Advertisement

Diagnostic test accuracy of magnetic resonance imaging and ultrasound for detecting bone erosion in patients with rheumatoid arthritis

  • Haozheng Tang
  • Xinhua QuEmail author
  • Bing YueEmail author
Original Article
  • 45 Downloads

Abstract

Objective

To evaluate and compare the diagnostic test accuracy of magnetic resonance imaging (MRI) and ultrasound (US) for bone erosion in rheumatoid arthritis (RA) patients for a specific and efficient diagnostic recommendation.

Method

To evaluate the diagnostic accuracy, the sensitivity, specificity, area under the summary receiver operating characteristic curve, positive likelihood ratio, negative likelihood ratio, and diagnostic odds ratio of MRI and US for detecting bone erosion were calculated. Subgroup analyses were conducted to evaluate the performance of these values with different standard references when compared with types of machines and scanning positions.

Results

Data from 26 articles were extracted for calculation. The comprehensive values of sensitivity and specificity were 0.77 (95% CI 0.63, 0.87)/0.89 (95% CI 0.80, 0.95) and 0.61 (95% CI 0.43, 0.77)/0.95 (95% CI 0.88, 0.98) for MRI and US, respectively. The 1.5-T Signa MRI system, General Electric© (sensitivity 0.66; specificity 0.90), and different models of LOGIQ US units and General Electric© (sensitivity 0.66; specificity 0.91) had better diagnostic capability to detect bone erosion, while the 2nd metacarpophalangeal joint (sensitivity 0.70; specificity 0.98) showed the best diagnostic performance among the hand joints with US.

Conclusions

Neither MRI nor US showed satisfactory diagnostic test accuracy in detecting bone erosion. However, the 1.5-T Signa MRI system, General Electric©, and different models of LOGIQ US units and General Electric© showed similarly good performance in detecting bone erosion in RA patients, while the 2nd metacarpophalangeal joint is the best recommended scanning position during US.

Key Points:

• In this study, we evaluated the diagnostic accuracy of US and MRI for bone erosion in RA patients, neither MRI nor US showed perfect diagnostic test accuracy.

• 1.5-T Signa system and the LOGIQ units both from General Electric© are the machine types of MRI and US with the greatest performance, respectively.

• The 2nd MCP joint is the scanning position recommended during US test.

• Different reference standards will greatly influence the judgment of the results.

Keywords

Bone Erosion Diagnosis Magnetic resonance imaging Rheumatoid arthritis Ultrasound 

Abbreviations

RA

Rheumatoid arthritis

MRI

Magnetic resonance imaging

US

Ultrasound

ACR

American College of Rheumatology

EULAR

European League Against Rheumatism

CR

Computed radiography

CT

Computed tomography

TP

True positivity

FN

False negativity

FP

False positivity

TN

True negativity

ROC

The receiver operating characteristic curves

AUC

The area under the curve of summary ROC

PLR

Positive likelihood ratio

NLR

Negative likelihood ratio

DOR

Diagnostic odds ratio

MCP

Metacarpophalangeal

PIP

Proximal phalangeal

DIP

Distal interphalangeal

MTP

Metatarsophalangeal

Introduction

Rheumatoid arthritis (RA) is a chronic persistent systemic disease that primarily affects the small joints, causing articular synovitis and pain, and impairs patients’ daily activities [1]. Previous studies have maintained that early diagnosis and treatment can provide more control over structural changes and better long-term prognosis [2]. It is known that synovitis and bone erosion are the two pivotal early imaging findings in RA patients. While the diagnostic accuracy of imaging methods for detecting synovitis in RA patients is widely researched, the efficacy of bone erosion detection lacks rigorous evidence.

The American College of Rheumatology (ACR) and European League Against Rheumatism (EULAR) have reported and updated widely accepted criteria for clinical diagnosis of RA [3]. The imaging techniques proven to enable bone erosion detection in RA are computed radiography (CR), computed tomography (CT), magnetic resonance imaging (MRI), and ultrasound (US). Currently, CR has been phased out for its limited capability for early bone erosion diagnosis in clinical practice; CT has been criticized for its risk of radiation exposure [4]. There is a need for more accurate, efficient, and convenient imaging methods to diagnose bone erosion in RA.

An authoritative RA diagnostic score system using MRI has already been proposed and updated by OMERACT in 2015, which authenticates the high diagnostic capability of MRI [5]. Nevertheless, based on published literature, especially for studying bone erosion detection, the diagnostic accuracy of MRI still differs notably, as the values of sensitivity and specificity vary from 0.28 to 1.00 and from 0.31 to 0.99, respectively; besides, there still are several limitations of MRI that cannot be overlooked, like its high cost, an always long waiting list, and a low penetration rate in remote areas or some developing countries [6, 7].

US is a common, non-invasive, and cost-effective imaging modality. Its ability to diagnose synovitis in RA has been proven; moreover, the use of US to detect bone erosion in RA patients has been receiving increasing attention among researchers [8]. However, similar to the varied diagnostic performance of MRI, the values of sensitivity (from 0.19 to 1.00) and specificity (from 0.44 to 1.00) of US in bone erosion diagnosis seen in different studies are far from consistent [6, 9]. Additional studies are required to evaluate its reliability and validity systematically.

In this study, we calculated the diagnostic test accuracy of these two methods separately and compared the results to identify their diagnostic capability for bone erosion. We aimed to discover a reliable method or even a specific type of machine or position for precise diagnosis of bone erosion in RA patients, which will be helpful to the establishment of a more accurate and effective imaging strategy for detecting RA in the future.

Materials and methods

Overview

This meta-analysis was performed according to the Cochrane Collaboration guidelines [10]. Both prospective and retrospective studies were included if they provided data for both sensitivity and specificity of one or more methods for the detection of bone erosion in RA patients. All the articles retrieved were published between January 2000 and September 2018. Due to the study-based nature of this study, ethical approval and patient consent were not deemed applicable.

Study selection

To obtain studies from January 2000 to September 2018, we systematically searched the database of PubMed and EMBASE. The search string used in this study was “‘Magnetic Resonance Imaging’ AND ‘Ultrasonography’ AND ‘Arthritis, Rheumatoid’ AND ‘bone erosion’” referring to MeSH terminology.

Previously published reviews and original studies included were physically re-evaluated. Researchers (H.T., X.Q.) screened and reviewed all the articles extracted independently, articles retrieved by the independent reviewers were scrutinized through full-text reading, and the controversial results were resolved by a third reviewer (B.Y.). Final inclusion was decided after resolving discrepancies among all the investigators.

