Table 3 Analysis of the false positives predicted by the CPM-Net model output before and after post-processing, for our private dataset (*: foramen magnum, ICA: internal carotid artery, A1: A1 segment of anterior cerebral artery, A2: A2 segment of anterior cerebral artery, ACA: anterior cerebral artery).

From: Automated, anatomy-based, heuristic post-processing reduces false positives and improves interpretability of deep learning intracranial aneurysm detection models

 

All FP

FP removed by

Method 1

Method 2

Method 3

Method 4

Method 5

Vein

71

11

67

67

67

67

Vein of Galen

33

0

33

33

33

33

Venous sinus

8

0

8

8

8

8

Cavernous venous sinus

3

0

0

0

0

0

Other intracranial veins

17

1*

16

16

16

16

Extracranial veins

10

10

10

10

10

10

Artery

15

2

6

1

6

2

Cervical artery

2

2

2

1

2

2

ICA infundibulum/branch points

2

0

0

0

0

0

MCA branch points

3

0

1

0

1

0

Caliber change at A1/A2 or ACA

4

0

1

0

1

0

Vertebrobasilar confluence

2

0

2

0

2

0

ICA occlusion

1

0

0

0

0

0

Ectatic basilar artery

1

0

0

0

0

0

Vessel

8

7

5

5

8

8

Intracranial

1

0

1

1

1

1

Extracranial

7

7

4

4

7

7

Tissue

32

8

5

5

12

12

Choroid plexus

19

0

2

2

2

2

Pineal gland

4

0

2

2

2

2

Other intracranial

1

0

0

0

0

0

Extracranial

8

8

1

1

8

8

Total

126

28

83

78

93

89

Intracranial

99

1

66

62

66

62

Extracranial

27

27

17

16

27

27