromatic computerized perimetry. NFA-I, however, is unable to distinguish these patients from normals, at least using these parameters, because of the considerable overlap.     

Spatial analyses of glaucomatous visual fields; a comparison with traditional visual field indices.

Interpretation of numeric automated threshold visual field results is often difficult. A large amount of data is obtained for every single field tested. Various approaches to summarize this data have been suggested, most commonly the mean and standard deviation of departures from age-corrected normal threshold values. These visual field indices differ substantially from subjective field interpretation where spatial relationships are important. We have previously devised two methods for automated field interpretation which take spatial information into account--regional up-down comparisons and arcuate cluster analysis. We now studied the merits of using these new spatial methods and compared them to traditional visual field indices for discrimination between normal and glaucomatous field results. Central static 30 degree field results in 101 eyes of 101 normal subjects and 101 eyes of 101 patients with glaucoma were discriminated using logistic regression analysis. The best field classification was obtained with a spatial visual field model combining up-down differences and arcuate clusters. The advantages of the spatial model were confirmed in an independent material of 163 eyes of 163 normal subjects and 76 eyes of 76 patients with glaucoma where eyes with large field defects had been removed. In this material the spatial model gave 87% sensitivity and 83% specificity while the best non-spatial model gave 82% sensitivity and 80% specificity. Visual field interpretation in glaucoma may be significantly enhanced if detection is focused on circumscribed field loss rather than on averages of differential light sensitivities and similar indices which do not take spatial relationships into consideration.     

中心视野光阈值的短期波动

借助Humphrey视野分析仪研究了正常眼和早期青光眼光阈值的短期波动，不同年龄组的正常眼短期波动有统计学差异，性别和眼别对此没有影响，有视野缺损的早期青光眼，短期波动明显增加；无视野缺损的早期青光眼，短期波动无明显增加，讲座了正常眼和早期青光眼短期波动的原因，划分了正常和异常短期波动的范围。     

Multifocal objective perimetry in the detection of glaucomatous field loss.

PURPOSE: To test the ability of a new type of multifocal objective perimetry to identify glaucomatous visual field defects. METHODS: A multichannel visual evoked potential was recorded using the ObjectiVision Accumap perimeter. One hundred patients (age, 62.2 +/- 9.8 years, mean MD -6.5 +/- 4.17 dB) with open-angle glaucoma and confirmed glaucomatous visual field defects were tested and compared with the normal database of 100 normal subjects (age, 58.9 +/- 10.7 years). Both eyes were tested, but for determining sensitivity the eye with the lesser field defect was chosen if both qualified. The amplitude and intereye asymmetry coefficient for each zone of the field were calculated. A mean amplitude and multifocal objective perimetry severity index was calculated for each subject. RESULTS: In 95 of 100 (95%) patients with glaucoma Humphrey field defects were correlated with visual evoked potential amplitude reductions identifying a cluster of three or more abnormal zones. In two of five remaining patients with glaucoma the defect was detected on the intereye asymmetry analysis. Topographic location was well correlated with Humphrey fields. Mean amplitude was significantly reduced in 86 of the glaucoma cases (86%). The glaucoma severity index was abnormal in 93 glaucoma cases and showed a correlation with Humphrey MD (r = 0.67 right eyes, 0.69 left eyes). In 37 glaucoma cases with no scotoma by definition in the fellow eye, 22 (59.4%) had an abnormal multifocal objective perimetry, whereas only eight had some other aspect of their Humphrey visual field flagged as abnormal. CONCLUSIONS: Multifocal objective perimetry can assess the visual field and identify glaucomatous visual field defects. It may have the potential for identifying defects earlier than conventional perimetry.     

