The problem of diagnosis and treatment of acute pancreatitis with destructive changes of the pancreas and its various complications is the most complex and controversial in surgical gastroenterology, which determines the need for further improvement of both early diagnosis and general surgical tactics of patient management.
The aim of the study was to improve the results of surgical treatment of acute pancreatitis by implementing scientifically justified criteria of stratification of the course of the disease, and development of the new methods of its treatment.
The total pool of the current study was 225 patients, among which 126 were retrospective and 99 were prospective. The group of patients who have been hospitalized within the first day from the onset of the disease consisted of 74 people.
At the second stage of the study it was determined that on the first day of acute pancreatitis, the following indicators are related to the severity of the disease: the C (A/M) pulse wave amplitude indicator, which is determined by laser Doppler flowmetry; clinical and visual signs of effusion in the pleural and/or abdominal cavity; signs of systemic inflammatory response syndrome; indicator of the hematocrit number.
Based on the results of the analysis of the single and complex relationship of indicators of general clinical, laboratory, instrumental and special research methods, with the help of machine learning technology, an own system of criteria for assessing the severity of the condition of patients with acute pancreatitis was designed by the method of building a classification and regression tree, or an acceptance tree solutions Taking into account the indicators included in the system of prognostic criteria, it was named PanHELS (Pancreatitis, Hematocrit, Effusion, LDF, SIRS). A simplified system called PanHES (Pancreatitis, Hematocrit, Effusion, SIRS) was created according to a similar technique for prospective use by centers without existing equipment for determining microcirculation.
It was found that for the definition of acute severe pancreatitis, the PanHELS and PanHES criteria systems had 100% sensitivity and 100% specificity during the first day after admission, which in a comparative analysis was the indicator of the highest efficiency (BISAP - 90% sensitivity and 85.71% specificity, HAPS - 100% sensitivity and 76.19% specificity, Panc3 - 60% sensitivity and 85.71% specificity). For the definition of moderate acute pancreatitis, the PanHELS and PanHES criteria systems had 100% sensitivity and 43.48% specificity within the first day after admission, and for the definition of mild acute pancreatitis, PanHELS and PanHES had 84.62% and 76.92% sensitivity and 94.44% specificity, respectively, with a tendency to overdiagnosis. In comparison with the systems proposed today, the new criteria proved to be no less effective, and sometimes more effective, when used in patients with mild and moderate acute pancreatitis during the first day after admission.
At the third stage of the study, archived medical records of 21 patients were selected, who made up the comparison group, and who were treated according to the standard scheme with open surgical intervention. Along with this, a research group of 17 patients was formed, who were treated according to the proposed method. After surgical treatment of patients, in the early postoperative period, the clinical condition was monitored, in particular, the following indicators were determined: the total length of stay in the hospital, the length of hospital treatment from the moment of the first surgical intervention, complications and mortality. It was found that with the use of an improved method of surgical treatment of acute severe pancreatitis, the mortality rate from the disease decreased from 42.9% to 17.7%, the average length of stay in inpatient treatment decreased from 60 to 47.8 bed days, partial or full recovery became possible avoiding such complications as intestinal fistula and erosive bleeding.
For the first time, the impact of perfusion dysfunction on the course of acute pancreatitis in the first hours from the onset of the disease was determined and scientifically substantiated.
For the first time, the optimal zone for studying the state of microcirculation in patients with acute pancreatitis was determined and scientifically substantiated.
Based on the analysis of the isolated and quantitative influence of clinical and laboratory indicators and indicators of microcirculation, for the first time, with the help of machine learning technology, a prognostic regression and classification tree of the severity of acute pancreatitis was developed, which takes into account the state of the microcirculatory bed.