The inclusion criteria for the studies were as follows: (1) studies that provided data on sensitivity and specificity of MRI and US from adults patients with confirmed RA diagnosis or others who were diagnosed during follow-up treatment; (2) studies that provided sufficient data to calculate values of true-positive (TP), false-negative (FN), false-positive (FP), and true-negative (TN) with more than 10 patients included in the study; (3) studies that reported clear gold standard chosen as reference for different kinds of imaging methods. When the same group of patients was suspiciously included in different articles, we evaluated the patient characteristics in all these articles and finally selected the most detailed article.

The exclusion criteria for the studies were as follows: (1) studies that reported non-RA disease; (2) studies without extractive data of diagnostic test accuracy like TP, FP, FN, and TN; (3) studies that provided equivocal gold standard or reference diagnostic method; and (4) non-English studies with inextricable data.

Data extraction and study quality assessment

One of the main investigators (H.T.) collected the data, using predetermined criteria, including study design (prospective or retrospective studies, blinded, gold standards, and endpoints), patient characteristics (number of target patients and healthy group, mean age, mean disease duration, numbers of the joints or quadrants, target scanning joints), and imaging parameters (position undergoing scan, MRI field, US array transducer, machine types, including brand and companies). The sensitivity and specificity of each study were also extracted for calculating the TP, FP, FN, and TN. The QUADAS tool was used to evaluate the quality of the articles in this research; after discussing with all researchers, all studies were found to have moderate qualities [11].

Primary outcomes

The primary outcomes were diagnostic test accuracy of MRI and US for bone erosion in patients defined using other different imaging methods or diagnosed using the 2010 ACR/EULAR classification criteria or the 1987 ACR criteria for RA [12]. When studies reported more than two subgroups (divided according to different joints or other basic characteristics), we included all groups and evaluated them separately using subgroup analysis.

Index and reference detection

The main index detections in this study were MRI and US in any mode. Determining a definite gold standard for RA detection among imaging methods is still an academic dilemma [13] although MRI has shown good performance in detecting synovitis and erosion [5]. Therefore, we ignored the boundaries of different reference standards for a comprehensive result as far as possible; we also compared the accuracy of US and MRI with a same standard reference (CR, CT, or ACR Criteria) for a further subgroup analysis.

Statistical analysis and heterogeneity assessment

This study was performed using the STATA SE version 15.1, and the R software was used for additional statistical calculations.

One reviewer (H.T.) calculated the TP, FP, FN, and TN values using the relevant data extracted from studies included previously. Sensitivity and specificity of MRI and US for detection of bone erosion were calculated to evaluate the diagnostic accuracy to differentiate RA from other forms of early arthritis or unidentifiable joint disease. The receiver operating characteristic curves (ROC) and the area under the summary receiver operating characteristic curve (AUC) were both calculated. We also compared the positive likelihood ratio (PLR), negative likelihood ratio (NLR), and diagnostic odds ratio (DOR) among different detective methods [14].

Conventionally, heterogeneity was tested using Cochran’s Q-test, and the results were considered to have significant heterogeneity when I2 > 50% [15]. The efficacy of MRI and US for detecting bone erosion was compared with the gold standard chosen in each study by calculating the sensitivity and specificity with the 95% CI. Thereafter, the random effects model was used to evaluate the value of diagnostic accuracy when the result showed heterogeneity (I2 > 50%) [15]. We also divided the original sample into different subgroups based on mean age, disease duration, the disparities of machine types, different scanning positions, for heterogeneity analysis and evaluated the heterogeneity using the forest plot analysis. Fagan’s nomograms were depicted for demonstrating post-test probability. Publication bias was estimated using Deek’s funnel plot asymmetry test [14].

Results

Study flow and characteristics

The search and review process of the included studies are demonstrated in Fig. 1. After the first-round selection, 663 relative abstracts in total were extracted by two researchers. Following full-text scrutiny, only 24 studies were retained after the second-round review. Articles were discarded for the following reasons: design not eligible for determining diagnostic accuracy (n = 468); population information unavailable like in juvenile RA patients or others (n = 90); no clear gold standard or reference (n = 56); not in English (n = 21); and suspiciously similar individual samples (n = 4). After including two additional studies through manual reviewing, the final analysis of diagnostic accuracy of bone erosion in RA was performed on 26 articles in total (Table 1); 15 articles on MRI and 15 on US techniques (Fig. 1).
Fig. 1

Flow diagram for study selection

Table 1

Characteristics of the 26 studies in meta-analysis for the diagnosis of bone erosion using US and MRI

Study

Year

Country

Participant number

Mean age

Joint/Quadrants

Disease durationb (years)

Joints

Method

Ref standard

Study design

QUADAS score

Wakefield et al. [16]

2000

UK

120

53.7

NAc

2.5

MCP

US

CR/ACR Criteria

Retrospective

12

Lindegaard et al. [17]

2001

Denmark

28

55

700

0.3

MCP/wrist

E-MRI

CR

Retrospective

11

Alarcon et al. [18]

2002

UK

15

45

28

NAd

MCP/MTP

US/MRI

CR

Retrospective

10

Lopez-Ben et al. [19]

2004

USA

15

45

120

NA

MCP/MTP

US

ACR Criteria

Retrospective

11

Szkudlarek et al. [20]

2004

Denmark

60

56

1200

2

MCP/MTP

US

MRI

Retrospective

12

Ejbjerg et al. [21]

2005

Denmark

65

52

1431

52

MCP/wrist

LF-MRI

HF-MRI

Retrospective

12

Bajaj et al. [22]

2006

USA

26

42.6

168

1

MCP/PIP/MTP

US

ACR Criteria

Prospective

11

Yao et al. [23]

2006

USA

11

48

165

0.5

Wrist

MRI

CT

Retrospective

12

Døhn et al. [24]

2006

Denmarks

21

52

336

8

MCP

US/MRI

CT

Retrospective

11

Szkudlarek et al. [25]

2006

Denmark

60

58

1832

5

MCP/PIP

US

MRI

Retrospective

12

Døhn et al. [7]

2007

Denmark

21

52

84

8

MCP

MRI

CT

Retrospective

12

Freeston et al. [26]

2007

UK

15

56.5

86

11

MCP/wrist

LF-MRI

HF-MRI

Retrospective

10

Duer-Jensen et al. [4]

2008

Denmark

25

55

550

9

MCP/wrist

MRI

CT

Retrospective

12

Døhn et al. [27]

2008

Denmark

21

51

315

7

Wrist

MRI

CT

Retrospective

11

Duer et al. [28]

