Abstract: The aim of the this study was to introduce the design of an expert system which was able to fully diagnoses and treat Pancreas, Thyroid and Parathyroid glands diseases; furthermore, it gave first aids in emergency cases caused by diabetes. Since, diabetes diseases are widely spreads in Gaza, we chose it to be the primary target from the endocrine diseases. Our expert system was not meant to replace the human physician but using such system may be useful in cases like overcoming the problems of the shortage in human physicians and accuracy and speed in processing facts. This system can be used to help the physician in their work. Our expert system was initially evaluated with existing classical test cases. The result of the evaluation was accurate and promising.
INTRODUCTION
The endocrine system is classified to be among the most important systems in the human body show the endocrine system (Fig. 1), the glands of the endocrine system and the hormones they release not only influence organs in human body but also influence every cell and the overall functions of our bodies.
Fig. 1: | Endocrine system |
Most diseases nowadays are mainly caused by the inadequate performance of the endocrine system, regulating mood, sexual function, metabolism and growth are all depends on the functionality of the endocrine system.
Generally, the endocrine system is in charge of the body processes that occur slowly, such as cell growth. Faster processes similar to breathing and body movement are managed by the nervous system. But even though the nervous system and endocrine system are distinct systems, they frequently work together to help the body function appropriately (www.kidshealth.org/parent/general/bodybasics/endocrine.html).
Due to the importance of the Endocrine system, we designed and developed an expert system for the followings:
• | Diagnosing cases related to Pancreas, Thyroid and Parathyroid glands diseases and give possible treatments |
• | Helping newly graduated physician in diagnoses patients cases and learn from it |
• | Physicians could use the system to follow up with their patients treatments or to use for the very urgent cases |
• | Concession students can use the expert system for training instead of going to the hospitals which are always busy and cant bear a large number of students |
• | In ambulances the expert system can be used to serve urgent cases of fainting |
• | In emergency when all physicians are busy with other cases |
Artificial Intelligence (AI) is a subfield of computer science concerned with symbolic reasoning and problem solving (Russel and Norvig, 2002; Abu-Naser et al., 2008). Expert Systems (ES) which is a branch of AI are computer systems that applies reasoning methodologies to knowledge in a specific domain to render advice or recommendation much like a human expert (Durkin, 1994; Giarratano and Riley, 2004; Jackson, 1999).
In our expert system we use the Java Expert System Shell (JESS) to perform its functions, facts and procedures. It is a rule based engine for the Java language platform which is a superset of CLIPS programming language developed by Ernest Friedman-Hill of Sandia National Lab. (Giarratano, 2002). It was first written in late 1995. It provides rule-based programming suitable for automating an expert system and is often referred to as an expert system shell. In recent years, intelligent agent systems have been developed also, which depend on a similar capability. This system requires having Java version 1.6.0 02 or 1.5.0 running on Windows XP SP2 or Windows Vista Home Premium.
MYCIN
It was the first well known medical expert system developed by Shortliffe
at Stanford University (Buchanan and Shortliffe, 1984)
to help doctors, not expert in antimicrobial drugs, prescribe such drugs for
blood infections. The limitation of MYCIN was: its knowledge base is incomplete
since, it does not cover anything like the full spectrum of infectious diseases.
Running it would have required more computing power than most hospitals could
afford at that time (1976). Doctors do not relish typing at the terminal and
require a much better user interface than that provided one.
Easy Diagnosis
It is an expert system software that provides a list and clinical description
of the most likely conditions based on an analysis of your particular symptoms
(Martin, 2004). Easy diagnosis focuses on the most common
medical complaints that account for the majority of physician visits and hospitalizations.
Easy Diagnoses
It is a poorly designed user-interface, the user is required to answer a
large number of questions without any notion that gives him the feeling that
his data is accepted and will be diagnosed.
PERFEX
It is a medical expert system that support solving problems clinicians currently
have in evaluating perfusion studies (Ezquerra et al.,
1992). The heart of the PERFEX system is the knowledge base, containing
over 250 rules. They were formulated using the expertise of clinicians and researchers
at Emory University Hospital. PERFEX limitation resides in its output. It is
mostly numerical.
INTERNIST-I
It is a rule-based expert system designed at the University of Pittsburgh
in 1974 (Kumar et al., 2009) for the diagnosis
of complex problems in general internal medicine.
