Automatization of Dry Eye Syndrome Tests

0 downloads 0 Views 2MB Size Report
phenomena as a color texture pattern, using image-processing techniques. .... theory of human color vision, proposed by Hering [33] in the 1800s, is used in ... opponent colors, are red versus green, green versus red, and blue versus yellow.
16 Automatization of Dry Eye Syndrome Tests Manuel G. Penedo, Beatriz Remeseiro, Lucía Ramos, Noelia Barreira, Carlos García-Resúa, Eva Yebra-Pimentel, and Antonio Mosquera CONTENTS 16.1 Tear Film.............................................................................................................................. 294 16.1.1 Structure and Function.......................................................................................... 295 16.1.1.1 Lipid Layer................................................................................................ 295 16.1.1.2 Aqueous Phase......................................................................................... 295 16.1.1.3 Mucous Layer........................................................................................... 295 16.1.2 Composition............................................................................................................ 296 16.2 Dry Eye Syndrome.............................................................................................................. 296 16.2.1 Semiology................................................................................................................. 296 16.2.2 Epidemiology.......................................................................................................... 297 16.2.3 Classification............................................................................................................ 297 16.2.3.1 ADDE......................................................................................................... 297 16.2.3.2 EDE............................................................................................................. 297 16.3 Morbidity of Dry Eye.......................................................................................................... 298 16.3.1 Impact of Dry Eye on Visual Function and Quality of Life.............................. 298 16.4 Tear Film Assessment........................................................................................................ 299 16.4.1 Biomicroscopic Examination................................................................................. 299 16.4.2 Clinical Tests............................................................................................................300 16.4.2.1 Quantitative Tear Film Tests...................................................................300 16.4.2.2 Qualitative Tear Film Tests..................................................................... 301 16.4.2.3 Osmolarity of Tear Film.......................................................................... 302 16.4.3 Laboratory Tests...................................................................................................... 303 16.5 Evaluation of Interference Lipid Pattern......................................................................... 303 16.5.1 Extraction of the Region of Interest......................................................................305 16.5.2 Texture Analysis.....................................................................................................305 16.5.2.1 Butterworth Filters...................................................................................306 16.5.2.2 Gabor Filters..............................................................................................306 16.5.2.3 Discrete Wavelet Transform...................................................................306 16.5.2.4 Markov Random Fields........................................................................... 307 16.5.2.5 Co-Occurrence Features......................................................................... 307 16.5.3 Color Analysis......................................................................................................... 307 16.5.3.1 Lab Color Space........................................................................................ 307 16.5.3.2 RGB Color Space: Opponent Colors......................................................308 16.5.4 Results......................................................................................................................308

293

294

Image Analysis and Modeling in Ophthalmology

Downloaded by [Beatriz Remeseiro] at 09:45 11 March 2014

16.6 Evaluation of the Tear Film Breakup Time..................................................................... 310 16.6.1 Location of Measurement Areas........................................................................... 311 16.6.2 Extraction of the ROI.............................................................................................. 312 16.6.3 BUT Measurement.................................................................................................. 313 16.6.4 Results...................................................................................................................... 316 References...................................................................................................................................... 317

Dry eye is a symptomatic disease that affects activities of daily living, adversely impacting important tasks such as computer use, driving, and others. Based on data from the largest studies of dry eye to date, the Physicians’ Health Study (PHS) and other studies, it has been estimated that about 3.23 million women and 1.68 million men (a total of 4.91 million Americans) 50 years and older have dry eye. In practice, there are several clinical tests to diagnose this syndrome by means of analyzing tear film quality. This chapter describes automatic image-processing methodologies to perform two clinical tests: analysis of the interference lipid pattern and the tear film breakup time test.