2008

Denmark

41

55

NA

1.5

MCP

MRI

ACR criteria

Prospective

12

Funck-Brentano et al. [29]

2009

Denmark

114

50.3

682

0.3

MCP/MTP

US

CR

Prospective

11

Døhn et al. [30]

2010

Denmark

52

61

832

7

MCP/wrist

US/MRI

CT

Retrospective

12

Rahmani et al. [31]

2010

Iran

12

45.3

216

0.3

MCP/PIP

US

MRI

Retrospective

11

Tamai et al. [32]

2012

Japan

51

52

1501

0.5

MCP/PIP/wrist

MRI

Gd-DTPA MRI

Retrospective

11

Kawashiri et al. [6]

2012

Japan

69

53.6

NA

0.3

MCP/PIP

US/MRI

ACR criteria

Prospective

12

Døhn et al. [33]

2012

Denmark

49

61

784

7

MCP

US

CT

Retrospective

11

Amin et al. [34]

2012

Egypt

50

45.6

50

6

Shoulders

US

MRI

Prospective

12

Aoki et al. [35]

2013

Japan

41

59

NA

NA

MCP/PIP/DIP/Wrist

Gd-DTPA MRI

ACR criteria

Prospective

12

Ji et al. [36]

2014

China

31

51.1

NA

1

MCP/wrist

Gd-DTPA MRI

ACR criteria

Prospective

11

Peluso et al. [9]

2015

Italy

20

52

220

< 1

MCP/PIP/wrist

3D-US

CT

Prospective

12

Ji et al. [37]

2017

China

94

56.7

NA

1

MCP/PIP/wrist

US

ACR criteria

Prospective

12

Ref standard, reference standard; NA, not available; LF-MRI, low-field MRI; HF-MRI, high-field MRI; ACR, American College of Rheumatology; Gd-DTPA MRI, gadolinium-diethylenetriaminepentacetate-enhanced MRI; MCP, metacarpophalangeal joint; PIP, proximal phalangeal joint; MTP, metatarsophalangeal joint; DIP, distal interphalangeal joint

aDisease (symptom) duration reported a mean value or a broad range from different original study

bStudies with unavailable number of joints provided the number of patients for calculation

cStudies not reporting precise number of mean RA duration provided whether the RA patients they studied were early diagnosis, instead

As shown in Table 1, we evaluated 16 studies from Europe (12 from Denmark [4, 7, 17, 20, 21, 24, 25, 27, 28, 29, 30, 33], 3 from the UK [16, 18, 26], and 1 from Italy [9]), 7 studies based on Asians (3 from Japan [6, 32, 35], 2 from China [36, 37], and 1 each from Egypt [34] and Iran [31]) and the remaining 3 from the USA [19, 22, 23]. Among 26 studies, 25 had calculated the data in hand joints, including the metacarpophalangeal (MCP) joints, proximal phalangeal (PIP) joints, and distal interphalangeal (DIP) joints, while 1 study had focused on the shoulder joint. Thirteen studies included patients with short disease or symptom duration (< 3 years), while patients with a mean age over 60 years were only found in 2 studies. All the studies were published between January 2000 and September 2018 and patients or members of the public were involved in their studies voluntarily. Only one study was published as a letter; all the other studies were published as original articles.

Diagnostic test accuracy of MRI and US in bone erosion detection

The comprehensive values of the diagnostic test accuracy of MRI and US were calculated separately.

MRI

There were 15 studies with 507 individuals in total. The calculated sensitivity and specificity of MRI were 0.77 (95% CI 0.63, 0.87; I2 = 93.27%) and 0.89 (95% CI 0.80, 0.95; I2 = 94.74%), respectively. Based on the PLR value of 7.2 and NLR of 0.25, the positive and negative MRI results suggested a moderate to slight change in erosion probability separately; and the negative post-test probability was 6% and positive post-test probability 97% in MRI (Online Resource 1a). Further, an AUC of 0.91 (95% CI 0.88, 0.93) and DOR of 28 (95% CI 12, 67) indicated that the MRI had good diagnostic test accuracy as a diagnostic method for RA joints (Fig. 2a).
Fig. 2

Summary receiver operating characteristic curves and forest plot. In the ROC graph, the diamond represents the distribution of different studies, the black solid line is the pooled receiver operating characteristic curves, the dash line represents 95% confidence contour, the dot line represents 95% prediction contour, and the black diamond indicates the pooled effect estimates. In the forest plot graph, the black horizontal bars represent 95% confidence intervals and the black dots indicate the pooled effect estimates. a Magnetic resonance imaging. b Ultrasound

US

Among the 15 studies, including 1476 patients and healthy volunteers, the PLR was 11.2 (95% CI 5.7, 22.2) indicating a high increase in the positive results of RA bone erosion when patients were detected using US; and the negative post-test probability of US was 9% and positive post-test probability 98% (Online Resource 1b). The calculated comprehensive sensitivity and specificity of US were 0.61 (95% CI 0.43, 0.77; I2 = 93.66%) and 0.95 (95% CI 0.88, 0.98; I2 = 94.93%), respectively; besides, the DOR was 27 (95% CI 14, 55) and AUC was 0.90 (95% CI 0.87,0.92). This indicated that US has modest diagnostic efficacy when used as a screening imaging test for RA bone erosion, although its efficacy is slightly lower than that of MRI (Fig. 2b).

Diagnostic test accuracy with different standard references

In order to analyze the diagnostic test accuracy of MRI and US when comparing with different standard references (CR, CT, and ACR Criteria), we also calculated the synthetic results with separate subgroup analyses; the results showed unstable trends in both two methods. When compared with CR, both MRI and US have excellent performance in bone erosion detection, with the sensitivity and specificity being 0.99 (95% CI 0.50, 1.00) and 0.85 (95% CI 0.69, 0.93) in MRI and 0.82 (95% CI 0.49, 0.95) and 0.84 (95% CI 0.64, 0.94) in US, respectively. The sensitivity and specificity of MRI, when compared with CT, is 0.66 (95% CI 0.59, 0.73) and 0.92 (95% CI 0.89, 0.95), respectively, while same values are 0.44 (95% CI 0.41, 0.47) and 0.94 (95% CI 0.93, 0.95) in US, respectively. Furthermore, when the ACR clinical diagnosis criteria were used as a reference, the sensitivity and specificity are 0.87 (95% CI 0.63, 0.96) and 0.62 (95% CI 0.41, 0.80) in MRI and 0.44 (95% CI 0.24, 0.65) and 0.97 (95% CI 0.35, 1.00) in US (Table 2), respectively.
Table 2

