Cloppenborg et all reported models for predicting developmental and intellectual outcomes of a pediatric patient undergoing epilepsy surgery. The authors are from Bielefeld University and University Medical Center Utrecht in Europe.
Patient selection: pediatric patient undergoing epilepsy surgery
Parameters:
(1) preoperative intelligence quotient (IQ), from 40 to 140
(2) preoperative developmental quotient (DQ), from 0 to 80
(3) extent of surgery
(4) invasive diagnostics
(5) etiology
points for preoperative IQ =
= (preoperative IQ) - 40
points for preoperative DQ =
= (1.25 * (preoperative DQ))
Parameter
|
Finding
|
Points
|
surgery
|
multi-lobar
|
0
|
|
focal
|
11.3
|
invasive diagnostics
|
no
|
7.5
|
|
yes
|
0
|
etiology
|
malformation of cortical development, vascular, encephalitis
|
0
|
|
other
|
15.1
|
score for postoperative IQ =
= (points for preoperative IQ) + (points for surgery) + (points for invasive diagnostics)
value of X for postoperative IQ =
= (0.1122 * (score for postoperative IQ)) - 8.574
probability IQ will be >= 85 at postoperative month 24 =
= 1 / (1 + EXP((-1) * X))
score for postoperative DQ =
= (points for preoperative DQ) + (points for etiology)
value of Y for postoperative DQ =
= (0.09148 * (score for postoperative DQ)) - 5.934
probability DQ will be >= 50 at postoperative month 24 =
= 1 / (1 + EXP((-1) * Y))
Performance (Figure 1):
• The area under the ROC curve for postoperative IQ is 0.90.
• The area under the ROC curve for postoperative DQ is 0.85.