Importance of Developing a Decision Support
System for Diagnosis of Glaucoma
Murat Durucu
Industrial Engineering Department, Management Faculty, Istanbul Technical University,
Istanbul, Turkey
[email protected]
Abstract Glaucoma is a condition of irreversible blindness, early diagnosis and
appropriate interventions to make the patients able to see longer time. In this study,
it addressed that the importance of developing a decision support system for glaucoma diagnosis. Glaucoma occurs when pressure happens around the eyes it causes
some damage to the optic nerves and deterioration of vision. There are different
levels ranging blindness of glaucoma disease. The diagnosis at an early stage allows
a chance for therapies that slows the progression of the disease.
By using Optical Coherence Tomography (OCT) images and pattern recognition
systems, it is possible to develop a support system for doctors to make their decisions on glaucoma. Thus, in this recent study we develop an evaluation and support
system to the usage of doctors. Pattern recognition system based computer software
would help the doctors to make an objective evaluation for their patients. It is intended that after development and evaluation processes of the software, the system
is planning to be serve for the usage of doctors in different hospitals.
Keywords Decision Support System, Glaucoma, Image Processing, Pattern Recognition.
Introduction
This study focused on necessity for developing an objective decisions support system for evaluating level and occurrence of glaucoma disease. Glaucoma is diagnosed with considering patients’ family history and by using clinical techniques,
such as tonometry, ophthalmoscopy, perimetry, gonioscopy and pachymetry. Glaucoma is an irreversible blindness for the patients in late stages. The diagnosis at an
early stage allows for therapies that slow the progression of the disease. Also diagnosis at an early stage can decrease the socio-economic wages for the patients and
country they live in (Mazhar 2013). However, problems have been experienced in
2
the diagnosis of the disease with clinical examination, so it is necessary to develop
new techniques for the diagnosis can be made at an early stage. Often encountered
in this disease especially between 40-80 years old, despite depends on genetic factors. 3:54% of the population worldwide is glaucoma patients. Worldwide, according to data from 2013, 64.3 million people that glaucoma patients between 40-80
years of age, in 2020 this number increased by 76 million (Quigley and Broman
2006), and is expected to reach 111.8 million in 2040 (Tham, Li et al. 2014). In
some studies, glaucoma affects 44.7 million people across the world, 2.8 million of
them lived in the United States, and as a result, 1.6 million direct, and indirect costs,
including 0.9 million in total; It is reported that generate $ 2.5 billion in costs. Development of the glaucoma can be seen in Figure 1. Increased blood pressure inside
the eye pupil, is damages the optical nerves.
Fig. 1. Development of glaucoma
The major symptoms of glaucoma include (1) Blurred vision (2) Severe pain in
the eye (3) Rainbow hallows with light Headache (4) Brow pain Nausea (5) Vomiting with Red Eye. The intraocular pressure is also identified as one of the risk
factors which develop the glaucomatous damage and lower pressure leads to progressive retinal degenerative change. (Murthi and Madheswaran 2014).
There are different levels of the glaucoma in the literature. The comparison between normal vision and the glaucoma eye can be seen in Figure 2. Description of
the levels of the glaucoma for ophthalmologist can be found in Table 1.
3
Fig. 2. Level samples for patients’ visions
Table 1. Glaucoma levels
Light or early level glaucoma
Glaucoma is defined as the optic nerve abnormalities that are
similar, but there is no abnormality in the visual field white upon
white field test.
Mid-level glaucoma
Glaucoma is an optic nerve abnormalities compatible with. Encountered a hemisphere and 5 is very hard glaucoma anomaly.
Extreme, last term level glaucoma
Glaucoma is consistent with glaucoma optic nerve and visual
field abnormalities anomalies are found in both hemispheres, are
seen at least 5 degrees loss of vision in one hemisphere fixed.
Uncertain glaucoma
Visual fields were not yet occurred, or is not suitable for patients
with visual field testing or visual field tests applicable / unreliable state.
In recent years, imaging technology from Heidelberg retinal tomography (HRT),
Stereoscopic disc Photo (SDP) and Optical Coherence Tomography (OCT) imaging
technology such as is used for the diagnosis of glaucoma (Mwanza and Budenz
2016). This better accuracy and faster imaging techniques in response technique of
OCT has become the most common method used by experts. Retinal Nerve Fiber
Layer (RNFL), optic nerve head (ONH) and reasonable analysis are applied to detect any glaucoma damage OCT (Bai, Niwas et al. 2016). Due to all economic facts
that, early diagnosis of glaucoma is very important as patients’ life quality as economical costs. Clinically, the diagnosis of Glaucoma can be done through measurement of CDR, defined as the ratio of the vertical height of the optic cup to the vertical height of the optic disc. An increment in the cupping of Optic Nerve Head
(ONH) corresponds to the increased ganglion cell death and hence CDR can be used
to measure the probability of developing the disease. A CDR value that is greater
than 0.65 indicates the high glaucoma risk (Li and Chutatape 2003).
In the Figure 2 you can find the medical imaging for normal eye and affected eye
(Murthi and Madheswaran 2014). With the defect of optical nerves irreversible
blindness will be begun.
4
Fig. 3. Medical imaging of normal eye and affected eye (Murthi and Madheswaran 2014)
Methodology
To date, procedures that have been employed for detection of glaucomatous visual
field progression may be broadly grouped into four categories: subjective clinical
judgment, defect classification systems, trend analyses, and event analyses.