Diagnostic value of the subgroup analysis for the diagnosis of bone erosion using US and MRI with different standard reference

 

Number of studies

SEN (95% CI)

SPE (95% CI)

AUC (95% CI)

DOR (95% CI)

CR as standard reference

MRI

By joints

2

0.99 (0.50, 1.00)

0.85 (0.69, 0.93)

0.99 (0.97, 0.99)

752 (6, 87847)

Ultrasound

3

0.82 (0.49, 0.95)

0.84 (0.64, 0.94)

0.90 (0.87, 0.92)

23 (8, 70)

CT as standard reference

MRI

By joints

3

0.66 (0.59, 0.73)

0.92 (0.89, 0.95)

0.88 (0.85, 0.90)

23 (17, 31)

Ultrasound

6

0.44 (0.41, 0.47)

0.94 (0.93, 0.95)

0.67 (0.62, 0.71)

13 (10, 16)

ACR criteria as standard reference

MRI

By patients

3

0.87 (0.63, 0.96)

0.62 (0.41, 0.80)

0.81 (0.77, 0.84)

11 (4, 31)

Ultrasound

4

0.44 (0.24, 0.65)

0.97 (0.35, 1.00)

0.77 (0.73, 0.81)

26 (0, 1642)

CR, computed radiography; CT, computed tomography; ACR, American College of Rheumatology; SEN, sensitivity; SPE, specificity; AUC, area under the ROC curve; DOR, diagnostic odds ratio

Diagnostic value with different types of imaging machines

Based on the limited available information from the original studies, we selected three types of each of the methods. Among three types of MRI machines, the 1.5-T Signa MRI system (sensitivity 0.77; specificity 0.93) from General Electric© showed the best performance, while the other two types, 0.6T Panorama, Philips© (sensitivity 0.66; specificity 0.90) and 0.2T Artoscan, ESAOTE Biomedica© (sensitivity 0.66; specificity 0.91), showed lower, similar value of diagnostic test accuracy. Regarding US machine, 6 studies used different models of LOGIQ units, General Electric©, for imaging detection; this machine has the best performance, with the sensitivity being 0.71 (95% CI 0.50, 0.86) and the specificity being 0.94 (95% CI 0.84, 0.98). In comparison, the other two types of US machine, the HDI 3000 unit from Advanced Technology Laboratories (sensitivity 0.66; specificity 0.94) and Esaote US unit (sensitivity 0.55; specificity 0.89), had unsatisfactory values of diagnostic test accuracy (Table 3).
Table 3

Subgroup analysis for the diagnosis of bone erosion using US and MRI with different machines

 

Number of studies

SEN (95% CI)

SPE (95% CI)

AUC (95% CI)

PLR (95% CI)

NLR (95% CI)

DOR (95% CI)

Magnetic resonance imaging

Overall study

15

0.77 (0.63, 0.87)

0.89 (0.80, 0.95)

0.91 (0.88, 0.93)

7.2 (3.9, 13.5)

0.25 (0.15, 0.42)

28 (12, 67)

Machine type

1.5T Signa, GE

4

0.77 (0.41, 0.94)

0.93 (0.70, 0.99)

0.93 (0.90, 0.95)

10.4 (2.3, 47.8)

0.25 (0.08, 0.82)

42 (6317)

0.6T Panorama, Philips

4

0.66 (0.62, 0.70)

0.90 (0.75, 0.96)

0.68 (0.63, 0.71)

6.5 (2.5, 17.3)

0.38 (0.32, 0.44)

17 (6, 52)

0.2T Artoscan, Esaote

4

0.66 (0.26,0.92)

0.91 (0.85, 0.95)

0.92 (0.89, 0.94)

7.3 (3.1, 17.4)

0.37 (0.12, 1.17)

20 (3137)

Ultrasound

Overall study

15

0.61 (0.43, 0.77)

0.95 (0.88, 0.98)

0.90 (0.87, 0.92)

11.2 (5.7, 22.2)

0.41 (0.27, 0.62)

27 (14, 55)

Machine type

LOGIQ, GE

6

0.71 (0.50, 0.86)

0.94 (0.84, 0.98)

0.91 (0.89, 0.94)

11.7 (4.6, 30.3)

0.31 (0.17, 0.57)

38 (13, 107)

HDI 3000

4

0.66 (0.16, 0.95)

0.94 (0.54, 1.00)

0.92 (0.89, 0.94)

11.9 (1.6, 87.1)

0.36 (0.09, 1.49)

33 (6185)

Esaote

2

0.55 (0.39, 0.70)

0.89 (0.86, 0.91)

0.89 (0.86, 0.91)

5.1 (3.5, 7.3)

0.50 (0.35, 0.72)

10 (5, 20)

SEN: sensitivity; SPE: specificity; AUC: area under the ROC curve; PLR: positive likelihood ratio; NLR: negative likelihood ratio; DOR: diagnostic odds ratio; LOGIQ, GE: General Electric’s LOGIQ ultrasound unit for imaging detection; HDI 3000: HDI 3000(5–10 MHz, Advanced Technology Laboratories, Bothell, WA); Esaote: Esaote ultrasound unit from Genova, Italy; 0.2 T Artoscan, Esaote: a 0.2 T Artoscan system (ESAOTE Biomedica, Italy); 0.6 T Panorama, Philips: Philips Panorama with 0.6 T unit (Philips Medical Systems, Helsinki, Finland); 1.5 T Signa, GE: 1.5 T General Electric’s Signa system

Diagnostic value of US in different hand joints when compared with MRI

US test of the small joints in the hand, such as the MCP, PIP, and DIP, is convenient and can offer an intuitionistic vision of lesions in RA patients. Based on the information gained from two studies that reported specific information about the US test for different MCP joints compared with MRI, we evaluated the ability of US to test bone erosion in the 2nd to 5th MCP joints, which are the most commonly affected joints. According to the results shown in Table 4, we found that all four MCP joints have an AUC over 90% and the 2nd MCP joint showed the best performance with a sensitivity of 0.70 (95% CI 0.59, 0.80) and a specificity of 0.98 (95% CI 0.97, 0.99) (Table 4).
Table 4

Detection of bone erosion in different MCP joints with US using MRI as reference standard

 

Number of studies

SEN (95% CI)

SPE (95% CI)

AUC (95% CI)

DOR (95% CI)

2nd MCP

2

0.70 (0.59, 0.80)

0.98 (0.97, 0.99)

0.97 (0.95, 0.98)

132 (56, 311)

3rd MCP

2

0.50 (0.39, 0.61)

0.98 (0.96, 0.99)

