Skip to main content

Table 2 Agreement comparisons among different segmentation methods for spontaneous ICH volume in two groups

From: Efficiency of a deep learning-based artificial intelligence diagnostic system in spontaneous intracerebral hemorrhage volume measurement

Agreement Statistics

ICH without IVH group

ICH with IVH group

Algorithm versus CTP

ABC/2 score versus CTP

Algorithm versus ABC/2 score

Algorithm versus CTP

Difference, mL

Range (min, max)

− 7.00,7.670

− 8.77,21.35

− 22.04, 7.59

− 14.02,10.80

Mean

− 0.10

1.53

− 1.63

− 0.11

Median

− 0.30

− 0.33

− 0.60

− 0.03

IQR

− 0.77,0.47

− 0.50,2.25

− 2.51,0.54

− 0.94,1.43

95% LOA (low, high)

− 4.38,4.18

− 7.90,10.96

− 11.22,7.96

− 7.05,6.82

CCC [95% CI]

0.993

[0.989 to 0.996]

0.968

[0.948 to 0.980]

0.967

[0.946 to 0.980]

0.996

[0.993 to 0.998]

P

0.218*

0.658*

0.007*

0.941†

  1. ICH indicates intracerebral hemorrhage; IVH, intraventricular hemorrhage; CTP, CT− based planimetry; LOA, limits of agreement; IQR, interquartile range; CCC, concordance correlation coefficient; and CI, confidence limit
  2. *Friedman test, followed by pairwise comparisons
  3. †Wilcoxon signed−rank test
Back to article page

BMC Medical Imaging

ISSN: 1471-2342

Contact us