Using unsupervised learning with independent component analysis to identify patterns of glaucomatous visual field defects

PURPOSE: Clustering by unsupervised learning with machine learning classifiers was shown to segment clusters of patterns in standard automated perimetry (SAP) for glaucoma in previous publications. In this study, unsupervised learning by independent component analysis decomposed SAP field patterns into axes, and the information represented by these axes was evaluated. METHODS: SAP fields were used that were obtained with the Humphrey Visual Field Analyzer (Carl Zeiss Meditec, Dublin, CA) from 189 normal eyes and 156 eyes with glaucomatous optic neuropathy (GON) determined by masked review with stereoscopic optic disc photographs. The variational Bayesian independent component analysis mixture model (vB-ICA-mm) partitioned the SAP fields into the most informative number of clusters. Simultaneously, the model learned an optimal number of maximally independent axes for each cluster. RESULTS: The most informative number of clusters in the SAP set was two. vB-ICA-mm placed 68.6% of the eyes with GON in a cluster labeled G and 98.4% of the eyes with normal optic discs in a cluster labeled N. Cluster G optimally contained six axes. Post hoc analysis of patterns generated at -1 SD and +2 SD from the cluster G mean on the six axes revealed defects similar to those identified by experts as indicative of glaucoma. SAP fields associated with an axis showed increasing severity, as they were located farther in the positive direction from the cluster G mean. CONCLUSIONS: vB-ICA-mm represented the SAP fields with patterns that were meaningful for glaucoma experts. This process also captured severity in the patterns uncovered. These findings should validate vB-ICA-mm as a data-mining technique for new and unfamiliar complex tests.     

The Peritest automatic perimeter in screening for glaucomatous visual field defects

The ability of the Peritest to screen for glaucomatous defects of the visual field was studied by using this automatic perimeter to examine the visual fields of 69 eyes with chronic open-angle glaucoma and 46 normal eyes after manual perimetry had been performed. The Peritest correctly identified 62 (90%) of the glaucomatous eyes as having visual field abnormalities; the other 7 eyes (10%) had false-positive results. Of the 46 normal eyes, 36 (78%) were correctly identified as having normal visual fields; the other 10 eyes (22%) were incorrectly identified as having visual field abnormalities. Thus, the overall rate with which eyes were correctly classified as having normal or abnormal visual fields was 85%, and the rate with which fields were falsely identified as having glaucomatous abnormalities was 15%.     

变化角膜补偿器偏振光激光扫描仪对青光眼的临床观察

目的　评价使用变化角膜补偿器的偏振光激光扫描仪 (GDx -VCC)对青光眼患者的定量观察。方法　对青光眼患者 91人 165眼根据有无视野异常及程度分为 :青光眼视野正常组 87眼 ,早期青光眼 5 6眼及中晚期青光眼 2 2眼 ,正常人组 3 6人 3 6眼进行GDx -VCC检查。统计各组平均视神经纤维层厚度 (RNFL)、上方及下方神经纤维层厚度、平均视盘周神经纤维层厚度标准差及视神经纤维索引因素 (NFI)。对各组值进行多组比较统计处理。结果　正常人平均视神经纤维层厚度、上方及下方神经纤维层厚度与视野有异常的青光眼各组值方差分析P <0 0 0 1;与视野正常的青光眼组各对应区域的神经纤维层厚度比较P =0 0 0 9、 0 0 0 5及 0 0 64。结论　使用变化角膜补偿器的偏振光激光扫描仪测量视神经纤维层厚度对正常人与青光眼患者有区别能力 ,能更早于视野反映视神经纤维层异常。     

Luminance-modulated adaptation of global flash mfERG: fellow eye losses in asymmetric glaucoma

PURPOSE: To use the global flash multifocal electroretinogram (mfERG) in patients with asymmetric glaucoma to determine whether retinal function is affected in fellow eyes that have no glaucomatous visual field defects. METHODS: Forty normal subjects and 12 patients with asymmetric glaucoma were recruited for visual field and mfERG measurement. The mfERG was assessed by using a global-flash stimulation paradigm with four video frames: 103 scaled hexagonal elements followed by a dark frame, a global-flash frame, and a dark frame. The localized luminance difference was set at 96%, 65%, 49%, and 29% display contrast during the four different test conditions, respectively. The first-order kernel response was measured, and the "adaptive index" which has been used previously was calculated. RESULTS: In fellow eyes with normal visual fields, the amplitude of the induced component (IC) was significantly reduced, and the adaptive index was reduced by a factor of almost 10 (P < 0.0001), as it was in the glaucomatous eyes. Although the adaptive index in the better (fellow) eye of the patients with glaucoma was slightly higher than in the eyes with diagnosed glaucoma, these differences were not statistically significant. CONCLUSIONS: The significant reduction of the adaptive index in the better eyes in subjects with asymmetric glaucoma shows that the fast adaptive mechanism(s) were reduced in these eyes. This implies that eyes that have functionally normal visual acuity and visual fields have abnormal fast-adaptive mechanisms.     