0.93 (0.90, 0.95)

41 (20, 85)

4th MCP

2

0.60 (0.42, 0.76)

0.99 (0.98, 1.00)

0.98 (0.96, 0.99)

182 (54, 621)

5th MCP

2

0.66 (0.45, 0.82)

0.98 (0.96, 0.99)

0.99 (0.97, 0.99)

121 (39, 370)

MCP, metacarpophalangeal joints; SEN, sensitivity; SPE, specificity; AUC, area under the ROC curve; DOR, diagnostic odds ratio

Diagnostic value in different research strategies and patients with different stages

As shown in Online Resource 2, the diagnostic accuracy in the prospective study subgroup was much lower in both MRI and US, which reinforces the comprehensive results that the diagnostic accuracy of MRI and US is unsatisfying. Moreover, the diagnostic capability of MRI between not only short-term (< 3 years) and long-term (≥ 3 years) RA disease, but also younger (< 50) and elder (≥ 50) patients were both similar, while the diagnostic capability of US seems unsatisfying in elder patients group. The results were listed in Online Resource 3.

Publication bias

We used Deek’s test to assess the publication bias; MRI (p = 0.13) and US (p = 0.93) had no evidence of publication bias based on the results we calculated (Online Resource 4).

Discussion

RA is a chronic autoimmune disease characterized by synovitis and bone erosion in small, non-weight-bearing joints; it has an overwhelming, painful, and agonizing long-term course in patients who are not diagnosed and treated in time [38]. As a result, the systematic review of accuracy and application of different methods of imagological RA diagnosis is a significant study. The OMERACT group [5] has performed extensive research on assessing radiographic outcomes using MRI in RA, while Takase-Minegishi et al. [8] conducted a convincing research on synovitis imaging detection in RA patients. Synovitis and bone erosion are typical imaging symptoms. Bone erosions are not pathognomonic only for RA (overlapping with osteoarthritis and gout); however, active erosive lesions also need to be detected on US/MRI imaging in RA. Hassan et al. [39] have reported a systematic review on diagnosis accuracy of US in detecting RA that provided only cursory values of synthetic sensitivity and specificity without a systematic exploration of the diagnostic capabilities of US.

Therefore, to validate the existing evidences and collect additional information, we conducted a comprehensive and systematical analysis for diagnostic test the accuracy of MRI and US in diagnosis RA-related bone erosion. We used the random effects model to evaluate the data from all 26 studies, calculated synthetic sensitivity and specificity of MRI and US as well as the values of AUC using ROC. In addition, we conducted a series of subgroup analyses to evaluate the sources of heterogeneity and to explore more potential information.

According to our findings, we determined that even if MRI showed a slightly better diagnostic ability than US, both methods showed only limited diagnostic test accuracy, which is confirmed by a prospective-study subgroup analysis (Fig 2 and Online Resource 2). However, based on the results of the subgroup analyses, we obtained considerable information that may have guiding significance for clinical practice. First, the subgroup analyses based on different standard references showed that although MRI showed more satisfactory diagnostic performance than US in all three groups, which is consistent with existing evidence and the main comprehensive conclusions mentioned above, MRI and US results showed significant differences among the three subgroups. When CR was selected as a reference, both MRI and US showed excellent bone erosion detection capability. When MRI and US were compared with CT separately, the specificity of each modality was high while the sensitivity was far from satisfying. Finally, if the ACR criteria were used as a standard reference, MRI had a high value of sensitivity but a lower value of specificity while the values of US are opposite (Table 2). This result showed that using different references as the standard methods has a significant impact on the outcome judgment and diagnostic performance evaluation. Therefore, we suggest that, as more eligible and available data may be reported in the future, further studies should use the same gold standard to evaluate the diagnostic test accuracy of MRI and US for bone erosion in RA. Second, subgroup analysis results for different machine types showed that the 1.5-T Signa MRI system and the different models of LOGIQ US units from the General Electric Company may have better diagnostic performance; besides, the sensitivity and specificity of these two types had similar values despite the imaging methods being different (Table 3). Therefore, it would be irresponsible to report a conclusion only relying on a classification based on different imaging methods that MRI has better diagnosis test accuracy of bone erosion in RA than US. To solve this problem, more specific information about the parameters of each methods should be summarized accurately and these evidences should be evaluated using a more scientific strategy. Moreover, as US cannot scan the patient’s limbs in multiple dimensions like MRI, the diagnostic efficacy of musculoskeletal US is closely related to the scanning plane and position specialists selected. Therefore, we compared the different hand joints for US scanning. The results suggested that the 2nd MCP joint, which is also the joint most commonly involved in early RA, is the finger joint with the greatest diagnostic test accuracy (Table 4). As shown in Online Resource 3, there is no obvious difference in the diagnostic performance of both MRI and US between early and long-term RA patients, which means the length of the disease duration may not significantly interfere the MRI and US diagnosis of bone erosion in RA; between younger (< 50 years) and elder patients (≥ 50 years), US showed better diagnostic performance in the former. The interferon may come from the bone lesion caused by other degenerative diseases like osteoarthritis, though more details remain unknown. These results provide a possible reference for the subsequent development of a diagnostic strategy or scoring system for RA.

The limitations of this study are as follows. One, the available data is insufficient for comparison with a similar gold standard; this requires more supporting evidence from further studies. Similarly, differences among various types of equipment and their specific parameters should also be considered. As this is a study-based research, it is difficult to explain the mechanism of different types of equipment at the physics/biomechanics and biotechnology level; this would require experts in these specific fields to offer their professional explanation. Furthermore, although MRI showed a better diagnostic performance, its high price and low penetration rate are the limitations that prevent the development of MRI; on the contrary, although US is more convenient and time-saving, its overall diagnostic accuracy in bone erosion is still far from satisfactory. Therefore, in future studies, researchers should focus on discovering a not only accurate but also convenient imaging method to replace both MRI and US as a method for efficiently diagnosing bone erosion or other lesions in RA.

In conclusion, according to the comprehensive results, neither MRI nor US showed satisfactory diagnostic accuracy in the diagnosis of bone erosion in RA. However, according to the results of subgroup analysis, the 1.5-T Signa MRI system, General Electric©, and different models of LOGIQ US units, General Electric©, showed similar good results in detecting RA bone erosion, while the 2nd MCP joint is the best recommended scanning position during US test. Different standard references tend to significantly influence the judgment of the results. Further rigorous clinical researches and convincing evidences of imaging diagnosis of RA are required to further knowledge about this topic.

Notes

Acknowledgments

The authors would like to thank the authors of all the studies included in this research.

