The processes of diagnosis and treatment of inflammatory bowel diseases are characterized by a high level of uncertainty in information about the causes of the disease, its etiology, the influence of external and internal factors on the patient's condition, the patient's individual response, and the standard treatment regimen. The paper presents the results of studies of the process of using intelligent agents of the monitoring information system to adapt the control influences of treatment regimens for Crohn's disease and ulcerative colitis to the mechanisms of individual interaction of processes occurring in the patient's body. As a result of joint research of scientists in the field of medicine and information technologies of Ukraine and Azerbaijan, a methodology for developing treatment regimens for inflammatory bowel diseases using the results of intellectual monitoring of the patient's condition was presented and experimentally tested. The paper formulates a number of hypotheses, for testing which experiments were carried out: an approach to the process of forming indicators to describe the patient's condition is described; available laboratory research methods are used; the problem of classification of patients' conditions is being solved; a method for non-invasive diagnostics of the influence of indicators on the patient's condition and interpretation of monitoring results has been developed. As a result of this work, experimental confirmation of the effectiveness of the process of using the results of intelligent monitoring in the practice of treating inflammatory bowel diseases was obtained.
Inflammatory bowel diseases (IBD), including ulcerative colitis (UC) and Crohn's disease
(CD), are an urgent problem in modern medicine. According to the WHO, the incidence of IBD is
growing worldwide. Despite the high level of health care, only in the United States - 1.5, and in the
European Union - 2 million patients [
The diagnosis of diseases is established on the basis of generally accepted criteria in
accordance with the recommendations of the European Crohn's and Colitis Organization (ECCO) [
At the same time, laboratory diagnostics, based mainly on taking into account only two parameters
(C-reactive protein and calprotectin), in our opinion, needs to be expanded and modernized.
Insufficient information content of the patient's test results does not allow using existing methods of
processing observation results and using them when choosing or constructing treatment regimens. In
such conditions, the effectiveness of treatment is largely determined by the doctor's intuition. Taking
this into account, we made an attempt to create a system for assessing the patient's condition on the
basis of other laboratory tests and indicators, building an information system for intelligent
monitoring [
Information technology of intelligent monitoring [
Each intelligent agent forms and adapts its structure following the local tasks that are assigned
to it. The main element of the agent structure is the model knowledge base [
A significant process for intelligent monitoring is the formation of a list of indicators that are used to
assess the patient's condition. For the formation of signs of the condition of patients with IBD, the
results of invasive studies are used, by processing images after endoscopy [
Today, for the diagnosis of IBD, protocols are used that take into account the results of a complete clinical, instrumental, laboratory, and pathomorphological examination of patients. At the same time, the main emphasis in instrumental studies is placed on radiation (CT, MRI, ultrasound) imaging methods and a thorough endoscopic examination with mandatory multiple biopsies of at least 5 sections of the intestine, and in laboratory diagnostics, in addition to routine studies and the exclusion of opportunistic infections, on the determination of calprotectin (a marker of damage to the intestinal mucosa) in feces and a marker of the acute phase of inflammation - "C" reactive protein (CRP) in the blood.
This approach, with undoubted advantages, has two significant drawbacks: the high cost of the
study and the remoteness of the timing of the diagnosis (waiting for the results of the
pathomorphological study). An important factor is also the patient's adherence to research (MRI, CT,
endoscopy). At the same time, laboratory diagnostics is based mainly on only two indicators: the
content of calprotectin in feces and CRP in the blood. In the literature, the possibility of using
additional methods of laboratory diagnostics is widely discussed [
The published materials contain the results of studies to describe the condition of patients at different stages of monitoring - expert determination of the list of signs, procedures for determining the characteristics of these signs, processing, and expert interpretation of the results. These results were used to diagnose the patient's condition. It was not possible to find a description of the use of the results of intelligent monitoring in the practice of treating IBD. 4. Aim
Investigation of the processes of using the results of intelligent monitoring to support decisionmaking in the practice of treating inflammatory bowel diseases using the example of the treatment of Crohn's disease and ulcerative colitis.
In the course of the research, several hypotheses were put forward.
Hypothesis 1. There are distinctive features that characterize patients with IBD, and MIS can be used in the process of constructing an individual treatment regimen for each of these patients.
