Wang et al developed the BRAIN algorithm for predicting the risk of hematoma growth following acute intracerebral hemorrhage (ICH). This can help to identify a patient who may benefit from more aggressive management. The authors are from around the world and members of the INTERACT trial.
BRAIN is a mneumonic for the 5 parameters.
Parameters:
(1) baseline ICH volume in mL
(2) recurrent ICH
(3) anticoagulation with warfarin at onset
(4) intraventricular extension
(5) number of hours to baseline CT from symptom onset
|
Parameter
|
Finding
|
Points
|
|
baseline ICH volume
|
< 10 mL
|
0
|
|
|
10 to 20 mL
|
5
|
|
|
> 20 mL
|
7
|
|
recurrent ICH
|
no
|
0
|
|
|
yes
|
4
|
|
anticoagulation
|
no
|
0
|
|
|
yes
|
6
|
|
intraventricular extension
|
no
|
0
|
|
|
yes
|
2
|
|
number of hours to baseline CT
|
<= 1 hour
|
5
|
|
|
1.01 to 2 hours
|
4
|
|
|
2.01 to 3 hours
|
3
|
|
|
3.01 to 4 hours
|
2
|
|
|
4.01 to 5 hours
|
1
|
|
|
> 5 hours
|
0
|
total score =
= SUM(points for all 5 parameters)
Interpretation:
• minimum score: 0
• maximum score: 24
• The higher the score the greater the risk of hematoma growth.
|
Score
|
Percent
|
|
0
|
3.4%
|
|
1
|
4.2%
|
|
2
|
5.1%
|
|
3
|
6.3%
|
|
4
|
7.7%
|
|
5
|
9.4%
|
|
6
|
11.3%
|
|
7
|
13.7%
|
|
8
|
16.4%
|
|
9
|
19.5%
|
|
10
|
23.1%
|
|
11
|
27.2%
|
|
12
|
31.6%
|
|
13
|
36.4%
|
|
14
|
41.5%
|
|
15
|
46.7%
|
|
16
|
52.1%
|
|
17
|
57.4%
|
|
18
|
62.5%
|
|
19
|
67.4%
|
|
20
|
71.9%
|
|
21
|
76%
|
|
22
|
79.7%
|
|
23
|
83%
|
|
24
|
85.8%
|
The algorithm can also be expressed by logistic regression analysis.
X =
= (1.007 if ICH volume > 20 mL) + (1.6009 if volume 10.1 to 20 mL) + (0.7705 for recurrent ICH) + (1.025 for anticoagulation) + (0.4316 for ICH extension) - (0.2141 * (hours for CT from onset) - 2.1644
where:
• In the implementation set maximum hours scored to 6 with MAX(-1.2846,(hours*(-0.2141))).
probability of extension =
= 1 / (1 + EXP((-1) * X))
Performance:
• The area under the ROC curve was 0.73.
