Table 1

Magnetic resonance imaging (MRI) erosions predicting subsequent radiographic erosions and MRI bone marrow edema predicting subsequent MRI erosions

MRI erosions predicting subsequent radiographic erosions

Time interval between studies

Patients

New x-ray erosions preceded by MRI erosion

MRI erosions NOT leading to x-ray erosion

New x-ray erosion NOT preceded by MRI erosion

Studies not having an x-ray or MRI erosion

Sensitivity

Specificity

PPV

NPV

Odds ratio

(CI)

6 months

43

3

32

1

430

0.75

0.93

0.086

0.998

^a40.3

(4.08–398.7)

12 months

39

3

27

2

417

0.6

0.94

0.10

0.995

^a32.2

(3.7–144.6)

24 months

41

3

15

1

218

0.75

0.94

0.17

0.995

^a43.6

(4.27–445)

MRI bone marrow edema predicting subsequent MRI erosions

Time interval between studies

Patients

New MRI erosions preceded by BME

BME NOT leading to MRI erosion

New MRI erosion NOT preceded by BME

Studies not having an MRI erosion or BME

Sensitivity

Specificity

PPV

NPV

Odds ratio

(CI)

6 months

43

4

12

4

410

0.50

0.97

0.25

0.99

^a34.17

(7.62–153.1)

12 months

39

6

6

3

204

0.67

0.97

0.50

0.99

^a68.0

(13.6–338.9)

The data for the major endpoints of the study are summarized here. The number of patients at each time interval varies because not all patients completed all studies. All radiographic erosions not preceded by an MRI erosion were preceded by MRI defects. There were only five new radiographic erosions for the entire cohort. ^aStatistically significant. BME, bone marrow edema; CI, confidence interval; NPV, negative predictive value; PPV, positive predictive value.

Mundwiler et al. Arthritis Research & Therapy 2009 11:R94   doi:10.1186/ar2737