Hypothesis 2. The body of each patient individually reacts to the onset of the disease and its course under the influence of drugs. The choice of a treatment regimen should depend on the priority factors that determine the patient's condition.
Hypothesis 3. Using the results of intelligent monitoring can improve the effectiveness of treatment.
A series of experiments were carried out to experimentally test the hypotheses put forward. Based on the research results, the monitoring information system was tasked with determining which processes and in which subsystems of the individual patient's body determine the condition of the patient with IBD.
The aim of the study is formalized as the task of supporting decision-making in the process of adapting treatment regimens by classifying the condition of patients based on the results of medical testing and determining the individual influence of factors. The list of classes of patient conditions, factors, and properties of patients was obtained by expert advice with the involvement of practicing doctors with scientific research experience, who are co-authors of this work.
The assessment of the condition of each patient was carried out especially for him by the intelligent agent MIS. The results of the analyzes were submitted to the agent's input in the form of a table of the input data array. The model knowledge base generated a conclusion about the patient's condition at a given time and the influence of the factors presented in Table 1 on the patient's condition. The assessments of the influence of factors were used by an expert physician as information about the individual characteristics of the patient in the process of adapting the treatment regimen. The effectiveness of using the results of intelligent monitoring was assessed according to the list of typical indicators before the application of the treatment regimen, built using assessments of the influence of factors, and after its application.
The input data array was formed based on the results of clinical studies. The list of indicators of the patient's condition was formulated expertly. When choosing the list of indicators, we proceeded from the fact that the basis of IBD is immunological, aseptic inflammation.
The term "endothelium" was proposed in 1865 to designate the lining of blood and lymphatic vessels, heart, serous, synovial and meninges, posterior chamber of the eye, respiratory tract. Currently, this term is used only to refer to the inner cell lining of the vascular bed. The endothelium, according to modern concepts, is the largest active endocrine organ in the human body, diffusely located in all organs and tissues.
The endothelium - the inner lining of blood vessels - consists of approximately 1-6 × 1013 cells. The endothelium of the vascular intima performs barrier, secretory, hemostatic, vasotonic functions, plays an important role in the processes of inflammation and remodeling of the vascular wall. Endothelial cells create a barrier between blood and tissues and, with the help of the factors they synthesize, perform many important regulatory functions, contributing to the maintenance of homeostasis. It is generally accepted that endothelial dysfunction (ED), as a typical pathological process, is a key link in the pathogenesis of many diseases and their complications, incl. with IBD. The inflammatory process in the intestinal mucosa, in particular leukocyte infiltration, contributes to damage to the vascular endothelium of the intestinal mucosa, causing a violation of microcirculation in it with the appearance of microthrombi and further trophic changes.
It is known that systemic endothelial dysfunction is reflected in damage to the wall of the
glomerular apparatus of the kidneys, which in turn leads to increased excretion of albumin in the
urine. It is believed that microalbuminuria (MAU) is an early marker for the development of
endothelial dysfunction [
To assess endothelial dysfunction, various parameters are currently being studied, such as
homocysteine, thrombocytosis, von Willebrand factor, endothelin, high-sensitivity CRP, changes in
the lipid spectrum and interleukin series, PAI-1, PAI-2, ICAM-1, NO, P- and E-selectins and many
others [
Our results allow us to conclude that the severity of endothelial dysfunction indicators directly correlates with the severity of the patient's condition.
Taking into account the fact that there is no significant difference between the groups of patients with UC and CD, we understand that these changes have a low level of specificity, and, therefore, can be applied only in cases of a previously established diagnosis.
We have selected the most accessible, both in practical terms (availability and availability in the laboratory network) and economically, to study indicators of endothelial dysfunction. The list of these indicators is presented in table 1.
The advantages of using this list of indicators include: 1) economic feasibility, 2) speed of calculation, 3) ease of implementation for the patient, 4) wide availability in outpatient practice.
To obtain the values of the indicators presented in Table 1, in the period from August 2015 to December 2018, 246 patients with IBD were examined at the clinical base of the Department of Therapy of AzSATI, the Department of Invasive Diagnostics and Treatment of the National Center of Oncology, the Medical Center "Memorial Klinika". Of this contingent, 44 people refused to participate in the study, and 19 people were excluded due to comorbid conditions (6 with arterial hypertension, 2 with chronic renal failure, 11 due to previous surgical interventions).
