AUCTORES
Globalize your Research
Review Article | DOI: https://doi.org/10.31579/2768-0487/053
*Corresponding Author: U N Musevi, Azerbaijan State University of Oil and Industry, Baku AZ1010, Azerbaijan.
Citation: U N Musevi. (2021). Differential Diagnosis of Parasitic Diseases of the Gastrointestinal Tract Using Artificial Neural Networks. Journal of Clinical and Laboratory Research. 3(5); DOI:10.31579/2768-0487/053
Copyright: © 2021 U N Musevi. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Received: 25 September 2021 | Accepted: 21 October 2021 | Published: 26 October 2021
Keywords: gastrointestinal tract; parasites; symptoms; neural network; structure; testing; prediction; error
Disorders of the functional state of the gastrointestinal tract associated with the influence of various parasites are considered. The symptoms of diseases caused by parasites and their location in the gastrointestinal tract are given. The possibility of using neural network technology in diagnosing illnesses as a result of the influence of various parasites is estimated. The structure of the neural network is given, indicating the set of inputs and outputs, as well as the result of its training. For the created neural network, test results for the respective symptoms and disease prediction results for these symptoms were obtained.
Often, a violation of the functional state of the gastrointestinal tract (GIT) is associated with the influence of various parasites. Parasites have a more complex structure and have well-oiled defense mechanisms directed against the human immune system (encapsulation, antigenic mimicry, antigenic "drift," inactivation of enzymes and biologically active substances, etc.), which allows them to exist for a long time in various human organs and tissues ... Also, there are objective difficulties in identifying, isolating, and obtaining immunoreagent specific antigens of parasites.
Therefore, for example, the immune response in giardiasis is for the most part due not to the surface proteins of the parasite, but to antigens that enter the human body along with the products of their vital activity. Thus, in the laboratory diagnosis of many pests, serological research methods are only of auxiliary value [1].
The World Health Organization has proven that 95% of humanity has a variety of parasites in the body. These living organisms are not as harmless and safe as it might seem in the first place. Most of them are localized in the organs of the gastrointestinal tract (the eggs of the worms get here along with contaminated water and food). Still, there are also so-called extra intestinal forms of invasion - parasites can live in the lungs, heart and even the human brain [2].
Parasites weaken the immune system, lowering the release of immunoglobulin, and their presence constantly stimulates the system's response and, over time, can loosen this vital immune mechanism, opening the way for bacterial and viral infections to enter the body.
These symptoms are just a few of them. In reality, the symptoms of diseases caused by parasites in the digestive tract are more extensive. The most challenging thing about this is that these symptoms of different infections are very close and require additional techniques to clarify the diagnosis.
Probable etiological factors of gastrointestinal tract dysfunction are mainly parasites: Entamoeba, Giardia lamblia, Balantidium colitis, Ascaris lumbricoides, Enterobius vermicularis, Taenia solium (saginata), Strongyloides stercoralis, Cryptosporidium parvum [3].
Colitis is an inflammation of the colon, observed in several diseases, namely, in chronic inflammatory bowel diseases, pseudomembranous colitis, and infections caused by bacteria, parasitic protozoa (amoeba) and viruses. Irritable bowel syndrome, otherwise called mucosal or spastic colitis, is not associated with inflammation of the colon, although it has similar symptoms [13].
The causative agent of giardiasis in humans is Lamblia intestinalis (Giardia intestinalis, Giardia lamblia). Giardiasis is an adequately widespread invasion throughout the world that affects all age groups, but children suffer from this disease more often than others [14].
Balantidium colitis is a type of ciliates parasitizing in the large intestine of some mammals: as a rule, in pigs, less often in rats, dogs, and also in humans. It causes a disease called balantidiasis or ciliated dysentery [6].
Ascariasis - intestinal invasion from the group of nematodes, the causative agents of which are roundworms (Ascaris lumbricoides). Ascaris parasitizes in the small intestine [7].
Enterobiasis is an intestinal invasion by the pinworm Enterobius vermicularis, usually found in children [15].
Cysticercosis is the most common parasitic disease of the central nervous system. The invasion of the central nervous system by the larvae of the pork tapeworm Taenia solium occurs when eating food contaminated with helminth eggs [16].
Strongyloidiasis is an invasion caused by Strongyloides stercoralis [15].
Cryptosporidiosis is a parasitic disease caused by protists of the genus Cryptosporidium from the Apicomlex type. Cryptosporidiosis, as a rule, manifests itself as an acute and short-term infection and is spread by the nutritive route, often through contaminated water [17].
Echinostoma infects the gastrointestinal tract in humans and can cause a disease known as echinostomes [18].
A neural network model for predicting gastrointestinal diseases caused by parasites
Artificial neural networks are effectively used in the diagnosis of various diseases [19, 20]. Neural network technologies are also used for the diagnostic of diseases of the gastrointestinal tract, for example, for the differential diagnosis of liver diseases [21] and in predicting the development of abdominal sepsis in patients with severe acute pancreatitis [22, 23].
The experiment was carried out on a NeuroPro network emulator. NeuroPro0.25 beta version allows you to implement the following basic operations:
- the creation of neuroprojects;
- connect data files with a neuroproject;
- adding layer architectures to neural projects from 1 to 10 layers, with up to 100 neurons in each;
- train a neural network to solve forecasting and classification problems;
- testing of a neural network based on database files, calculating the significant indicators of input signals;
- simplify the neural network;
- selection of learning algorithms, determination of forecasting for a given accuracy, etc.
A neural network that determines the prognosis of diseases using the symptoms of diseases.
For the experiment, we select the symptoms of various gastrointestinal forgetfulness, progressing with parasites. Twenty-four indications were selected (at the request of doctors, the number of symptoms can be increased, since these systems are open) and nine diseases (it should be noted that the number of illnesses created by parasites is quite large, the most common of them were selected) (table 2). For the experiment, a neural network simulator NeuroPro 0.25, was used.
The input parameters of the neural network are the symptoms shown in Table 2, the set of inflows includes 9 varieties, and the output of the net system will be the solution of the neural network according to the training rules. Figure 1 shows a neural network corresponding to the first variant of the experiment.
After displaying the input and output parameters of the network, the network is trained.
The analysis shows that the most optimal algorithm for learning a multilayer perceptron is the back propagation algorithm [24].
To test the created neural network, selected symptoms were chosen (Table 5).
Table 6 shows the results of a neural network for predicting diseases by symptoms.
Thus, an efficient type of structure of an artificial neural network designed to solve problems of medical diagnostics and prognosis is a perceptron with sigmoid activation functions, the input of which is information about the symptoms of a patient's diseases, and the output is a diagnosis of the disease. According to the results obtained by the neural network, it is possible to confidently clarify the illness that corresponds to the "Echinostomosis" output, created by parasites of the small intestine, which as a result, leads to a violation of the functional state of the gastrointestinal tract.
Clearly Auctoresonline and particularly Psychology and Mental Health Care Journal is dedicated to improving health care services for individuals and populations. The editorial boards' ability to efficiently recognize and share the global importance of health literacy with a variety of stakeholders. Auctoresonline publishing platform can be used to facilitate of optimal client-based services and should be added to health care professionals' repertoire of evidence-based health care resources.