Clinical Judgment
Clinical judgment consists of simple subjective observation of sequential visual
field test results and represents the oldest method for identification of progressive
visual field defects. This approach is advantageous for a number of reasons: 1) it
demands no additional computation; 2) it is highly flexible as observers with any
degree experience may apply it to the results of any instrumentation; and 3) it is
easy to perform. However, the subjectivity of this approach means that it is also
poorly controlled, and criteria can vary considerably from one evaluator to another.
A comparison of visual field series evaluated for deterioration, stability, or improvement by six expert observers illustrates clinical judgment’s disadvantages (Werner,
Bishop et al. 1988). When using clinical judgment to assess data, care should be
taken when a patient has been examined with a variety of threshold estimation algorithms.
5
Defect Classification Systems
Visual field defect classification systems use predetermined criteria to grade single test results, providing a discrete score for each visual field test result. The advantages of this approach are that test results are immediately stratified into broadly
similar defect magnitudes, interpretation is relatively simple, and progression can
be easily defined as worsening of the score over time. There are, however, a number
of drawbacks to use of classification systems. They do not provide information on
the spatial configuration of defects and may not be scaled linearly, for example, a
change from 0 to 3 may not be equal to a change from 10 to 13.
Trend Analyses
Trend analyses evaluate test parameters sequentially to determine temporal patterns that may exist within the data (Holmin and Krakau 1980, Holmin and Krakau
1982). Such analyses are of value because they are capable of determining longterm characteristics with use of information from all visual field examinations performed on a patient, and therefore have the potential to discriminate subtle progressive loss from considerable degrees of test variability (Fitzke, Hitchings et al. 1996).
Event Analyses
Event analyses are valuable because they attempt to identify single events of
significant change relative to a reference examination (Hitchings 1994). Event analyses can be relatively simple, and can look for statistically significant differences
between one examination and another, such as used within the DELTA program of
the Octopus perimeter. This particular method employs a paired t test to determine
whether significant differences are present between one test result and another.
Conclusion
Although OCT images or HRT precision and quickness, especially in the early
stages, difficulty and mistakes are experienced in diagnosis of glaucoma. To be in
the discretion of the doctor's diagnosis and placement process, it is difficult to obtain
objective results. It is very important to develop an objective decision support system for diagnosis and level the glaucoma disease for patients.
6
In recent years computer aided diagnosis (CAD) is playing a major role in
screening the glaucoma. The CAD system is simple, repetitive, not prone to inter or
intra observer variability and fast in diagnosis. Also CAD can screen many patients
in a small time. There is a scarcity of ophthalmologists in many developing countries, where CAD can be very useful. The proposed decision support system for
glaucoma can differentiate normal and glaucoma classes accurately
By using OCT images and pattern recognition systems, it is possible to develop
a support system for doctors to make their decisions on glaucoma. For this purpose,
an evaluation and support system will be developed, and will be offered to usage of
doctors. Pattern recognition system based computer software will be evaluating the
level of glaucoma between none to severe, end-stage glaucoma. After evaluation
processes of the software, the system is planning to be serve for the usage of medical
personnel in different hospitals.
References
Bai, X. L., S. I. Niwas, W. S. Lin, B. F. Ju, C. K. Kwoh, L. P. Wang, C. C. Sng, M. C. Aquino and
P. T. K. Chew (2016). "Learning ECOC Code Matrix for Multiclass Classification with Application to Glaucoma Diagnosis." Journal of Medical Systems 40(4): 10.
Fitzke, F. W., R. A. Hitchings, D. Poinoosawmy, A. I. McNaught and D. P. Crabb (1996). "Analysis of visual field progression in glaucoma." The British Journal of Ophthalmology 80(1): 4048.
Hitchings, R. A. (1994). "Perimetry--back to the future?" The British Journal of Ophthalmology
78(11): 805-806.
Holmin, C. and C. E. Krakau (1980). "Visual field decay in normal subjects and in cases of chronic
glaucoma." Albrecht Von Graefes Arch Klin Exp Ophthalmol 213(4): 291-298.
Holmin, C. and C. E. Krakau (1982). "Regression analysis of the central visual field in chronic
glaucoma cases. A follow-up study using automatic perimetry." Acta Ophthalmol (Copenh)
60(2): 267-274.
Li, H. and O. Chutatape (2003). A Model-Based Approach For Automated Feature Extraction In
Fundus Images. Proc. of the 9th IEEE International Conference on Computer Vision.
Mazhar, S. (2013). "Nuggets in clinical approach to diagnosis of glaucoma." Journal of Clinical
Ophthalmology and Research 1(3): 175-181.
Murthi, A. and M. Madheswaran (2014). "Medical Decision Support System to Identify Glaucoma
using Cup to Disc Ratio." Journal of Theoretical and Applied Information Technology 68(2):
406-413.
Mwanza, J. C. and D. L. Budenz (2016). "Optical coherence tomography platforms and parameters
for glaucoma diagnosis and progression." Current Opinion in Ophthalmology 27(2): 102-110.
Quigley, H. A. and A. T. Broman (2006). "The Number of People With Glaucoma Worldwide In
2010 and 2020." Br. J. Ophthalmol 90(3): 262-267.
Tham, Y.-C., X. Li, T. Y. Wong, H. A. Quigley, T. Aung and C.-Y. Cheng (2014). "Global Prevalence of Glaucoma and Projections of Glaucoma Burden through 2040: A Systematic Review
and Meta-Analysis." Ophthalmology 121(11): 2081-2090.
Werner, E. B., K. I. Bishop, J. Koelle and et al. (1988). "A comparison of experienced clinical
observers and statistical tests in detection of progressive visual field loss in glaucoma using
automated perimetry." Archives of Ophthalmology 106(5): 619-623.