Funding information

This work was supported by National Natural Science Foundation of China [81472119 and 81672196], Shanghai “Rising Stars of Medical Talent” Youth Development Program (Youth Medical Talents – Specialist Program), and Shanghai municipal education commission-Gaofeng clinical medicine grant support [20161423].

Compliance with ethical standards

Disclosures

Nones.

Supplementary material

10067_2019_4825_Fig3_ESM.png (1.9 mb)
Online Resource 1:

Likelihood ratio plot gram and Fagan’s nomogram. The solid diamond represents the pooled positive likelihood ratio (LRP) & negative likelihood ratio (LRN) for index text with 95% confidence intervals in Likelihood ratio plot; and in Fagan’s nomogram, the solid arrow represents the relationship between pre- and post-test probability of LRP, while the dash arrow represents the LRN. A: magnetic resonance imaging; B: ultrasound (PNG 90 kb)

10067_2019_4825_MOESM1_ESM.tif (2.7 mb)
High resolution image (TIF 2737 kb)
10067_2019_4825_MOESM2_ESM.docx (0 kb)
Online Resource 2: Diagnostic value of the subgroup analysis for the diagnosis of bone erosion using US and MRI in different research strategies. (DOCX 14 kb)
10067_2019_4825_MOESM3_ESM.docx (0 kb)
Online Resource 3: Diagnostic value of the subgroup analysis for the diagnosis of bone erosion using US and MRI at different stages. (DOCX 16 kb)
10067_2019_4825_Fig4_ESM.png (455 kb)
Online Resource 4:

Funnel plots for the included studies. The solid line is the regression line and the circles represent the studies included. A: magnetic resonance imaging; B: ultrasound (PNG 18 kb)

10067_2019_4825_MOESM4_ESM.tif (677 kb)
High resolution image (TIF 676 kb)