16.1  Tear Film The tear film is a complex layer of liquid covering the anterior surface of the eye. Classically, it is described as a trilaminar structure consisting of a thin anterior lipid layer (0.1–0.05 μL), an intermediate aqueous layer (7 μL) and an innermost mucous layer (0.02–0.04 μL) [1] (Figure 16.1). All of these layers must work properly to keep the eye moist and free from dry eye [2]. The tear film is not evenly distributed over the ocular surface. The total volume of the tear film is 7.0 ± 2.0 μL, with a thickness ranging from 6 to 10 μm. Along the upper and lower lids it forms a tear meniscus or marginal tear strips. This represents 70% of the total volume of tear fluid within the palpebral aperture [1]. A small proportion lies beneath the eyelids between the palpebral and bulbar conjunctivae, and the remainder covers the cornea and the exposed bulbar conjunctiva [2].

Lipid layer

Aqueous phase Mucous layer Corneal epithelium FIGURE 16.1 Trilaminar structure of the tear film.

0.1–0.05 µL

7 µL

0.02–0.04 µL

Automatization of Dry Eye Syndrome Tests

295

16.1.1  Structure and Function The structure of the tear film is classically described as a trilaminar structure comprising a superficial lipid layer, an intermediate aqueous phase, and an underlying mucous layer.

Downloaded by [Beatriz Remeseiro] at 09:45 11 March 2014

16.1.1.1  Lipid Layer This is the outermost and thinnest layer of the tear film and is mainly secreted by the meibomian glands, embedded in the upper and lower tarsal plates [3]. The normal lipid layer is divided into two layers: the anterior and posterior lipid layers. The anterior lipid layer is formed by nonpolar lipids, mainly mixed wax esters and sterol esters (more than 60% of the total lipids). The posterior lipid layer is formed by highly polar lipids: free sterols, free fatty acids, and phospholipids [2]. The main function of the lipid layer is the reduction of evaporation from the aqueous phase. Furthermore, the nonpolar structure of the lipid layer is important in preventing surface contamination of the film with highly polar skin lipids, which could disrupt the tear film. 16.1.1.2  Aqueous Phase The aqueous phase is the major component of the film, comprising around 98% of its total thickness. It is a complex dilute solution of both inorganic electrolytes and low- and highmolecular-weight organic substances. This is the major, intermediate phase of the tear film and is approximately 6.5–7.5 μm thick [1]. This phase is mainly secreted by lacrimal glands situated in the superior temporal angle of the orbit. This phase contains many ions and molecules such as electrolytes, hydrogen ions, proteins, enzymes, and metabolites, which provide the proper functions of the tear film. 16.1.1.3  Mucous Layer The preocular tear film of the human eye is dependent on a constant supply of mucus, which must be of adequate chemical and physical quality to maintain the corneal and conjunctival surfaces in the proper state of hydration [2]. Secreted by goblet cells (sited in the conjunctiva), it represents 0.2% of the whole tear film, is formed by glycoproteins, and performs several functions. The main function of these mucous glycoproteins is to lower the surface tension of tears from about 70 dynes/cm to about half of that value. Lubrication of the cornea is also an important function, allowing the lids to slide smoothly with minimal friction during blinking and wetting the ocular surface to maintain a stable tear film [1]. The tear film is an essential component of the eye that fulfils some important functions [1,2]. The first function is the lubrication of the eyelids, which allows the eyelid margins and palpebral conjunctiva to slide smoothly over one another with minimal friction during blinking. The visual function is another role of the tear film. The corneal epithelium is a rough surface and the tear film fills the small irregularities in the corneal epithelium, providing a smooth, regular optical surface to guarantee high optical quality. The tear film also has a cleaning function. It takes desquamated epithelial cells, debris, and so forth from the epithelium and, due to the action of the lids, these are “washed” from the surface of the eye. In addition, since the corneal surface is avascular, the nutrition is driven by the tear film. Oxygen dissolves in the tear fluid and nutrients, such as glucose, are passed from the palpebral conjunctival vessels into the tear film. Finally, the tear film contains proteins (such as lysozyme or lactoferrin) that inhibit microbiological contamination.

