Mapping Ultrasound Features of Hand Osteoarthritis: Links to Classifications, Clinical Signs, and Self-Reported Outcomes

Abstract

Background Hand osteoarthritis (OA) is a heterogeneous condition characterised by structural joint damage, low-grade inflammation, and considerable variability in clinical expression. Ultrasound (US) imaging provides sensitive modality for detecting early OA features, particularly inflammatory changes; however, the interplay between US-detected abnormalities, systemic inflammatory activity, clinical signs, and self-reported symptoms is not fully understood. This thesis aimed to: (1) map the anatomical distribution of US-detected features in hand OA and (2) examine their associations with hand OA classifications, clinical signs, and self-reported outcomes. Additionally, a systematic review of published literature was conducted to evaluate the relationship between US features and laboratory markers of inflammation. Methods This study was divided into two parts. The first was a systematic review of published manuscripts. A systematic literature search was performed in MEDLINE, EMBASE, CINAHL, and Web of Science (inception–June 2025) to identify observational studies assessing associations between US-detected features and laboratory inflammatory biomarkers in hand OA. Data were extracted on study characteristics, US features, biomarkers, and reported association measures. Due to substantial heterogeneity in study design, US scoring methods, biomarker panels, and statistical approaches, findings were synthesised narratively rather than meta-analytically. Study quality was appraised using the Newcastle–Ottawa Scale (NOS). The second was an observational study involving 100 participants recruited from primary care. Bilateral US assessment of DIP, PIP, MCP and 1st CMC joints was performed to identify osteophytes, synovial hypertrophy (SH), synovial effusion (SE), power Doppler (PD), and superb microvascular imaging (SMI) vascularity. Clinical signs were assessed physically (finger nodes, tenderness, thumb-base squaring) and photographically (finger nodes, deformity, thumb-base squaring). Pain and aesthetic impact were measured using visual analogue scales (VAS), and functional limitation using the Functional Index of Hand Osteoarthritis (FIHOA). Participants were classified according to erosive, nodal, radiographic, and symptom status. Joint-level frequencies were estimated. Associations between US features and clinical signs were examined using multilevel mixed-effects logistic regression, with multivariable linear regression for self-reported outcomes and multivariate regression for hand OA classifications. Results Out of 5,128 citations, four studies (546 participants, 91.75% female, mean age 56.1–66.3 years) scored >5 on the NOS. Within erosive hand OA (EHOA), one study reported positive correlations (r = 0.30–0.57) between serum inflammatory markers (including TNF, MIP-β, PDGF-bb, and IP-10) and grey scale synovitis (GSS); the other three studies did not assess the same biomarker panel, so this finding was not replicated across studies. No significant correlations were observed between other US-detected features (e.g., PD signals, osteophytes, effusion, cartilage thickness) and inflammatory biomarkers, with coefficients generally <0.2. In the observational study, US abnormalities followed the classical OA pattern, with highest involvement at the 2nd–3rd DIPJs, 1st PIPJs, and 1st CMCJs. Osteophytes were the most prevalent feature. Nodal OA demonstrated the strongest and most consistent association with overall US burden across all domains, whereas erosive OA (EOA) exhibited feature-specific relationships (Verbruggen erosions with osteophyte; OARSI erosions with SH). Radiographic and symptomatic classifications showed no independent associations. Clinical signs were primarily associated with osteophytes and, to a lesser extent, SH. Participant-level symptoms demonstrated weak and inconsistent relationships with US features, with osteophytes emerging as the only independent predictor of functional limitation. In explanatory models, tenderness was the strongest predictor of pain, function, and aesthetic concern, while US added no additional explanatory value beyond clinical and radiographic assessment. Conclusions US provides anatomical and phenotypic depiction of hand OA, highlighting structural burden as the dominant driver of clinical presentation. Nodal status is a key marker of global US-defined disease severity, while inflammatory biomarkers show limited correspondence with US features. These findings support a model in which biomechanical loading shapes both structural and inflammatory expression in hand OA and emphasise the need for longitudinal and mechanistically oriented research to refine phenotyping and target early disease processes.