Reliability of echographic examination for the study of optic nerve cupping

Purpose: To investigate the reliability and applicability of echography to detect optic nerve cupping smaller than or equal to 0.3 and greater than or equal to 0.9. Patients and Method: Fifty-three eyes from 30 patients, with transparent optic media, underwent contact B-scan echography to evaluate optic disc cupping. On biomicroscopy, twenty-six eyes presented small (Group 1) and 27 large (Group II) optic nerve cupping. Results: B-scan images of the optic disc were evaluated by two observers and ocular echography was considered reliable with a Kappa index of 0.77 (± 0.14). Sensitivity and specificity were 71.4 % (53.4 81.1) and 96.2 % (78.4 99.8), respectively, with a confidence interval of 95 % • The positive and negative predictive values found for the sample were 95.2 % (7 4.1 99.8) and 78.1 % (59.6 90.1), respectively. Conclusion: Echography is a reliable method to evaluate optic disc cupping when s; 0.3 and � 0.9. However, as a screening method it would be more important in the case of negative results.


INTRODUCTION
Primary open angle glaucoma is a chronic disease whose diagnostics is based, among others, on the triad intraocular pressure, optic nerve cupping and visual ' field. Sometimes the evaluation of optic nerve cupping is difficult or even impossible due to the opacity of the media. ln these cases, the echographic study of optic nerve head could be useful for the visual prognosis evaluation. Cohen et ai. 1 showed that, on echographic examination, cuppings greater than or equal to O. 7 usually appear as a concavity at the posterior pole of the eye, opposite the acoustically void area corresponding to the optic nerve 1 • ln addition there is a discontinuity at the vitreous-retinal interface corresponding to the area ofthe optic disc 1 , indicating presence of increase cupping.
The optic nerve is considered normal on echographic examination if there is a triangular image in the retrobulbar region which is echogenically void, without irregularity at the vitreous-retinal interface 1 -3 • Darnley-Fisch et ai. 4 showed that echography is a faulty method if used alone to classify the size of cupping and that great cuppings (:2: O. 7) are more easily diagnosed. ln addition shallow cuppings, although great, may be also difficult to detect.
The purpose of this study is to investiga te the possibility of evaluating optic nerve cuppings smaller than or equal to 0.3 and greater than or equal to 0.9 using echography, its reliability and applicability, in the attempt to distinguish between physiological and great cuppings.

PATIENTS AND METHODS
This study was carried out at the Department of Ophthal mology of the Santa Casa de Misericórdia de São Paulo, from December 1994 to December 1995.
Fifty-three eyes from 30 patients, with transparent optical media were selected, and submitted to biomicroscopy of the optical disc with a 90-diopter Volk lens by the sarne examiner. Of these, 26 were eyes without a diagnosis of glaucoma, with a ratio cupping/optic disc ::; 0.3 (Group 1). The remaining 27 were from patients with ocular hypertension or glaucoma with a ratio cupping/optic disc � 0.9 (Group II).
The age of the patients with small cupping (::; 0.3) ranged from 53 to 80 years, with a mean of 67 .5 ± 12.8 years. Group II presented ages ranging from 53 to 62 years, with a mean of 64.7 ± 8.5 years.
Ali eyes were submitted to B-scan echography in order to evaluate the cupping by the sarne examiner. The photographs of the echographic examination of the two groups were then randomly evaluated by two observers (A and B), and classified into positive or negative echographic images regarding glaucomatous cupping. The image was considered positive regarding advanced glaucomatous cupping when a depression limited to the papilla, anterior to the optic nerve or discontinuity at the vitreous-retinal interface was observed (Figures 1 and 2). It was considered negative regarding cupping when no irregularity anterior to the optic nerve was observed ( Figure 3).
The echographic examination was performed with the   Statistical analysis was based on reliability and vali dity of the examination. ln order to test reliability, the Kappa ( K) test was used. For the validity test, sensitivity and specificity of ocular echography in detecting advanced cupping was calculated in addition to calculation of positive and negative predictive values according to different prevalences, conside ring a confidence interval of 95 % 5• 6 . Table 1 shows the results of echographic examination of the eyes of Groups 1 and II according to the analysis by examiners A and B. It can be seen that with respect to the positive echographic image regarding cupping there was agreement in 19 eyes and in 28 eyes in the case of negative image. Therefore there was a 88.7% concordance. Kappa índex was 0.77 ± 0. 14; Z = 5.60 and p = 0.0000. ln order to evaluate the validity of the examination, the result of biomicroscopic evaluation by observer A was used: of the 37 eyes with cupping � 0.9, 20 presented positive and 7 negative echography. Of the 26 eyes with small cupping, 01 presented positive and 25 negative echography ( Table 2).
PPV and NPV were calculated on the basis of sensitivity and specificity of 74. 1 % and 96.2%, respectively, according to prevalence estimate for advanced glaucomatous cupping, as shown in Table 3 and Graph 1.

