A FEASIBILITY STUDY ON A NOVEL COMBINED THERMAL IMAGING AND CLINICAL JOINT ASSESSMENT APPROACH USING ULTRASOUND DETECTED JOINT INFLAMMATION OUTCOMES IN RHEUMATOID ARTHRITIS

Background: Thermal imaging (TI) is a portable, low cost imaging tool with high feasibility for use. Clinical joint assessment. Is routinely performed in rheumatoid arthritis (RA) patient care. Objectives: To assess a combined TI and clinical joint assessment (CTCA) approach in comparison with TI alone using ultrasound (US) detected joint inflammation outcomes as a gold standard. Methods: Bilateral (BL) hand and wrist (22 joint sites) were assessed in this cross-sectional study. For TI (performed in a draft free room with a controlled temperature of around 22°C), the adjusted maximum (Tmax), minimum (Tmin) and average (Tavg) temperatures were derived by subtracting a control temperature (lowest Tmin at the joints per subject) from the Tmax, Tmin and Tavg per joint. US power Doppler (PD) and greyscale (GS) joint inflammation were graded semi-quantitatively (0-3) using validated scoring methods. Joint swelling and tenderness were graded as yes = 1 or no = 0. To increase the relative weightage of CTCA-MAX, CTCA-MIN and CTCA-AVG on the CTCA scores, if the joint was swollen and/or tender, the adjusted Tmax, Tmin and Tavg at each joint were multiplied by a factor of 2; otherwise, they remained unchanged. Receiver operating characteristic (ROC) analysis assessed the performance of TI and CTCA in identifying joints with US PD score > 1 and GS score > 1. A parameter was selected as a univariate predictor if statistically significant (P < 0.05) with area under the ROC curve (AUC) ≥ 0.70. Results: This study included 814 joints from 37 RA patients (mean disease duration, 30.9 months; mean DAS28, 4.43). For both TI and CTCA, out of the 22 joints sites, 3 joint sites were evaluated for PD score > 1 and 14 joint sites for GS score > 1; the remaining joint sites had AUC results unavailable due to small number of outcomes. For TI ( Table 1 ), 3 joint sites had ≥ 1 predictive parameter for either PD score > 1 and/or GS score > 1 as follows: left (L) wrist and right (R) MCPJ 1, AUCs (0.813 to 0.897) for PD score > 1; L wrist and R MCPJs 1 and 3, AUCs (0.808 to 0.947) for GS score > 1. For CTCA ( Table 1 ), 6 joint sites had ≥ 1 predictive parameter for either PD score > 1 and/or GS score > 1 as follows: BL wrists, AUCs (0.726 to 0.899) for PD score > 1; BL wrists, MCPJs 2 and 3, AUCs (0.739 to 0.931) for GS score > 1. Table 1. Identifying joints with ultrasound PD score >1 & GS score >1 Thermal Imaging alone CTCA Joint UScriterion Parameter (AUC ≥ 0.7& P <0.05) AUC(95% CI) Cut-off Joint UScriterion Parameter (AUC ≥ 0.7& P <0.05) AUC (95% CI) Cut-off L R L R L PD score >1 Adjusted Tmax **0.841 (0.691, 0.992) 4.7 L & R PD score >1 CTCA-MAX **0.899 (0.797, 1) **0.776 (0.578, .973) 9.4 7.3 Wrist Adjusted Tmin **0.813 (0.669, 0.958) 2.85 Wrist CTCA-MIN **0.861 (0.735, 0.987) *0.726 5.7 4.45 (0.526, 0.926) Adjusted Tavg **0.849 (0.714, 0.985) 3.9 CTCA-AVG **0.889 (0.781, 0.997) *0.761 7.3 5.95 (0.563, 0.959) GS score >1 Adjusted Tmax **0.827 (0.687, 0.966) 4.7 GS score >1 CTCA-MAX **0.918 (0.833, 1) **0.813 8 7.3 (0.632, 0.994) Adjusted Tmin **0.808 (0.67, 0.947) 2.85 CTCA-MIN **0.873 (0.761, 0.986) **0.766 4.4 4.45 (0.581, 0.951) Adjusted Tavg **0.837 (0.707, 0.967) 3.9 CTCA-AVG **0.913 **0.802 5.5 5.95 (0.824, 1) (0.62, 0.985) R PD score >1 Adjusted Tmax *0.897 (0.726, 1) 5.7 L & R GS score >1 CTCA-MAX - *0.758 - 9.8 (0.494, 1) MCPJ 1 MCPJ 2 GS score >1 Adjusted Tmax *0.936 (0.813, 1) 7.2 CTCA-MIN *0.902 *0.739 2.75 3.9 (0.775, 1) (0.443, 1) Adjusted Tmin *0.932 (0.793, 1) 3.95 CTCA-AVG *0.931 **0.763 4.7 5.5 (0.835, 1) (0.474, 1) Adjusted Tavg *0.947 (0.868, 1) 4.9 L & R GS score >1 CTCA-MAX *0.914 *0.873 6.35 12.2 (0.735, 1) (0.617, 1) R GS score >1 Adjusted Tmax *0.922 (0.76, 1) 4.6 MCPJ 3 CTCA-MIN - *0.902 - 3.15 (0.75, 1) MCPJ 3 CTCA-AVG - *0.902 - 4.1 (0.728, 1) Corresponding P-value: statistically significance at *P <0.05, **P<0.01. Conclusion: A novel CTCA approach helps discriminate the severity of US detected joint inflammation in RA at more joint sites when compared to TI alone; this includes the commonly affected BL wrists, MCPJs 2 and 3. Further validation work in a larger RA cohort will be required. Disclosure of Interests: None declared Citation: , volume 81, supplement 1, year 2022, page 1770Session: Diagnostics and imaging procedures (Publication Only)