ONCOCIN
It is a rule-based medical expert system for oncology protocol management
(Wiederhold et al., 2001) developed at Stanford
University. Oncocin was designed to assist physicians with the treatment of
cancer patients receiving chemotherapy.
Dxplain
It is a decision support system which uses a set of clinical findings (signs,
symptoms, laboratory data) to produce a ranked list of diagnoses which might
explain (or be associated with) the clinical manifestations (Elhanan
et al., 1996). The DXplain provides justification for why each of
these diseases might be considered, suggests what further clinical information
would be useful to collect for each disease and lists what clinical manifestations,
if any, would be unusual or a typical for each of the specific diseases.
PUFF
It is an expert system for the interpretation of pulmonary function tests
for patients with lung disease (Aikins et al., 1983).
PUFF was probably the first AI system to have been used in clinical practice.
Those Expert Systems suffer from limitation, bad interface or output format.
Our expert system is specialized in the diagnosis of endocrine system diseases with descriptive output and carefully designed interface.
KNOWLEDGE ACQUISITION
Basic information about the endocrine diseases, symptoms and treatment where
collected from experts (physicians), books, sites and special prepared notes
by clinical physicians. Knowledge elicitation was performed through interviews.
KNOWLEDGE REPRESENTATION
The environment of the system may affect its reliability. The use of some
Expert System programming languages make the system limited in specific features.
In our expert system, we used the Java Expert System Shell (JESS) to perform its functions, facts, rules and procedures. JESS is a rule based engine for the Java platform and it is a superset of CLIPS programming language. CLIPS (C Language integrated production System) was developed by Ernest Friedmanhill of Sandia National Labs (Giarratano, 2002). It was originally written in late 1995 and provided rule-based programming suitable for automating an expert system and is often referred to as an expert system shell. The following rule is an example of how knowledge is represented in CLIPS:
• | (defrule IDDM (Patient (name ?first ?last)(age ?age)) (test (< ?age 30)) (BPressure Hypotension) (Symptoms (ketonuria yes)) (exists (or (Symptoms (coma yes)) (Symptoms(CrackedLips yes)) (Symptoms(Tachycardia yes)) (Symptoms (Confusion yes) (Symptoms (polyuria yes) (polydipsia yes) (polyphygia yes))))) => (assert (Type (type IDDM))) (printout t ?first ?last Has Diabetes type 1 crlf)) |
Currently, our expert system has more than 60 rules which cover: Pancreas, Thyroid and Parathyroid diseases of the endocrine system. Here, is a brief identification of each of the three diseases that our expert system can help the user with:
Pancreas Diseases
The pancreas is a pinkish-grey organ that lies behind the stomach. The organ
is approximately 15 cm in length with a long, slender body connecting the head
and tail segments (Kumar et al., 2004; Martini,
2001; Moore et al., 2009).
The endocrine pancreas is separate from the exocrine pancreas. The endocrine pancreas is made up of small clumps of cells within the pancreas, called pancreatic islets, or the islets of Lange Hans. These account for only 1% of the pancreatic mass. It is composed of three distinct cell types each producing a different hormone. The two important hormones are:
• | Glucagon: Secretion of glucagon is controlled by the level of blood sugar, being released when levels are too low. This greatly increases the output of sugar from the liver and returns blood sugar levels to normal |
• | Insulin: Insulin is needed to convert sugar (glucose), starches and other food into energy needed for daily life. Insulin is designed to lower blood sugar levels when they become too high and is released in periods when there is a lot of sugar available, like after a meal |
Hyperglycaemia
Diabetes Mellitus is a clinical syndrome characterized by hyperglycaemia
due to absolute or relative deficiency of insulin. Lack of insulin, whether
absolute or relative, affects the metabolism of carbohydrates, protein, fat,
water and electrolytes. The Hyperglycaemia is introduced in terms of increasing
the glucose in the blood. That means that there is no insulin to reduce the
percent of the glucagose to its normal level.
Classification of Diabetes Mellitus
Insulin-Dependent Diabetes Mellitus (IDDM) is called the type one that means
it is the first type of diabetes mellitus and this type has characteristics
that different from the second one (Bryant et al.,
2006).