DISCUSSION
B-scan echography with a 1 O MHz probe which has a high frequency and presents sufficient penetrability to evaluate the optic nerve and its intraocular portion 2 was used in this study.
A Kappa índex ofü.77 ± 0. 14 and p = 0.0000 show that the test has a significance of 1 % and therefore echography was considered to be reliable in the evaluation of optic disc cupping. Thus echography may be used as an auxiliary method in the evaluation of optic nerve cupping in cases of patients with cataract associated with galucoma. This association is not uncommon, since both disorders are more frequent in the elderly (usually individuais above 50 years of age) and glaucoma, as well as cataract, is a silent and chronic disease.  K = 0.77 ± 0.14; Z = 5.60 and p = 0.0000 ln addition, echography is also useful in cases in which glaucoma diagnosis in patients with cataract can not be firmly established, as for instance in glaucoma cases with normal pressure where the diagnosis can only be established after cataract surgery.
Regarding validity we could observe that the sensitivity of ocular B-scan echography in the evaluation of advanced glaucomatous cupping was 74. 1 %, that is, echography was able to diagnose 74. 1 % of the advanced glaucomatous cuppings, an índex considered to be reasonable, since it presents 25.9% of false negatives. Specificity was 96.2%, that is, it detects 96.2% of the patients with small cuppings, with only 3.8% of false positives.
Thus echography is a method presenting a considerable percentage of false negatives and almost no false positives (3.8%) in the detection of papillary cupping. If we would use it as a predictive method for vision diagnosis of patients with opacity of visual media, it would be an acceptable method since it would not diagnose 25 .9% of the advanced cuppings.
The study by Darnley-Fish et ai 4 already suggested that echography alone is a faulty method for the classification of size of cupping.
The PPV found in the sample was 95.2%, with a confidence interval of 95%, that is, since the test is positive, the probability of it being actually positive is 95.2%. NPV was 78. 1 %, that is, when echography was negative, the chance of it being actually negative was 78.1 %.
Prevalence, the number of cases in the population, modifies the indicators (predictive values) ofthe performance of echography. Therefore, the values of the sample do not reflect reality, since prevalence of glaucoma in the population is approximately 1 %. ln order to estimate the validity of the echographic examination in the actual conditions, the positive and negative predictive values were calculated for each range of prevalence (Table 3 and Graph 1 ). We observed that, with increase in prevalence of advanced glaucomatous cupping, PPV increases and NPV decreases and that the prevalence rate at which predictive performance would be optimal is approximately 30%, where the 2 curves cross. At the prevalence of 1 %, PPV is very low and NPV quite high.
Thus, if we would use echography as a screening method to evaluate the optic disc in the population, it would be more  significant if the examination result would be negative, since the chance that it would be actually negative would be close to 100%. However, echography is an examination that can be applied in services such as that of the Santa Casa de Miseri córdia de São Paulo. Since prevalence of pathological cuppings is greater than in the population, because in is a reference service for glaucoma. Therefore, echography was shown to be a reliable exami nation for the diagnosis of the presence or not of great papillary cupping, becoming an ally in the presumption of vision prognosis of patients with ocular hypertension or glaucoma who, for some reason, have a difficult observation of their optic disc, or even as a method to help in the diagnosis of the disease.