296

Image Analysis and Modeling in Ophthalmology

16.1.2 Composition

Downloaded by [Beatriz Remeseiro] at 09:45 11 March 2014

The tear film is a matrix-like structure composed of water, electrolytes, immunoglobulins, antimicrobial molecules, mucins, and so on. The composition of the tear film is distributed into the three layers. The lipid layer comprises polar and nonpolar lipids secreted by meibomian glands. The major, intermediate watery phase of the tear film contains dissolved ions, organic solutes, and other metabolites. Finally, the mucous layer is made up of glycoproteins secreted by goblet cells of the conjunctiva. The concentration of each layer is as follows: • Lipid layer: The major lipid classes are wax esters, sterol esters (mainly cholesterol), diesters, triglycerides, free sterols, and polar lipids. Hydrocarbons and free fatty acids, both straight and branched chain, are also present [1]. • Aqueous phase: This layer contains many ions and molecules including electrolytes (sodium, potassium, calcium, chloride, bicarbonate, and phosphate ions), proteins (lysozyme, lactoferrin, albumin, lipocalin, secretory immunoglobulin A), enzymes (lactate dehydrogenase, malate dehydrogenase, pyruvate kinase), and metabolites (glucose, urea) [2]. • Mucous layer: This layer mainly comprises O-linked glycoproteins. The principal sugars in crude ocular mucus are sialic acid, galactose, glucose, and N-acetylglucosamine [2].

16.2  Dry Eye Syndrome Commonly, the term “dry eye” has been used to describe a variety of conditions of diverse origin that affect the tear film, the ocular surface, or both. However, there has always been a lack of consensus on the main characteristics of dry eye syndrome (DES), leading to confusion between clinicians. To solve this, the international committee of the Dry Eye WorkShop (DEWS) clarified the main aspects of DES and established a complete definition as follows [4]: Dry Eye is a multifactorial disease of the tears and ocular surface that results in symptoms of discomfort, visual disturbance, and tear film instability with potential damage to the ocular surface. It is accompanied by increased osmolarity of the tear film and inflammation of the ocular surface.

From this definition, some factors can be elucidated. First, DES is a multifactorial disease, which means that several factors are responsible for the condition. Therefore, in order to establish an accurate DES diagnosis, a battery of clinical tests may be necessary. Second, a stable tear film covering the ocular surface is essential to maintain eye physiology, otherwise the ocular surface will be damaged. Third, a poor tear film shows increased osmolarity. This hyperosmolarity is considered to be the primary cause of ocular surface damage and inflammation. Thus, DES represents an inflammatory status. 16.2.1 Semiology Nichols et al. have reported that the most common symptom of DES is dryness (98.7%), followed by ocular fatigue (85.1%), grittiness (78.5%), redness (71.6%), and soreness (64.8%) [3].

Automatization of Dry Eye Syndrome Tests

297

However, other studies have found ocular fatigue to be the most frequent symptom, followed by dryness [2]. Irritation, foreign-body sensation, burning, the presence of stringy mucous discharge, and transient blurring of vision also can affect dry eye patients [2]. DES cannot be diagnosed based on symptoms alone, since most of them can be caused by other conditions.