Table 1: | Classification of diabetes mellitus |
The patients with IDDM depend on the insulin in their treatments. Death may results from the absolute deficiency of insulin, so the patients must take the insulin to stay alive. Most IDDMs patients are from children and young people.
The second type of primary diabetes mellitus is the Non-Insulin-Dependent Diabetes Mellitus (NIDDM).
The patients under this type have relative deficiency of insulin and they may take drugs or make diets as treatments (Bryant et al., 2006). That does not mean that they do not take insulin but the insulin is the last choice that may be needed overtime. Most patients in this type are obese and old (Table 1) under excess endogenous production of hormonal antagonists to insulin topic, the probable actions of hormones countering the effect of insulin in humans.
Diagnostic Criteria for Diabetes Mellitus
Persons presented with clinical manifestations that are normally associated
with diabetes (such as polyuria, polydipsia, weight loss and blurred vision)
and/or major risk factors for diabetes, should be referred to the laboratory
for fasting plasma glucose (Rajala et al., 1995;
Garancini et al., 1995).
Hypoglycaemia
Hypoglycaemia, defined as a blood glucose concentration of less than 2.5
mmol L-1, occurs commonly in diabetic patients treated with insulin
and relatively infrequently in those taking a Sulphonylureas drug (Cryer,
2001; Cryer et al., 2009). In most instances
the patient has no difficulty in recognizing the symptoms of hypoglycaemia and
can take appropriate remedial action. However, in certain circumstances (e.g.,
during sleep) and particularly in certain type of patients (e.g., patients with
long duration of IDDM) warning symptoms are not always perceived by the patient
even when awake so that appropriate action is not taken and if no assistance
is available, unconsciousness is the result. Severe hypoglycaemia, defined as
hypoglycaemia requiring the assistance of another person for recovery, can result
in serious morbidity and has recognized mortality of 2 to 4% in insulin-treated
patients. The unrecognized mortality is probably significantly higher than this.
Sudden death in sleep otherwise healthy young patients with IDDM has been described
and has been attributed to hypoglycaemia-induced cardiac arrhythmia.
Recurrent severe hypoglycaemia is very disruptive and impinges on many aspects of the patients life including employment, driving and sport. Risk of hypoglycaemia is the most important single factor limiting attainment of the therapeutic goal, namely normal/near normal glycaemia in patients with IDDM.
Causes of Hypoglycaemia
The main causes of hypoglycaemia in patients taking insulin or a Sulphonylureas
drug are as follows (Cryer, 2001; Cryer
et al., 2009):
• | Missed, delayed or inadequate meal |
• | Unexpected or unusual exercise |
• | Alcohol |
• | Poorly designed insulin regime, particularly that predisposing to nocturnal hypoglycaemia |
• | Defective glucose counter-regulation/unawareness of hypoglycaemia |
• | Gastroparesis due to autonomic neuropathy |
• | Other endocrine disorder, e.g., Addisons disease |
• | Malabsorption |
• | Factitious hypoglycaemia |
• | Other causes include the following: |
GI surgery, Idiopathic, Hepatic disease, Islet cell tumor/extrapancreatic tumor, Exercise (in diabetic patients), Pregnancy, Renal Glycosuria, Ketotic hypoglycemia of childhood, Adrenal insufficiency, Hypopituitarism, Enzyme deficiency, Large tumors (e.g., mesenchymal tumors, epithelial tumors, endothelial tumors), Sepsis, Starvation and Artifact |
Symptoms of Hypoglycaemia
The symptoms of hypoglycaemia fall into two main groups (Cryer,
2001; Cryer et al., 2009) those related to
acute activation of the autonomic nervous system and those secondary to glucose
deprivation of the brain (neuroglycopenia). They are categorized as follows:
• | Autonomic (Sweating, Trembling, Pounding heart, Hunger, Anxiety) |
• | Neuroglycopenic (Confusion, Drowsiness, Speech, difficulty, Inability to concentrate, Incoordination) |
• | Non-specific (Nausea, Tiredness, Headache) |
Diabetic Ketoacidosis
Prior to the discovery of insulin more than 50% of diabetic patients died
in ketoacidosis (Eledrisi et al., 2006; Powers,
2005). Today this complication should account for less than 2% of deaths
among diabetics. However, both the incidence and the mortality rate are still
unfortunately high. Failure of patient to understand the disease and to appreciate
the significance of symptoms of poor control is the most common causes. Its
prevention is largely a matter of education of both patients and doctors. A
significant number of new patients still present in diabetic ketoacidosis and
in established diabetics a common course of events that patients may develop:
intercurrent infection, lose of their appetite and either stop or drastically
reduce their dose of insulin (on either their own initiative or their doctors
advice) by mistakenly belief that under these circumstances less insulin is
required. Any form of stress, particularly which produced by infection, may
precipitate severe ketoacidosis in even mildest case of diabetes.
