Huang et al reported a model for predicting symptomatic epilepsy in a patient with radiation-induced brain necrosis. This can help to identify a patient who may benefit from anticonvulsant therapy. The authors are from multiple institutions from around the world.
Patient selection: radiation-induced brain necrosis
Parameters:
(1) Dmax of the temporal lobe (Gray), from 62.1 to 82
(2) treatment
(3) sex
(4) radiation necrosis in volume in cubic cm, from 0 to 250 cc
(5) serum creatinine phosphokinase in U/L, from 0 to 900
(6) serum total cholesterol in mmol/L, from 2 to 9
(7) medical comorbidity
points for Dmax =
= (3.8556 * (Dmax)) - 239.4556
points for radiation necrosis volume =
= (0.1816 * (volume)) + 39.1
points for serum creatinine phosphokinase =
= (0.07267 * (CPK)) + 37.3
points for serum total cholesterol =
= 77.2143 - (6.0571 * (cholesterol))
Parameter
|
Finding
|
Points
|
treatment
|
pulse steroids
|
45
|
|
conventional steroids
|
64.5
|
|
bevacizumab
|
78.8
|
sex
|
female
|
21.8
|
|
male
|
45.3
|
medical comorbidity
|
no
|
45.5
|
|
yes
|
59.4
|
total score =
= SUM(points for all of the parameters)
Interpretation:
• minimum score: 210
• maximum score: 505.4
X =
= (0.04783 * (score)) - 17.73
Y =
= (0.0008748 * ((score)^2)) - (0.5809 * (score)) + 95.34
• The higher the score the greater the risk for epilepsy.
Score
|
3-Year Probability of Epilepsy
|
< 259
|
< 0.5%
|
259 to 360.23
|
1 / (1 + EXP((-1) * X))
|
360.23 to 393
|
1 / (1 + EXP((-1) * Y))
|
> 393
|
> 90%
|