Downloaded by [Beatriz Remeseiro] at 09:45 11 March 2014

16.2.2 Epidemiology One of the main problems in obtaining data on prevalence (the proportion with the disease within a population at a given point in time) is the different criteria of DES diagnosis between studies. There is consensus, however, that the prevalence of DES has increased, mainly due to current lifestyles [5]. Several factors can disrupt the tear film such as harmful environments (pollution, smoke, etc.), increased visual tasks that decrease the blink rate (especially when done with a computer), and increased use of antihistamines, antihypertensives, and so on [6]. These environmental and work factors, together with the ageing of the population (tear stability and secretion decrease with age) have increased the proportion of the population with DES [7]. Taking into account the lack of consensus discussed above, prevalence data found in the literature ranks from 10% to 35% [5]. If only contact lens wearers are considered, this prevalence is even greater [4]. Despite this discrepancy of data prevalence between studies, the weight of the evidence from large epidemiological studies indicates that female sex and older age (greater than 50 years) increase the risk for dry eye [2]. 16.2.3 Classification From an etiopathogenic point of view, DES can be mainly classified as aqueous teardeficient dry eye (ADDE) and evaporative dry eye (EDE). Dry eye can be initiated in either of these classes, but they are not mutually exclusive [4,8]. 16.2.3.1 ADDE ADDE implies that dry eye is due to a failure of lacrimal tear secretion. In any form of dry eye due to lacrimal acinar destruction or dysfunction, dryness results from reduced lacrimal tear secretion and volume. This causes tear hyperosmolarity and stimulates a cascade of inflammatory events. ADDE has two major subclasses: Sjogren’s syndrome dry eye and non-Sjogren’s syndrome dry eye. Sjogren’s syndrome is an exocrinopathy in which the exocrine glands (such as lacrimal and salivary glands) are targeted by an autoimmune process. Non-Sjogren’s syndrome dry eye does not present those systemic autoimmune features. The most common form of non-Sjogren’s syndrome dry eye is agerelated dry eye, although other factors can contribute to ADDE, such as lacrimal gland infiltration, sarcoidosis, lymphoma, obstruction of the lacrimal gland ducts, and reflex hyposecretion [4,8]. 16.2.3.2 EDE EDE is caused by excessive water loss from the exposed ocular surface in the presence of normal lacrimal secretory function. The volume and composition of the lacrimal fluid are adequate, with tear abnormality created by other periocular diseases, usually leading to increased tear evaporation. This is the type of dry eye most commonly found in young

298

Image Analysis and Modeling in Ophthalmology

Downloaded by [Beatriz Remeseiro] at 09:45 11 March 2014

to middle-aged people, and related to ambient conditions (air conditioning), contact lens wear, or both [4,8]. Current work conditions, such as computer use, have increased the proportion of people with EDE [4]. The main cause of EDE is meibomian gland dysfunction [9]. Meibomian glands, embedded in the upper and lower tarsal plates, are responsible for lipid secretion, which is essential to retard tear film evaporation [9]. Blink rate is an important factor in tear film stability. If the blink rate is reduced, the period between blinks is lengthened (the ocular surface is exposed to water loss), increasing tear film evaporation [10]. This mainly occurs during the performance of concentration tasks, such as reading, but most remarkably when using a computer. Contact lens wear may cause tear instability and dry eye because the contact lens disrupts the tear film. In fact, the primary reasons for contact lens intolerance are discomfort and dryness [4].

16.3  Morbidity of Dry Eye The high prevalence of dry eye among older people, together with the ageing of the population, makes DES a public health problem to be considered. Dry eye is a prevalent condition with the potential for a high economic burden. From a public health point of view, dry eye supposes an economic impact derived from costs due to health-care system utilization, including health-care professional visits, nonpharmacological therapies, pharmacological treatments, and surgical procedures, with the latter two categories being the major cost drivers. Other costs include complementary and alternative therapeutics [11], purchase of specialized eyewear, and other therapeutics, such as humidifiers. Given the prevalence of the condition, indirect costs derived by DES may be large [11]. These can include pharmacological therapies other than tear replacements, complementary medicine use, and the cost of complications of surgical procedures. Additionally, costs derived from impacts on daily capacity can include lost work time and productivity, alterations in work type or environment, decreased work time, and days off work with dry eye symptoms. Intangible costs include decreased leisure time, impaired physical functioning and quality of life, and impact on social interactions and mental and general health. In a survey designed to ascertain how much a patient’s everyday activities were limited by symptoms of dry eye, it was found that patients with DES were significantly more likely to report problems with reading, carrying out professional work, using a computer, watching television, driving during the day, and driving at night [12]. Overall, patients with DES were about three times more likely to report problems with common activities than those without DES. With increased severity, patients also reported deficits in general health perception and vitality, and the most severely affected patients reported worse health-related domains [12]. Dry eye is associated with contact lens intolerance and discontinuation of contact lens wear, can adversely affect refractive surgery outcomes, and may be associated with increased risk of infection and complications with ocular surgery. Severe dry eye may lead to ocular damage [13], although the natural history of dry eye remains to be determined. 16.3.1  Impact of Dry Eye on Visual Function and Quality of Life Dry eye limits and degrades visual performance and affects common vision-related daily activities. Visual complaints are highly prevalent among dry eye patients and are usually