For the study, 183 patients were selected who had no complaints and anamnestic indications of cerebro-, cardio- and nephrovascular pathology. The diagnosis of the disease was established based on generally accepted criteria following the ECCO recommendations. The severity of the disease was assessed using the CDAI, HBI, and TW criteria.
The age of the patients was from 17 to 60 years (42.3 ± 2.7). By sex: 81 women and 102 men. The duration of the disease before going to a specialist doctor is 1.2-9.4 years (3.4 ± 1.1). 167 patients were examined on an outpatient basis, and 16 were on inpatient treatment. Of the patients, 104 (56.8%) suffered from CD, and 79 (43.1%) had UC. The patients were under dynamic observation from 9 to 36 months (14.2 ± 3.8). Patients, if necessary, underwent repeated examinations (426 in total). The results obtained in the study of the main group of patients are presented in table 2.
As can be seen from Table 2, in the total group of IBD patients, out of 426 studies conducted, 369 (86.6%) cases had an increased blood level of homocysteine, 405 (95.0%) - the level of h/s CRP, 322 (75,5%) - thrombocytosis, in 411 (96.4%) - a decrease in vitamin D content, in 308 (72.3%) albumin was found in urine, and in 411 (96.4%) - and increased content of calprotectin in feces. A separate analysis of the detection of each of these indicators in the UC and CD groups did not reveal any difference (p ˃ 0.05). Also, no difference was found when analyzing the results by gender (p ˃ 0.05).
In the control group K1 (patients diagnosed with irritable bowel syndrome, N = 20), an increase in homocysteine was detected in 3 cases, a decrease in the content of vitamin D, in 3 cases an increase in the level of h/s CRP and in 1 case a slight increase in calprotectin. In the control group K2 (healthy individuals, N = 20), in 3 cases, a decrease in the content of vitamin D was revealed and in 1 - a moderate increase in homocysteine. There was no statistically significant difference between groups K1 and K2. At the same time, the results of studies in both control groups significantly differed from those in the main (p ˂0.01). The same tendency persisted when the control groups were separately compared with the groups of UC and CD patients (p ˂0.01).
We carried out statistical processing of the data obtained to search for a possible relationship between the indicators of endothelial dysfunction and the severity of the patient's condition, determined following the ECCO recommendations. The severity of changes in the studied indicators of endothelial dysfunction was assessed as a percentage of the permissible value of the norm (with an increase in the indicator - it's excess of the upper limit of the norm and a decrease in comparison with the lower limit - with a lower content).
For simplicity of calculation, a point-based system for assessing the significance of these parameters was developed. The results are shown in Table 3. 59,8 21,0 19,1 22,0 51,5 26,5
As a result, it was found that in the presence of the severity of endothelial dysfunction, estimated from 4 to 6 points, patients with IBD have mild, from 6 to 9 points - medium, and if there are more than 9 points, a high degree of severity of the clinical course of the disease. Clinical and endoscopic remission corresponded to 3 points or less. The degree of correlation was 0.863.
Thus, hypothesis 1 received experimental confirmation. It has been proven that the list of indicators proposed as modeling variables in Table 1 are significant in determining IBD.
The individual condition of the patient was determined based on the results of solving the MIS
classification problem. The results of the expert classification of the patient's condition were used as a
modeled indicator. It was proposed [
macroscopic changes during endoscopy
0 50 500 1000
The severity of the disease as a whole was determined by expert judgment, taking into account the
severity of the current condition, the presence of extraintestinal manifestations and complications,
refractoriness to treatment, in particular, the development of hormonal dependence and resistance
[
frequency usual usual
The Norm Easy vulnerability Moderate Severe condition of (1 point on the vulnerability vulnerability the mucous Schroeder scale) (2 points on the (3 points on the membrane Schroeder scale) Schroeder scale)
infiltrate Availability is doubtful
The sum of the ratings determines the class of condition of the patient.: ≤4 remission; 5-6 - light attack; 7-8- medium-heavy; ≥ 9 – heavy
Table 8 shows a fragment of the input data array, built according to the results of observations and testing described above.
