When evaluating expert systems to be used in clinical perimetry, various aspects of their performance as compared with that of human interpreters must be considered. In this investigation, the results produced by the new Octosmart diagnostic program have been compared with the performance of three interpreters with various amounts of experience in visual field analysis. The evaluations were based on 27 visual fields with glaucomatous damage, which had been examined with the Octopus program G1. It is shown that in borderline cases (i.e., neither clearly normal nor clearly pathological) where strict statistical criteria must be employed in order to distinguish between possible pathology and artifacts, the "personal styles" of human interpreters, more than standardized decision criteria, implicitly guide the decision process, resulting in unpredictable, non-standardized interindividual differences. A standardized expert system, based on constant, explicit, and logical criteria is therefore considered to be superior to unaided human interpretation. It is pointed out that the influence of the implicit decision criteria of human interpreters must be controlled carefully if expert systems are to be evaluated with reference to human interpreters.
Two adaptive perimetric strategies for the search and the analysis of scotomata have been developed and tested over the normal and pathologically disturbed blind spot area. Adaptive procedures automatically concentrate their search effort upon areas of pathological disturbance and avoid time loss used for declaring normal areas as such. Provided a first scan of low spatial resolution, detects only one edge of the blind spot, spatial resolution increases and analyses the blind spot, returning to low resolution as soon as the border of the blind spot is crossed and normal sensitivity is again attained. Two adaptive strategies are described. One reduces examination duration to 1/2, the other to about I/3 when compared to a similar non adaptive high resolution strategy. Only in the latter this time reduction has to be paid by some loss of information. The accuracy of the threshold determination method described is limited whereas spatial precision is high. Hence a more accurate threshold determination technique has to be added if the postulate of great threshold accuracy is to be fulfilled.
The new Octopus programs SARGON and DELTA are described. The SARGON program makes possible the arbitrary distribution of a maximum number of 66 test locations with a resolution of up to 0.2 degrees across the 60 degrees visual field. Eighty user-defined self-created programs can be permanently stored and recalled later at will from the program diskette.. The program DELTA evaluates perimetric examination results with the aid of a series of statistical tests and investigates whether the data base is sufficient to assume a significant deviation, or alternatively, whether a visual field defect can be explained simply by spontaneous fluctuations or if a true pathologic defect should be assumed at a particular level of significance. Furthermore, the program tests changes in the visual field as a function of time for their statistical significance.
The local mean and the average difference of four pairs of test locations within the 26 degrees visual field, situated above and below the horizontal nasal meridian, were used to predict the global field indices MD and CLV of the Gl glaucoma program. Out of 539 examinations (194 eyes suspected of having glaucoma), the local indices NDIFF (describing asymmetrical behavior around the nasal horizontal meridian), ND0 (the mean defect in the nasal region), and the global indices MD and CLV were calculated. Seven hundred fifty-five examinations (446 normal eyes) served as a control group. First and second examinations of 146 glaucoma suspect eyes were used to calculate the retest reliability scores for the indices in question. When analyzing the glaucoma suspects, the local index NDIFF, together with the local mean defect, ND0, yielded highly reliable estimates of the global indices MD and CLV, with a retest correlation r = 0.86 for NDIFF, and r = 0.96 for ND0. The covariance of NDIFF with CLV was r = 0.67, while the co-variance of MD with ND0 was r = 0.95. The ranges of the local indices ND0 and NDIFF were each classified into 'normal range' and 'range of suspected pathology', in analogy to the normal and pathological ranges of the global field indices. Equivalence of the local indices with the corresponding ranges of MD and CLV was investigated and the results are shown. The establishment of local indices may prove to be a powerful tool in early detection of glaucomatous damage.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.