299

Downloaded by [Beatriz Remeseiro] at 09:45 11 March 2014

Automatization of Dry Eye Syndrome Tests

described as disturbed vision or blurry, foggy vision that clears temporarily with blinking. These transient changes can be profound, resulting in marked drops in contrast sensitivity and visual acuity, thus affecting workplace productivity and vision-related quality of life [14]. In this sense DES affects quality of life in different ways. Pain and irritative symptoms negatively disrupt the welfare of the patient. Also, DES affects ocular and general health. Perception and visual function may be reduced, which impacts visual performance. For example, the irritative symptoms of dry eye can be debilitating and result in both psychological and physical effects that impact quality of life. Dry eye also limits and degrades performance of common vision-related daily activities, such as driving. The need for frequent instillation of lubricant eye drops can affect social and workplace interactions. From an ocular surface health point of view, corneal surface irregularity due to epithelial desiccation, tear film instability, and evaporation can occur. An uneven, disrupted tear film in the central cornea can result in transient vision changes in the dry eye patient. Optical aberrations created by tear film breakup between blinks contribute to a decline in retinal image quality that can be measured by both objective and subjective methods [14].

16.4  Tear Film Assessment DES diagnosis is very difficult to achieve, mainly because this is a multifactorial syndrome, so several tests are necessary to obtain a clear diagnosis. Fortunately, there are a wide number of tests to evaluate different aspects of the tear film. Some clinical tests measure the quality and quantity of tears, but most have high variability and demonstrate poor to fair diagnosis repeatability [4]. 16.4.1  Biomicroscopic Examination This is an essential part of the protocol of eye examination, consisting of a biomicroscope (designed to view the anterior eye with high magnification) and a slit lamp to change the illumination technique (Figure 16.2a). It is not exclusive to tear film assessment and

(a)

(b)

FIGURE 16.2 (a) Eye examination by biomicroscopy. (b) Appearance of anterior eye dyed with fluorescein under biomicroscopic observation with cobalt blue and yellow filters. Light dots indicate damage of corneal epithelium surface.

Downloaded by [Beatriz Remeseiro] at 09:45 11 March 2014

300

Image Analysis and Modeling in Ophthalmology

provides information about the health of the ocular surface. Furthermore, biomicroscopy is essential during the performance of various tear film tests. Biomicroscopic examination is essential for grading ocular surface staining, which is a sign of epithelial damage. Various stains are used, sodium fluorescein and lissamine green being the most common. Sodium fluorescein is a dye indicated to observe the corneal epithelium. Under biomicroscopic observation with cobalt blue and yellow filters, we can see potential corneal epithelium damage as green dots (Figure 16.2b) that, in severe cases, can form a confluent path. Lissamine green dye is used to observe the conjunctival epithelium, but in this case filters are not necessary. Dry eye is a symptomatic disease, so the presence and status of symptoms need to be ascertained. Symptoms vary in severity according to the state of instability of the tear film and damage of the ocular surface [2]. One problem that can arise is the subjective interpretation of symptoms by the patient. In order to solve this, standardized symptom questionnaires have been developed for use in dry eye diagnosis. The aim of these questionnaires is to offer objective symptom evaluation and guarantee comparisons. Various questionnaires are available but the most used are McMonnies questionnaire, the Ocular Surface Disease Index (OSDI) questionnaire, and the Dry Eye Questionnaire (DEQ). 16.4.2  Clinical Tests Clinical tests are those that can be performed in clinical settings as part of a routine eyecare examination. Classically, these tests have been divided into two groups: quantitative tests and qualitative tests. Quantitative tests assess quantity of tear film and are related to tear secretion, whereas qualitative tests assess the stability of tear film [8]. 16.4.2.1  Quantitative Tear Film Tests Quantitative tear film tests are related to the lacrimal gland secretion function. Defective lacrimal function is usually demonstrated by showing reduced aqueous tear volume and tear flow. The most used quantitative tear film tests are listed here. 16.4.2.1.1  Tear Meniscus Assessment It has been estimated that the tear meniscus holds 75%–90% of the total volume of the tear film. Therefore, careful examination of the tear meniscus provides a useful indication of the tear volume [15]. Tear meniscus parameters most commonly used for tear volume examination are tear meniscus height, tear meniscus radius of curvature, and subjective examination of the meniscus. The main advantage of this tear meniscus examination is its noninvasive nature. TMH is the parameter most used for assessing tear menisci. It can be assessed in clinical settings but biomicroscopy with high magnification is necessary to obtain enough resolution. With the aid of a graticule eyepiece, it is possible to measure TMH, from the lid to the top of the meniscus [15] (Figure 16.3a). The cutoff point between normal and dry eye is ≤0.1 mm. However, measuring the TMH with these settings could be difficult due to the lack of contrast to distinguish the highest limit of the meniscus [12]. Tearscope Plus, a device designed to evaluate the anterior lipid layer (Figure 16.3b) (see description below), can be used to evaluate TMH. Subjective assessment of the TMH is frequently used by clinicians and has been advocated as part of routine ocular assessments, where the main features evaluated