Pattern of visual field defects in normal-tension and high-tension glaucoma

There are probably two major types of causative factors in open-angle glaucoma: pressure-dependent and pressure-independent. If clinical features such as the pattern of visual field defects differ between normal-tension and high-tension glaucoma, the differences may provide an insight for discriminating between the pressure-dependent and the pressure-independent damage in open-angle glaucoma. This article gives a brief review of the most recent studies including reports wherein progression or pattern of visual field defects in normal-tension and high-tension glaucoma or primary open-angle glaucoma are addressed. Further deterioration of the visual field in 5 years is expected in about 50% of eyes with normal-tension glaucoma in which intraocular pressure is one of the contributing factors. This figure may be greater than that seen in eyes with primary open-angle glaucoma where intraocular pressure is controlled with surgery in the middle teens. When eyes with normal-tension glaucoma and high-tension glaucoma or primary open-angle glaucoma were matched for extent of overall visual field loss, many studies noted a difference in the pattern of visual field defects between the two groups. Visual field defects in normal-tension glaucoma are relatively more localized and closer to fixation, especially in the nasal superior quadrant and may be more predominant in the lower hemifield. Results of other psychophysical tests also appear to support the above findings.     

Polarimetric measurement of retinal nerve fiber layer thickness in glaucoma diagnosis

PURPOSE: To evaluate reliability and diagnostic value of polarimetric measurements of the retinal nerve fiber layer (RNFL) thickness in the diagnosis of glaucoma. METHODS: The study included 81 eyes with perimetric glaucoma with glaucomatous changes of the optic disc and visual field defects; 52 eyes with preperimetric glaucoma with glaucomatous optic disc abnormalities and normal achromatic visual fields; and 70 normal eyes. For determination of reliability, four examiners repeated polarimetric measurements five times in ten normal subjects. RESULTS: The polarimetric variables were significantly correlated with increasing mean visual field defect and decreasing neuroretinal rim area. In correlation analyses with visual field defects, correlation coefficients were highest for the variable "superior/nasal ratio" and "the Number," a variable calculated by the neural network of the device. In correlations with neuroretinal rim area, correlation coefficients were highest for measurements of the inferior nerve fiber layer thickness. The preperimetric glaucoma group and the control group differed significantly in the variables "superior/nasal ratio" and "the Number" and, to a smaller degree, in the variables "superior/temporal ratio" and "superior/inferior ratio." The Number variable had a sensitivity of 82% and 58% at a predefined specificity of 80% in separating perimetric glaucoma patients and preperimetric glaucoma patients, respectively, from control subjects. Reproducibility of the polarimetric measurements ranged between 70% and 89%. CONCLUSION: Polarimetric measurements of the RNFL thickness can detect glaucomatous optic nerve damage in patients with visual field loss, and in some patients with preperimetric glaucomatous optic nerve damage. Considering the fast performance, easy handling, and low maintenance costs, RNFL polarimetry may be helpful in glaucoma diagnosis.     

正常人及青光眼患者中心30°视野光敏感度的定量分析

目的检测正常人及青光眼（有或无早期视野损害）患者中心视野光敏感度的差异,并提出早期视野损害的识别标准。方法对７５例（１０８只眼）正常人,年龄２０～７８岁;５３例（８２只眼）早期开角型或闭角型青光眼患者,年龄３０～７０岁;采用美国Ｈｕｍｐｈｒｅｙ６１０型视野分析仪,中心３０２程序,检测３０°视野内７６个点的光敏感度。结果正常组与青光眼组间性别与眼别光敏感度差异无显著性。正常眼组２０～４５岁与４６～７８岁组间的光敏感度差异有显著性。计算正常年龄组２０～４５岁和４６～７８岁两个年龄组中心视野每一检测点的８０％正常范围的下限值,距中心每隔１０°光敏感度降低２或３ｄＢ,下方视野光敏感度高于上方视野１或２ｄＢ,鼻侧与颞侧间的光敏感度差异无显著性。有视野缺损的青光眼患者,视野中多个检查点显示光敏感度下降４～８ｄＢ;无视野缺损患者与正常对照组比较,光敏感度降低１或２ｄＢ。结论根据正常眼与８０％正常范围的光敏感度下限值比较结果,提出青光眼视野损害分级标准：Ⅰ级：视野中有数个聚集点的光敏感度下降１～３ｄＢ;Ⅱ级：视野中有数个聚集点的光敏感度下降４～６ｄＢ;Ⅲ级：视野中有数个聚集点的光敏感度下降＞６ｄＢ。     

