The case-based reasoning method has a high potential for solving tasks of intelligence decision-support. To implement it, it is necessary to solve the problem of comparing situations and selecting the one that is most similar to the current situation in the knowledge base. The problem arises in the case of heterogeneous objects and situations with many different types of parameters and their possible uncertainty. In this paper, an approach based on machine (deep) learning is investigated for this task. It is proposed to carry out the process of selecting situations and solutions from the knowledge base in two stages: recognition of the states of the elements of a complex object and the relationships between them, then the formation of a representation of the situation in the state space and its use for comparing situations using neural networks. An ensemble neural network model based on a multi-layer network is proposed. It successfully simulates the cognitive functions of a human (expert), correctly selects similar situations and ranks them according to the similarity parameter. Proposed neural network models provide the implementation of a hybrid-CBR approach for decision-making on complex objects.

Participants were asked to wear their Fitbit TM daily during waking hours (removing for water-based activities), and sync the device at least every three days during the study period of 90 days. We downloaded step count data from each participant's Fitbit TM using the product interface (www. fitbit.com). For each day of the 90-day period, a minimum of 1500 steps was needed to be recorded by the device for that day to be considered a valid day. Average daily step count was presented as mean (±SD). Participants were classified as physically inactive (5000-7499 steps/day), moderately active (7500-9999 steps/day), physical active (10,000 steps/day), and very active (12,500 steps/day) (Tudor-Locke & Basset, Sports Med, 2004). The associations between step count and health-and knee-related quality of life were evaluated with Pearson correlation coefficients (r). Results: Thirty-two participants with PF OA were included (Table 1). Participants reported moderate to severe knee-related symptoms, with the greatest limitations reported on the pain and function in activities of daily living subscales of KOOS. The mean (±SD) daily step count was 9660±3455 (range 5148 to 17380), and the average number of valid days was 87±6 (range, 67 to 90). 34% of participants were classified as physically inactive, 25% were moderately active, 22% were physically active, and Values represent median (range). g¼grams, IU¼international units, kcal¼kilocalories, mg¼milligrams, n¼number of participants, mg¼micrograms Abstracts / Osteoarthritis and Cartilage 28 (2020) S86eS527 S391

Peculiarities of the morphology of some phenotypes of experimental osteoarthrosis was studied in experiments on rats. Reorganization of the knee articular cartilage of Wistar rats during aging (age-associated phenotype), obesity (metabolic phenotype), circulatory disturbances (e.g., chronic heart failure), and their combinations (polymorbidity) was studied by hematoxylin and eosin staining, immunohistochemical staining for collagen II and caspase 3, and morphometry. High sensitivity of the cartilage to non-traumatic influence of different anthropomorphic factors was demonstrated; morphological changes in osteoarthrosis of different genesis. The most pronounced pathological changes were observed in polymorbid animals, which allowed developing a new pathogenetically substantiated model of nontraumatic osteoarthrosis.

Background:Osteoarthritis (OA) relevance is determined by its record prevalence with progredient growth throughout the world [1]. Clinical and pathogenic heterogeneity of disease actualizes problem of its stratification [2]. Lack of unified understanding of OA and its phenotype determination results in incredible number of attempts to group OA, using of different classification criteria in last decade.Objectives:To analyze and systematize available OA classifications, proposals and phenotypes, to highlight the most promising of them.Methods:We studied publications from MEDLINE / PubMed and Google Scholar databases found by the keywords “osteoarthritis”, “phenotypes”, “subphenotypes”, “classification”, “subtypes”, “subsets”, “subgroups”, “subpopulations”, “profiles” and “endotypes” in various combinations in English and Russian. We did not set a time frame, but aimed to include as many different methods as possible in order to reflect evolution of scientists’ views on structuring of this disease.Results:A total of 55 OA grouping methods were covered so that OA was structured by different determinants into 6 big boxes.First OA classifications were characterized by complex etiopathogenetic approach, while subsequent studies differed in joint-mediated approach, and the knee joint was undisputed “champion” in this “race”. One of the first attempts to group OA was division into primary, or idiopathic, and secondary, due to known causes. It is now obvious that it is becoming obsolete, and criteria for OA primacy are difficult to determining. Genomic highly specialized studies based on isolation of “favorable” and “unfavorable” genes develops prerequisites to genetic OA classifying. Clinical variants occupy central place as they are the most fully consistent with modern phenotype conception [3], considerating as subtypes of disease shared by underlying pathobiological and pain mechanisms and their structural and functional consequences. Trajectories of OA progression are distinguished by longitudinal design, that is, the determinants for grouping here are disease characteristics in dynamics. The ancestor of structural OA trajectories can be considered Kellgren-Lawrence grades; subsequent studies identified complex of clinical, laboratory and morphological factors contributing to development of trajectories. Structural OA variants are diverse depending on visualization methods, and many of them can be naturally considered phenotypes, since they drive certain clinical OA manifestations. Morphological changes were described at macro- and microscopic levels; it is interesting to note the absence of histopathological norm in patients without OA. Laboratory profiles of patients are determined by content of systemic (serum, urinary) or local, “proximal” (in synovial fluid) biomarkers, which seem to be more precise. Metabolomic analysis is perspective new direction of laboratory studies based on joint metabolic products identification in the synovial fluid. New trend in OA research is molecular phenotyping. The specific molecular pathway explaining observed phenotype properties is called “endotype”. Endotype is related to certain pathobiological scenario, and laboratory markers are potentially effective for its diagnosis.Conclusion:Thus, a large amount of accumulated information and its diversity soon will probably lead to qualitatively new knowledge level with deep understanding of phenotype-associated strategy for managing OA patients.References:[1]Wallace IJ, Worthington S, Felson DT, et al. Knee osteoarthritis has doubled in prevalence since the mid-20th century. Proc Natl Acad Sci USA. 2017 Aug 29;114(35): 9332-9336. doi: 10.1073/pnas.1703856114 Epub 2017 Aug 14.[2]Deveza LA, Nelson AE, Loeser RF. Phenotypes of osteoarthritis: current state and future implications. Clin Exp Rheumatol 37 Suppl 2019;120(5):64-72.[3]Van Spil WE, Bierma-Zeinstra SMA, Deveza LA, et al. A consensus-based framework for conducting and reporting osteoarthritis phenotype research. Arthritis Res Ther. 2020;22(1):54. doi:10.1186/s13075-020-2143-0Disclosure of Interests:None declared.