301

Automatization of Dry Eye Syndrome Tests

(a)

(b)

(c)

Downloaded by [Beatriz Remeseiro] at 09:45 11 March 2014

FIGURE 16.3 (a) Appearance of tear meniscus height by slit lamp biomicroscopy. (b) Appearance of tear meniscus height by Tearscope Plus. (c) Phenol red thread test in use.

qualitatively were the presence/absence of debris or foaming in the tear meniscus and its regularity. However, the reliability of this evaluation is affected by subjective criteria of the observer [15]. 16.4.2.1.2  Schirmer Test The Schirmer test estimates tear flow stimulated reflexly by insertion of a filter paper into the conjunctival sac. The strip is placed at the junction of the middle and lateral thirds of the lower eyelid and the patient is told to keep the eyes closed. After 5 min, the strip is replaced and the length of wetted strip is measured. The cutoff value to distinguish normal and dry eye subjects is established at 5 mm. Although this is a test widely employed, it is affected by a wide variability [2]. This test is very uncomfortable for the patient. 16.4.2.1.3  Phenol Red Thread Test The phenol red thread test consists of a thread impregnated with phenol red, which is pH sensitive and changes from yellow to red over the pH range of normal tears (Figure 16.3c). This test only needs to be hooked over the lower lid for 15 sec and is more comfortable than the Schirmer test (it is barely noticeable for the patient). The cutoff value between controls and dry eye is 10 mm [4]. 16.4.2.2  Qualitative Tear Film Tests Qualitative tear film tests are related to the ability of the tear film to remain stable, which is essential to cover the anterior eye and perform its functions (optical, nutritive, antimicrobial, and cleaning). Secretion of the lacrimal gland may be normal and the volume and composition of the lacrimal fluid adequate, but there could be tear abnormality driven by other factors that can lead to increased tear evaporation. The most used qualitative tear film tests are listed here. 16.4.2.2.1  Lipid Layer Pattern Assessment Tear film quality and lipid layer thickness can be assessed by noninvasively imaging the superficial lipid layer with interferometry. The Tearscope Plus, designed by Guillon, is the instrument of choice for rapid assessment of lipid layer thickness [16] (Figure 16.4a). With this instrument, a qualitative analysis of the lipid layer structure can be made.

302

Image Analysis and Modeling in Ophthalmology

Downloaded by [Beatriz Remeseiro] at 09:45 11 March 2014

(a)

(b)

FIGURE 16.4 (a) Tearscope Plus. This device projects a cylindrical source of cool white fluorescent light onto the lipid layer. (b) NIBUT test with Tearscope Plus. Grid pattern projected onto the precorneal tear film for the observation of distortion and/or abnormality in the image.

16.4.2.2.2  Tear Breakup Time (BUT) Tear BUT consists of measuring the time that the tear film remains stable without blinking. After fluorescein instillation, the patient is asked to keep the eye open until a sign of tear film rupture (dark spot) appears. This is the most commonly used test of tear film stability and the cutoff for dry eye diagnosis most used is 90 nm) are readily observed since they produce color and wave patterns. However, thin lipid layers (