Thyroid Disease
The thyroid is a small gland; shaped likes a butterfly that rests in the
middle of the lower neck (Fig. 1). Its primary function is
to control the body's metabolism (rate at which cells perform duties essential
to living). To control metabolism, the thyroid produces hormones, T4 and T3,
which tell the body's cells how much energy to use (Fatourechi,
2009; Villar et al., 2007).
A properly functioning thyroid will maintain the right amount of hormones needed to keep the body's metabolism functioning at a satisfactory rate. As the hormones are used, the thyroid creates replacements. The quantity of thyroid hormones in the bloodstream is monitored and controlled by the pituitary gland. When the pituitary gland, which is located in the center of the skull below the brain, senses either a lack of thyroid hormones or a high level of thyroid hormones, it will adjust its own hormone (TSH) and send it to the thyroid to tell it what to do. When the thyroid produces and releases more hormones than ones body needs, it is called Hyperthyroidism.
Parathyroid Disease
Parathyroid glands are small glands of the endocrine system which are located
in the neck behind the thyroid (Fig. 1). There are four parathyroid
glands which are normally the size and shape of a grain of rice (Eknoyan,
1995). Occasionally, they can be as large as a pea and still be normal.
Normal parathyroid glands are the color of spicy yellow mustard. Although, the
thyroid and parathyroid are neighbors and both are part of the endocrine system,
they are unrelated and do not have the same functions. Hyperparathyroidism is
the principle disease of parathyroid glands. It occurs when one of the parathyroids
develops a tumor which makes too much parathyroid hormone.
Expert System User Interface
Communication between the user and the expert system is done through the
user interface which was implemented in Arabic and English language to be easy
for the regular end user (Fig. 2). The user interface does
not require much tying.
When the user choose Clinical Examination for example, a new screen is displayed in the format of a dialogue, the expert systems ask a question and the user choose the best answer form the choices provided (Fig. 3).
Finally, the expert system informs the patient/user about the initial results of consultation of the phase one checkup (Fig. 4).
The expert system and the user as a second phase of checkup is started to determine a more accurate percentage that the user is diabetic (Fig. 5).
The final result of consultation of phase2 is displayed to the user (Fig. 6).
Fig. 2: | Expert system user interface |
Fig. 3: | General examination phase1 |
Fig. 4: | Initial results of consultation for phase 1 |
Fig. 5: | General examination phase 2 |
Fig. 6: | Results of consultation for phase 2 |
System Evaluation
In a preliminary evolution of the expert system, a few classical test cases
were used to test the expert system and the result of the system was accurate
when compared with the result of the Physicians; furthermore, some patients
having diabetes diseases tried this expert system, in order to evaluate it and
they were surprised by the accuracy of the diagnosis and treatment of the diabetes
diseases.
CONCLUSION AND FUTURE WORK
In this study, we have presented a medical expert system for Endocrine Diagnosis and Treatments. Even though, our expert system is similar to some previously implemented experts systems; but we managed to overcome the limitations they had. We have concentrated on three glands: Pancreas with diabetes disease, Thyroid with hyper and hypothyroidism diseases and parathyroid with hyper and hypoparathyroidism diseases. We used Java Expert System Shell (JESS) to perform its functions, facts and procedures. The JESS is a rule based engine for the Java language platform which is a superset of CLIPS programming language. The system can be improved to include more glands and to add more diseases to the system. We have tested the system by using classical diagnosed cases and checked if the systems results in agreement with the physicians diagnoses.
This expert system is considered to be a base of future ones; additional diseases and features like data recording of patients are planned to be added and to make it more accessible to users from anywhere at any time.
EXPERT SYSTEM SOURCE CODE