Factors were assessed according to standard MIS procedures. The structure of the agent of the
monitoring information system includes a model knowledge base and a model synthesizer [
Agent model synthesizers build of model knowledge base to classify the conditions of each
patient adaptively. The influence of the indicators presented in Table 1 was determined after
calculating the values of partial derivatives. Table 9 presents the influence of factors found in agent
models that were synthesized using the multi-line GMDH algorithm [
Table 9
The results presented in Table 9 make it possible to determine the prevalence of processes that determine the course of the disease in a patient whose name is encrypted in Tables 8 and 9. As a result of a detailed study of the data in Table 9, decisions were made on the individual correction of treatment regimens based on indicators that affect the course of the pathological process. For example, when correcting the treatment regimen for patient 245, it was taken into account that the prevailing factor at the time of diagnosis is an increase in the level of fecal calprotectin (influence 46%) with a concomitant increase in the level of h/s CRP (influence 35%) and an increase in the level of platelets (influence 12%). This means that in this patient, the course of the disease is determined by inflammatory processes in the intestinal mucosa, directly related to disturbances in microvascular hemodynamics, and as a consequence, correction is necessary not only in terms of escalating the dose of the basic drugs used, but also the correction of microvascular processes.
In patient 244, the prevailing factor at the time of diagnosis is also an increase in the level of fecal calprotectin (influence 62%), but with a concomitant increase in the level of homocysteine (influence 30%), which means that one of the triggers is deep intracellular hypoxia with a deficiency of folic acid and vitamin B12 and the correction of therapy in this patient must take these aspects into № account. In the majority of patients, one of the highly prevalent factors at the time of diagnosis, along with calprotectin, was h/s CRP, the drug correction of which significantly improved the treatment results in patients.
At 12 weeks of treatment, patients were asked to undergo re-examinations. Some of the patients refused to take a complete list of tests. Therefore, 59 patients remained in the control group. The results of assessing the condition of patients by intelligent agents, expertly confirmed by a doctor, after 12 weeks of treatment using adapted regimens are presented in Table 10.
correction of treatment regimens
The results of Table 10 suggest that the use of the results of intelligent monitoring of MIS agents increases the efficiency of treatment of patients with IBD.
At 12 weeks of treatment according to the adjusted scheme, 52 patients (88.0%) showed a significant improvement in well-being; at the same time, 10 (16.9%) patients showed remission, 34 (57.6%) patients showed a decrease in the severity of the disease course, and 8 (13.5%) patients showed subjective improvement in well-being was not accompanied by a noticeable improvement in the process activity indicators (false positive result). Another 4 (6.7%) patients did not notice an improvement in their condition, and 3 (5.1%) patients showed a deterioration. These patients were reexamined and it turned out that the patient under code 162 did not comply with the treatment period (he completely stopped taking one and reduced the dosage of other drugs). In patients under codes 131 and 160, opportunistic infections were detected (in 1- tuberculosis according to the quantiferon test, in 1 - herpes viruses HSV6 and EBV, detected by PCR of biopsies of the gastrointestinal mucosa). That is, 74.7% of patients achieved a positive treatment result.
Thus, we obtained experimental confirmation of hypothesis 2 and 3. It has been proved that taking into account the individual reactions of the patient's body when constructing a treatment regimen and using the results of intellectual monitoring increases the effectiveness of treatment.
The process of supporting decision-making by a doctor in the process of treating inflammatory bowel diseases consists in providing an intelligent agent with information about the patient's condition, the influence of well-known factors, and the prognosis of the results of the use of adapted treatment regimens.
The hypothesis of the existence of signs that characterize patients with IBD has been experimentally confirmed, and MIS can be used in the process of constructing an individual treatment regimen for each of these patients.
The source of increasing the effectiveness of the treatment of inflammatory bowel diseases in a patient is the correction of his treatment regimen, taking into account the priority factors that determine the patient's condition. As a result of combining methods for solving intellectual problems by agents of the monitoring intellectual system and correcting the methods of treating inflammatory bowel diseases used by a doctor, a systemic effect was obtained in the form of improving the health of patients.