The visual evoked potential in glaucoma and ocular hypertension: effects of check size, field size, and stimulation rate

In order to determine the optimum stimulus conditions for the detection of optic nerve damage due to glaucoma and ocular hypertension, checkerboard pattern reversal visual evoked potentials (VEPs) were recorded from 20 glaucoma patients, 20 ocular hypertensive patients, and 20 age-matched normals. Two check sizes (12' and 48'), two field sizes (14 degrees and 28 degrees), and two alternation rates (1.9 and 7.5 alt/sec) were used. All subjects had visual acuities of 20/40 or better in each eye and equal pupils of 2 to 5 mm diameter. The largest number of VEP abnormalities were found with large checks (48') reversing at a fast rate (7.5 alt/sec). After correcting for the effects of age, visual acuity, and pupil size, 16 of 30 eyes with glaucomatous visual field defects had abnormally long VEP latencies under this condition (beyond the 99% confidence limit of the normal subjects). Nine of 40 ocular hypertensive eyes also had abnormally long latencies. Increased pattern VEP latency was significantly correlated with both the severity and location of visual field defects and the degree of cupping and pallor of the optic disc. VEP latency was not significantly related to intraocular pressure.     

Überschwellige periphere Stimulation bei präperimetrischen Glaukomen

BACKGROUND AND PURPOSE: Early glaucomatous visual field defects can occur outside the central 30 degrees , which is usually examined in perimetric tests used for glaucoma diagnosis and screening. This study aimed to evaluate the diagnostic value of peripheral suprathreshold stimulation in open angle glaucoma before the development of reproducible visual field damage in standard 30 degrees automatic white-on-white perimetry. METHODS: A total of 352 eyes of 352 patients (ages 35-69 years; visual acuity 0.8 or better) from the Erlanger Glaucoma Registry were included in this study. They were divided into two groups: normal eyes and preperimetric glaucoma. All patients underwent a standardized glaucoma examination including Octopus 500EZ static perimetry (G1 program, all three phases); 95 eyes of 95 patients also received a 135-point suprathreshold test pattern of the Humphrey Field Analyzer (model 750i) for detecting peripheral visual field defects. Sensitivity and specificity were calculated for any single test point in phase 3 of the G1 test pattern and the Humphrey 135-point pattern. A score was calculated, and cluster analysis was performed. RESULTS: In 33 of 176 (18.8%) eyes with preperimetric glaucoma, the score was 3 or higher in phase 3 of the G1 program (normal eyes: 19 of 196; 9.7%). For both examination modalities, the highest sensitivity was found in test locations in the superior nasal midperiphery, corresponding to neuroretinal rim loss predominantly in the inferotemporal sector in early glaucomatous optic disc atrophy. CONCLUSION: Positive test results using suprathreshold stimulation in the midperiphery can be found in patients with preperimetric glaucoma at a significantly higher frequency than in normal subjects. Longitudinal studies will show whether such tests can be useful for predicting perimetric manifestation of the disease.     

Detection of early glaucoma with optical coherence tomography (StratusOCT)

