Skip to main content

Table 3 The performance evaluation of the datasets and subgroups

From: A meta-analysis of the diagnostic performance of machine learning-based MRI in the prediction of axillary lymph node metastasis in breast cancer patients

Dataset

No. of studies

Sensitivity

Specificity

PLR

NLR

DOR

Testing set

3

0.76 (0.64–0.86)#

0.82 (0.72–0.89)##

3.90 (2.49–6.11)##

0.29 (0.19–0.46)#

12.88 (5.99–27.71)##

Validation set

12

0.79 (0.74–0.84)#

0.77 (0.73–0.81)#

3.47 (2.91–4.14)#

0.27 (0.22–0.33)#

12.92 (9.34–17.87)#

Overall validation group

      

Magnetic field strength (T)

1.5 T

7

0.81 (0.75–0.87)#

0.76 (0.71–0.81)#

3.38 (2.70–4.24)#

0.25 (0.18–0.34)#

13.15 (8.37–20.66)#

3.0 T

5

0.86 (0.78–0.89)**

0.76 (0.70–0.82)#

3.46 (2.69–4.46)#

0.23 (0.11–0.48)**

17.78 (10.54–30.00)##

Sequence

T2-FS

3

0.67 (0.54–0.79)#

0.80 (0.71–0.88)#

3.43 (2.22–5.29)##

0.41 (0.28–0.59)#

8.01 (4.01–16.00)##

T1CE

8

0.80 (0.75–0.84)##

0.80 (0.75–0.84)#

4.04 (3.25–5.02)#

0.25 (0.19–0.31)##

16.27 (11.08–23.89)#

DWI

2

0.76 (0.60–0.89)#

0.75 (0.63–0.85)#

3.10 (1.96–4.91)#

0.31 (0.17–0.56)#

10.00 (3.97–25.17)#

Algorithm

SVM

5

0.77 (0.71–0.82)**

0.79 (0.73–0.83)##

3.67 (2.87–4.70)#

0.30 (0.19–0.47)*

13.43 (8.62–20.93)##

LR

4

0.70 (0.58–0.81)#

0.78 (0.68–0.85)#

3.20 (2.17–4.71)#

0.37 (0.26–0.55)#

8.60 (4.44–16.66)#

LDA

3

0.83 (0.77–0.88)#

0.75 (0.68–0.81)#

3.35 (2.58–4.35)##

0.22 (0.16–0.31)#

14.78 (9.01–24.26)#

Segmentation

2D

4

0.77 (0.70–0.83)##

0.76 (0.70–0.82)#

3.30 (2.55–4.28)#

0.29 (0.21–0.39)#

11.46 (7.05–18.62)#

3D

9

0.80 (0.75–0.84)##

0.78 (0.73–0.82)#

3.56 (2.88–4.39)#

0.26 (0.21–0.34)##

13.52 (9.27–19.72)#

  1. PLR, positive likelihood ratio; NLR, negative likelihood ratio; DOR, diagnostic odds ratio; SVM, support vector machine; LR, linear regression; LDA, linear discriminant analysis; T2-FS, fat-suppressed T2; T1CE, contrast-enhanced T1; #, 0–25% of I2 values; ##, 25–50%; *, 50–75%; **, > 75%; # and ##, fixed-effects model; * and **, random-effects model