Background/Purpose: While musculoskeletal ultrasound examination (MSUS) is used increasingly in the work-up of patients with inflammatory joint disease, the exact diagnostic utility has not been established. We sought to apply a prospective, probabilistic (Bayesian) approach to quantify the diagnostic utility of MSUS.
Methods: All patients referred to our clinic for evaluation of arthritis were eligible, unless a prior diagnosis was indicated or a certain diagnosis could be made based on the information in the referral letter. Patients were assessed by history and physical examination including joint examination, laboratory testing including acute-phase reactants, RF, and ACPA, and plain x-ray of hands, wrists and feet if clinically indicated. A diagnostic assessment was then performed by the responsible physician where the probability of a) any inflammatory joint disease and b) rheumatoid arthritis was given on a 5-point scale ranging from unlikely (0-20% probability) to very likely (80-100% probability). Subsequently, an ultrasound examination of the wrist, MCP, PIP 2-5 in both hands, MTP 2-5 in both feet and also symptomatic joints was performed by HR and the results of the examination presented to the responsible physician. The latter then assessed the diagnostic probabilities again, using the same scale. The proportions of patients with maximal and minimal diagnostic certainty pre-test and post-test were compared by Fisher exact test.
Results: 67 patients were included, 52 were female, average (SD) age 47.9 (16.5) years. Symptom duration 8.9 (4.0) months, 15 patients were positive for RF and 12 for ACPA. The pre-test and post-test probability distributions for (any) inflammatory joint disease and for RA are given in the table. The final diagnoses in these patients were RA (15), other inflammatory joint diseases (21), non-inflammatory joint disease (31). With regard to a diagnosis of (any) inflammatory joint disease, the proportion of patient for whom diagnostic certainty was maximal (<20% OR >80% likelihood) was 20/67 (29.9%) before MSUS and 42/67 (62.7%) after MSUS (p=0.0002). With regard to a diagnosis of RA, the proportions were 21/67 (31.3%) pre-test and 39/67 (58.2%) post-test (p=0.003). Parallel reductions were seen in the proportions of patients with greatest diagnostic uncertainty (40-60% likelihood), from 24/67 (35.8%) to 10/67 (14.9%) (p=0.0093) and from 17/67 (25.3%) to 4/67 (6.0%) (p=0.0035), respectively.
|
|
any inflammatory disease (n patients in group) |
rheumatoid arthritis (n patients in group) |
||
|
Diagnostic probability |
Pre-test |
Post-test |
Pre-test |
Post-test |
|
<20% |
4 |
14 |
12 |
27 |
|
20-40% |
14 |
10 |
26 |
17 |
|
40-60% |
24 |
10 |
17 |
4 |
|
60-80% |
9 |
5 |
3 |
7 |
|
>80% |
16 |
28 |
9 |
12 |
Conclusion: In this probabilistic (Bayesian) analysis, musculoskeletal ultrasound when added to routine physical and laboratory examination greatly increased the diagnostic certainty in patients referred for the evaluation of arthritis.
Disclosure:
H. Rezaei,
None;
E. af Klint,
None;
Y. Kisten,
None;
R. F. van Vollenhoven,
AbbVie, BMS, GSK, Merck, Pfizer, Roche, UCB,
2,
AbbVie, AstraZeneca, Biotest, BMS, GSK, Lilly, Merck, Pfizer, Roche, UCB, Vertex,
5.
« Back to 2013 ACR/ARHP Annual Meeting
ACR Meeting Abstracts - https://acrabstracts.org/abstract/the-diagnostic-utility-of-musculoskeletal-ultrasound-in-early-arthritis-a-probabilistic-bayesian-approach/