References

  1. 1.
    Scott DL, Wolfe F, Huizinga TW (2010) Rheumatoid arthritis. Lancet 376(9746):1094–1108.  https://doi.org/10.1016/S0140-6736(10)60826-4 CrossRefPubMedGoogle Scholar
  2. 2.
    van Nies JA, de Jong Z, van der Helm-van Mil AH, Knevel R, Le Cessie S, Huizinga TW (2010) Improved treatment strategies reduce the increased mortality risk in early RA patients. Rheumatology (Oxford) 49(11):2210–2216.  https://doi.org/10.1093/rheumatology/keq250 CrossRefGoogle Scholar
  3. 3.
    Aletaha D, Neogi T, Silman AJ, Funovits J, Felson DT, Bingham CO 3rd, Birnbaum NS, Burmester GR, Bykerk VP, Cohen MD, Combe B, Costenbader KH, Dougados M, Emery P, Ferraccioli G, Hazes JM, Hobbs K, Huizinga TW, Kavanaugh A, Kay J, Kvien TK, Laing T, Mease P, Menard HA, Moreland LW, Naden RL, Pincus T, Smolen JS, Stanislawska-Biernat E, Symmons D, Tak PP, Upchurch KS, Vencovsky J, Wolfe F, Hawker G (2010) 2010 Rheumatoid arthritis classification criteria: an American College of Rheumatology/European League Against Rheumatism collaborative initiative. Arthritis Rheum 62(9):2569–2581.  https://doi.org/10.1002/art.27584 CrossRefGoogle Scholar
  4. 4.
    Duer-Jensen A, Vestergaard A, Dohn UM, Ejbjerg B, Hetland ML, Albrecht-Beste E, Ostergaard M (2008) Detection of rheumatoid arthritis bone erosions by two different dedicated extremity MRI units and conventional radiography. Ann Rheum Dis 67(7):998–1003.  https://doi.org/10.1136/ard.2007.076026 CrossRefPubMedGoogle Scholar
  5. 5.
    Ostergaard M, Peterfy CG, Bird P, Gandjbakhch F, Glinatsi D, Eshed I, Haavardsholm EA, Lillegraven S, Boyesen P, Ejbjerg B, Foltz V, Emery P, Genant HK, Conaghan PG (2017) The OMERACT Rheumatoid Arthritis Magnetic Resonance Imaging (MRI) Scoring System: updated recommendations by the OMERACT MRI in arthritis working group. J Rheumatol 44(11):1706–1712.  https://doi.org/10.3899/jrheum.161433 CrossRefPubMedGoogle Scholar
  6. 6.
    Kawashiri SY, Suzuki T, Okada A, Yamasaki S, Tamai M, Nakamura H, Origuchi T, Mizokami A, Uetani M, Aoyagi K, Eguchi K, Kawakami A (2013) Musculoskeletal ultrasonography assists the diagnostic performance of the 2010 classification criteria for rheumatoid arthritis. Mod Rheumatol 23(1):36–43.  https://doi.org/10.1007/s10165-012-0628-7 CrossRefPubMedGoogle Scholar
  7. 7.
    Dohn UM, Ejbjerg BJ, Hasselquist M, Narvestad E, Court-Payen M, Szkudlarek M, Moller J, Thomsen HS, Ostergaard M (2007) Rheumatoid arthritis bone erosion volumes on CT and MRI: reliability and correlations with erosion scores on CT, MRI and radiography. Ann Rheum Dis 66(10):1388–1392.  https://doi.org/10.1136/ard.2007.072520 CrossRefPubMedPubMedCentralGoogle Scholar
  8. 8.
    Takase-Minegishi K, Horita N, Kobayashi K, Yoshimi R, Kirino Y, Ohno S, Kaneko T, Nakajima H, Wakefield RJ, Emery P (2018) Diagnostic test accuracy of ultrasound for synovitis in rheumatoid arthritis: systematic review and meta-analysis. Rheumatology (Oxford) 57(1):49–58.  https://doi.org/10.1093/rheumatology/kex036 CrossRefGoogle Scholar
  9. 9.
    Peluso G, Bosello SL, Gremese E, Mirone L, Di Gregorio F, Di Molfetta V, Pirronti T, Ferraccioli G (2015) Detection of bone erosions in early rheumatoid arthritis: 3D ultrasonography versus computed tomography. Clin Rheumatol 34(7):1181–1186.  https://doi.org/10.1007/s10067-015-2938-6 CrossRefPubMedGoogle Scholar
  10. 10.
    Maxwell L, Santesso N, Tugwell PS, Wells GA, Judd M, Buchbinder R (2006) Method guidelines for Cochrane Musculoskeletal Group systematic reviews. J Rheumatol 33(11):2304–2311PubMedGoogle Scholar
  11. 11.
    Whiting P, Rutjes AW, Reitsma JB, Bossuyt PM, Kleijnen J (2003) The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol 3:25.  https://doi.org/10.1186/1471-2288-3-25 CrossRefPubMedPubMedCentralGoogle Scholar
  12. 12.
    Nordberg LB, Lillegraven S, Aga AB, Sexton J, Olsen IC, Lie E, Berner Hammer H, Uhlig T, van der Heijde D, Kvien TK, Haavardsholm EA (2018) Comparing the disease course of patients with seronegative and seropositive rheumatoid arthritis fulfilling the 2010 ACR/EULAR classification criteria in a treat-to-target setting: 2-year data from the ARCTIC trial. RMD Open 4(2):e000752.  https://doi.org/10.1136/rmdopen-2018-000752 CrossRefPubMedPubMedCentralGoogle Scholar
  13. 13.
    Baillet A, Gaujoux-Viala C, Mouterde G, Pham T, Tebib J, Saraux A, Fautrel B, Cantagrel A, Le Loet X, Gaudin P (2011) Comparison of the efficacy of sonography, magnetic resonance imaging and conventional radiography for the detection of bone erosions in rheumatoid arthritis patients: a systematic review and meta-analysis. Rheumatology 50(6):1137–1147.  https://doi.org/10.1093/rheumatology/keq437 CrossRefPubMedGoogle Scholar
  14. 14.
    McGrath TA, Alabousi M, Skidmore B, Korevaar DA, Bossuyt PMM, Moher D, Thombs B, McInnes MDF (2017) Recommendations for reporting of systematic reviews and meta-analyses of diagnostic test accuracy: a systematic review. Syst Rev 6(1):194.  https://doi.org/10.1186/s13643-017-0590-8 CrossRefPubMedPubMedCentralGoogle Scholar
  15. 15.
    Chen B, Benedetti A (2017) Quantifying heterogeneity in individual participant data meta-analysis with binary outcomes. Syst Rev 6(1):243.  https://doi.org/10.1186/s13643-017-0630-4 CrossRefPubMedPubMedCentralGoogle Scholar
  16. 16.
    Wakefield RJ, Gibbon WW, Conaghan PG, O’Connor P, McGonagle D, Pease C, Green MJ, Veale DJ, Isaacs JD, Emery P (2000) The value of sonography in the detection of bone erosions in patients with rheumatoid arthritis: a comparison with conventional radiography. Arthritis Rheum 43(12):2762–2770.  https://doi.org/10.1002/1529-0131(200012)43:12<2762::AID-ANR16>3.0.CO;2-#
  17. 17.
    Lindegaard H, Vallo J, Horslev-Petersen K, Junker P, Ostergaard M (2001) Low field dedicated magnetic resonance imaging in untreated rheumatoid arthritis of recent onset. Ann Rheum Dis 60(8):770–776CrossRefGoogle Scholar
  18. 18.
    Alarcon GS, Lopez-Ben R, Moreland LW (2002) High-resolution ultrasound for the study of target joints in rheumatoid arthritis. Arthritis Rheum 46(7):1969–1970; author reply 1970-1961.  https://doi.org/10.1002/art.10310 CrossRefPubMedGoogle Scholar
  19. 19.
    Lopez-Ben R, Bernreuter WK, Moreland LW, Alarcon GS (2004) Ultrasound detection of bone erosions in rheumatoid arthritis: a comparison to routine radiographs of the hands and feet. Skeletal Radiol 33(2):80–84.  https://doi.org/10.1007/s00256-003-0693-2 CrossRefPubMedGoogle Scholar
  20. 20.
    Szkudlarek M, Narvestad E, Klarlund M, Court-Payen M, Thomsen HS, Ostergaard M (2004) Ultrasonography of the metatarsophalangeal joints in rheumatoid arthritis: comparison with magnetic resonance imaging, conventional radiography, and clinical examination. Arthritis Rheum 50(7):2103–2112.  https://doi.org/10.1002/art.20333 CrossRefPubMedGoogle Scholar
  21. 21.
    Ejbjerg BJ, Narvestad E, Jacobsen S, Thomsen HS, Ostergaard M (2005) Optimised, low cost, low field dedicated extremity MRI is highly specific and sensitive for synovitis and bone erosions in rheumatoid arthritis wrist and finger joints: comparison with conventional high field MRI and radiography. Ann Rheum Dis 64(9):1280–1287.  https://doi.org/10.1136/ard.2004.029850 CrossRefPubMedPubMedCentralGoogle Scholar