PURPOSE: To evaluate the performance of optical coherence tomography (StratusOCT) for discriminating eyes with early glaucoma from normal eyes. METHODS: Thirty eyes with established early glaucomatous visual field defects (EGVF group), 30 eyes with evidence of early glaucomatous optic neuropathy with normal standard achromatic perimetry [early glaucoma by disc (EGD)], and 33 age-matched normal eyes with good quality StratusOCT nerve fiber layer (NFL) images were enrolled. Average NFL thickness and NFL thickness at quadrants and sectors, areas under receiver operator characteristic curves, and sensitivities at 80% and 90% specificity were evaluated. RESULTS: The average (+/-SD) mean deviation in the EGVF group was -3.4 (+/-1.7) dB. Receiver operator characteristic curves showed areas under the curve (AUC) for NFL thickness in the superior quadrant (AUC=0.75+/-0.07) and in the inferior quadrant (AUC=0.94+/-0.03) to be the best StratusOCT parameters for discrimination of normal controls from EGD and EGVF eyes, respectively. The best parameter for detection of EGD eyes at 80% and 90% specificities was NFL thickness at superior quadrant (51% and 36% sensitivities, respectively). The best parameter for detection of EGVF eyes at 80% and 90% specificities was NFL thickness in the inferior quadrant (90% and 87% sensitivities, respectively). CONCLUSIONS: Optical coherence tomography (StratusOCT) showed good sensitivity and specificity in a group of glaucoma patients with early visual field loss. In patients with normal visual fields in whom the optic disc appeared glaucomatous to glaucoma specialists, half were confirmed to have StratusOCT findings consistent with damage from glaucoma.     

Association of retinal vessel caliber and visual field defects in glaucoma

PURPOSE: This study evaluates the asymmetry of peripapillary retinal vessel caliber between inferior and superior hemispheres in eyes with visual field defects predominantly in one hemifield. DESIGN: Observational case series. METHODS: In a retrospective study, 64 eyes of 64 patients with primary open-angle glaucoma who had a marked difference in visual field defects between hemifields and who had no history of diabetes, trauma, or vascular occlusive disease were studied. The diameters of the superior and inferotemporal vessels were measured at the optic disk border with calipers on an enlarged image. RESULTS: In 64 eyes, the average ratio of the superior temporal artery diameter to inferotemporal artery diameter was significantly greater in the eyes with predominantly superior visual field defects as compared with those with inferior defects (1.10 +/- 0.22 vs. 0.92 +/- 0.19, respectively, P =.002, two-tailed t test). This indicates that the arteriole corresponding to the hemifield with the greater visual field defect was narrower than the arteriole in the other hemifield. This relationship was confirmed using chi(2) analysis (P =.002) comparing the proportions of eyes with ratios greater or less than normal vessel caliber ratios (normal ratio = 0.95 from data reported by Jonas and associates to the location of the dominant field defect. No statistically significant relationship was detected between retinal vein diameter and localized visual field defects, as determined by both the unpaired t test and chi(2) analysis. CONCLUSION: In eyes with primary open-angle glaucoma, this study demonstrates a strong association between decreased peripapillary arteriole diameter and visual field defects in the corresponding hemifield. This reflects either an ischemic basis for glaucomatous damage or vascular constriction when there are fewer axons to nourish.     

Preperimetric glaucoma diagnosis by confocal scanning laser tomography of the optic disc

AIM: To evaluate the ability of confocal scanning laser tomography of the optic nerve head to detect glaucomatous optic nerve damage in ocular hypertensive eyes without visual field defects. METHODS: The study included 50 normal subjects, 61 glaucoma patients with glaucomatous changes in the optic disc and visual field, and 102 "preperimetric" patients with increased intraocular pressure, normal visual fields, and glaucomatous appearance of the optic disc as evaluated on colour stereo optic disc photographs. For all individuals, confocal scanning laser tomographs of the optic nerve head were taken using the Heidelberg retina tomograph (HRT; software 2.01). RESULTS: Almost all investigated HRT variables varied significantly (p < 0.05) between the normal eyes and preperimetric glaucoma eyes with pronounced overlap between the two study groups. Corresponding to the overlap, sensitivity and specificity values were relatively low when HRT variables were taken to differentiate between normal and preperimetric glaucoma eyes. At a given specificity of 95% highest sensitivities were found for the variables "rim area in the superior disc sector" (24.8%), "nerve fibre layer thickness in the inferior disc sector" (26.5%), and "rim volume in the superior disc sector" (25.5%). A multivariate approach increased sensitivity to 42.2% at a given specificity of 95%. For the glaucoma group highest sensitivity values were reached by rim volume in the superior disc sector (73.8%) and rim area (72.1%); the multivariate approach reached 83.6%. CONCLUSIONS: Owing to pronounced overlapping between the groups, confocal scanning laser tomography of the optic nerve head has relatively low diagnostic power to differentiate between normal eyes and preperimetric glaucoma eyes. One of the reasons may be the biological interindividual variability of quantitative optic disc variables.