A machine learning approach to knee osteoarthritis phenotyping: data from the FNIH Biomarkers Consortium

Abstract

Objective: Knee osteoarthritis (KOA) is a heterogeneous condition representing a variety of potentially distinct phenotypes. The purpose of this study was to apply innovative machine learning approaches to KOA phenotyping in order to define progression phenotypes that are potentially more responsive to interventions. Design: We used publicly available data from the Foundation for the National Institutes of Health (FNIH) osteoarthritis (OA) Biomarkers Consortium, where radiographic (medial joint space narrowing of 0.7 mm), and pain progression (increase of 9 Western Ontario and McMaster Universities Osteoarthritis Index [WOMAC] points) were defined at 48 months, as four mutually exclusive outcome groups (none, both, pain only, radiographic only), along with an extensive set of covariates. We applied distance weighted discrimination (DWD), direction-projection-permutation (DiProPerm) testing, and clustering methods to focus on the contrast (z-scores) between those progressing by both criteria ("progressors") and those progressing by neither ("non-progressors"). Results: Using all observations (597 individuals, 59% women, mean age 62 years and BMI 31 kg/m 2) and all 73 baseline variables available in the dataset, there was a clear separation among progressors and non-progressors (z ¼ 10.1). Higher z-scores were seen for the magnetic resonance imaging (MRI)-based variables than for demographic/clinical variables or biochemical markers. Baseline variables with the greatest contribution to non-progression at 48 months included WOMAC pain, lateral meniscal extrusion, and serum N-terminal pro-peptide of collagen IIA (PIIANP), while those contributing to progression included bone marrow lesions, osteophytes, medial meniscal extrusion, and urine C-terminal crosslinked telopeptide type II collagen (CTX-II). Conclusions: Using methods that provide a way to assess numerous variables of different types and scalings simultaneously in relation to an outcome of interest enabled a data-driven approach that identified key variables associated with a progression phenotype.

The Combination of an Inflammatory Peripheral Blood Gene Expression and Imaging Biomarkers Enhance Prediction of Radiographic Progression in Knee Osteoarthritis.

“…The nal baseline model included uCTXII and sNTXI and had an AUC of 0.586. Similar to the unbiased causal analysis of various biomarkers (Fig.2) in this study, Loeser and associates have also identi ed through in an unbiased machine learning approach also identi ed BML, osteophytes, and medial meniscal extrusion as potential biomarkers in identifying radiographic (>0.7mm at 48 months) and pain progressors in the OAI cohort (44). Dunn et al have studied the peripheral blood methylation status in the radiographic progressors relative to non-progressors in a small cohort of OAI subjects.…”

“…One of the limitations of these ndings is that none of the biomarkers studied, alone or in combination, predicted symptomatic worsening at 2 years -a time frame chosen to represent a feasible time period for clinical trials. In part, this may be due to the limited period of observation, as compared, for example, to the 48 month period of follow-up in the FNIH study (44). We note that BML and cartilage readings are semi-quantitative, and it is possible that with more advanced scoring systems, precise evaluation of BML and cartilage volume or size would further improve the progression prediction.…”

Differential correlation network analysis identified novel metabolomics signatures for non-responders to total joint replacement in primary osteoarthritis patients

“…Network science is a powerful tool to investigate the relationships and interaction patterns of a set of entities and has seen many applications to biomedical research as these types of analyses do not require a priori hypotheses and can thus be used to identify datadriven subgroups in complex disease that otherwise may not have been considered due to the large number of factors often considered in these types of analyses and lack of known relationship between factors. Such data driven methods have been applied to OA research by different research groups (Hu et al 2018;Nelson et al 2019). In this study, we applied a differential correlation network analysis method (Hu et al 2018) to the same dataset of our previous study (Costello et al 2019) to identify further metabolic markers and pathways for non-responders to TJR.…”

Effectiveness of drugs with delayed-release structurally modified action depending on the phenotype of osteoarthritis

“…Активные формы кислорода, индуцированные ишемией хряща и субхондральной кости вследствие нарушения сосудистых паттернов, способствуют активации патологического внутриклеточного каскада ядерного фактора каппа B [19,20,21], который, в свою очередь, запускает транскрипцию генов MMPs, цитокинов, хемокинов, приводящих к деградации внеклеточного матрикса [22,23]. Более того, активное свободнорадикальное окисление способствует повреждению ДНК и формированию патологического фенотипа хондроцитов [24,25].…”