  22. 22.
    Bajaj S, Lopez-Ben R, Oster R, Alarcon GS (2007) Ultrasound detects rapid progression of erosive disease in early rheumatoid arthritis: a prospective longitudinal study. Skelet Radiol 36(2):123–128.  https://doi.org/10.1007/s00256-006-0196-z CrossRefGoogle Scholar
  23. 23.
    Yao L, Magalnick M, Wilson M, Lipsky P, Goldbach-Mansky R (2006) Periarticular bone findings in rheumatoid arthritis: T2-weighted versus contrast-enhanced T1-weighted MRI. AJR Am J Roentgenol 187(2):358–363.  https://doi.org/10.2214/AJR.05.0098 CrossRefPubMedGoogle Scholar
  24. 24.
    Dohn UM, Ejbjerg BJ, Court-Payen M, Hasselquist M, Narvestad E, Szkudlarek M, Moller JM, Thomsen HS, Ostergaard M (2006) Are bone erosions detected by magnetic resonance imaging and ultrasonography true erosions? A comparison with computed tomography in rheumatoid arthritis metacarpophalangeal joints. Arthritis Res Ther 8(4):R110.  https://doi.org/10.1186/ar1995 CrossRefPubMedPubMedCentralGoogle Scholar
  25. 25.
    Szkudlarek M, Klarlund M, Narvestad E, Court-Payen M, Strandberg C, Jensen KE, Thomsen HS, Ostergaard M (2006) Ultrasonography of the metacarpophalangeal and proximal interphalangeal joints in rheumatoid arthritis: a comparison with magnetic resonance imaging, conventional radiography and clinical examination. Arthritis Res Ther 8(2):R52.  https://doi.org/10.1186/ar1904 CrossRefPubMedPubMedCentralGoogle Scholar
  26. 26.
    Freeston JE, Conaghan PG, Dass S, Vital E, Hensor EM, Stewart SP, Emery P (2007) Does extremity-MRI improve erosion detection in severely damaged joints? A study of long-standing rheumatoid arthritis using three imaging modalities. Ann Rheum Dis 66(11):1538–1540.  https://doi.org/10.1136/ard.2007.074708 CrossRefPubMedPubMedCentralGoogle Scholar
  27. 27.
    Dohn UM, Ejbjerg BJ, Hasselquist M, Narvestad E, Moller J, Thomsen HS, Ostergaard M (2008) Detection of bone erosions in rheumatoid arthritis wrist joints with magnetic resonance imaging, computed tomography and radiography. Arthritis Res Ther 10(1):R25.  https://doi.org/10.1186/ar2378 CrossRefPubMedPubMedCentralGoogle Scholar
  28. 28.
    Duer A, Ostergaard M, Horslev-Petersen K, Vallo J (2008) Magnetic resonance imaging and bone scintigraphy in the differential diagnosis of unclassified arthritis. Ann Rheum Dis 67(1):48–51.  https://doi.org/10.1136/ard.2006.063792 CrossRefPubMedGoogle Scholar
  29. 29.
    Funck-Brentano T, Etchepare F, Joulin SJ, Gandjbakch F, Pensec VD, Cyteval C, Miquel A, Benhamou M, Banal F, Le Loet X, Cantagrel A, Bourgeois P, Fautrel B (2009) Benefits of ultrasonography in the management of early arthritis: a cross-sectional study of baseline data from the ESPOIR cohort. Rheumatology (Oxford) 48(12):1515–1519.  https://doi.org/10.1093/rheumatology/kep279 CrossRefGoogle Scholar
  30. 30.
    Dohn UM, Ejbjerg B, Boonen A, Hetland ML, Hansen MS, Knudsen LS, Hansen A, Madsen OR, Hasselquist M, Moller JM, Ostergaard M (2011) No overall progression and occasional repair of erosions despite persistent inflammation in adalimumab-treated rheumatoid arthritis patients: results from a longitudinal comparative MRI, ultrasonography, CT and radiography study. Ann Rheum Dis 70(2):252–258.  https://doi.org/10.1136/ard.2009.123729 CrossRefPubMedGoogle Scholar
  31. 31.
    Rahmani M, Chegini H, Najafizadeh SR, Azimi M, Habibollahi P, Shakiba M (2010) Detection of bone erosion in early rheumatoid arthritis: ultrasonography and conventional radiography versus non-contrast magnetic resonance imaging. Clin Rheumatol 29(8):883–891.  https://doi.org/10.1007/s10067-010-1423-5 CrossRefPubMedGoogle Scholar
  32. 32.
    Tamai M, Kawakami A, Uetani M, Fukushima A, Arima K, Fujikawa K, Iwamoto N, Aramaki T, Kamachi M, Nakamura H, Ida H, Origuchi T, Aoyagi K, Eguchi K (2012) Magnetic resonance imaging (MRI) detection of synovitis and bone lesions of the wrists and finger joints in early-stage rheumatoid arthritis: comparison of the accuracy of plain MRI-based findings and gadolinium-diethylenetriamine pentaacetic acid-enhanced MRI-based findings. Mod Rheumatol 22(5):654–658.  https://doi.org/10.1007/s10165-011-0575-8 CrossRefPubMedGoogle Scholar
  33. 33.
    Dohn UM, Terslev L, Szkudlarek M, Hansen MS, Hetland ML, Hansen A, Madsen OR, Hasselquist M, Moller J, Ostergaard M (2013) Detection, scoring and volume assessment of bone erosions by ultrasonography in rheumatoid arthritis: comparison with CT. Ann Rheum Dis 72(4):530–534.  https://doi.org/10.1136/annrheumdis-2011-201114 CrossRefPubMedGoogle Scholar
  34. 34.
    Amin MF, Ismail FM, El Shereef RR (2012) The role of ultrasonography in early detection and monitoring of shoulder erosions, and disease activity in rheumatoid arthritis patients; comparison with MRI examination. Acad Radiol 19(6):693–700.  https://doi.org/10.1016/j.acra.2012.02.010 CrossRefPubMedGoogle Scholar
  35. 35.
    Aoki T, Yamashita Y, Saito K, Tanaka Y, Korogi Y (2013) Diagnosis of early-stage rheumatoid arthritis: usefulness of unenhanced and gadolinium-enhanced MR images at 3T. Clin Imaging 37(2):348–353.  https://doi.org/10.1016/j.clinimag.2012.07.004 CrossRefPubMedGoogle Scholar
  36. 36.
    Ji L, Li G, Xu Y, Zhou W, Zhang Z (2015) Early prediction of rheumatoid arthritis by magnetic resonance imaging in the absence of anti-cyclic citrullinated peptide antibodies and radiographic erosions in undifferentiated inflammatory arthritis patients: a prospective study. Int J Rheum Dis 18(8):859–865.  https://doi.org/10.1111/1756-185X.12420 CrossRefPubMedGoogle Scholar
  37. 37.
    Ji L, Deng X, Geng Y, Song Z, Zhang Z (2017) The additional benefit of ultrasonography to 2010 ACR/EULAR classification criteria when diagnosing rheumatoid arthritis in the absence of anti-cyclic citrullinated peptide antibodies. Clin Rheumatol 36(2):261–267.  https://doi.org/10.1007/s10067-016-3465-9 CrossRefPubMedGoogle Scholar
  38. 38.
    Rantalaiho V, Kautiainen H, Korpela M, Hannonen P, Kaipiainen-Seppanen O, Mottonen T, Kauppi M, Karjalainen A, Laiho K, Laasonen L, Hakola M, Peltomaa R, Leirisalo-Repo M, Group NE-RS (2014) Targeted treatment with a combination of traditional DMARDs produces excellent clinical and radiographic long-term outcomes in early rheumatoid arthritis regardless of initial infliximab. The 5-year follow-up results of a randomised clinical trial, the NEO-RACo trial. Ann Rheum Dis 73(11):1954–1961.  https://doi.org/10.1136/annrheumdis-2013-203497 CrossRefPubMedGoogle Scholar
  39. 39.
    Hassan R, Hussain S, Bacha R, Gillani SA, Malik SS (2019) Reliability of Ultrasound for the Detection of Rheumatoid Arthritis. J Med Ultrasound 27(1):3–12.  https://doi.org/10.4103/JMU.JMU_112_18 CrossRefPubMedPubMedCentralGoogle Scholar

Copyright information

© International League of Associations for Rheumatology (ILAR) 2019

Authors and Affiliations

  1. 1.Department of Bone and Joint Surgery, Department of Orthopedics, Renji Hospital, School of MedicineShanghai Jiao Tong UniversityShanghaiChina

Personalised recommendations