J Clin Med Res

Journal of Clinical Medicine Research, ISSN 1918-3003 print, 1918-3011 online, Open Access

Article copyright, the authors; Journal compilation copyright, J Clin Med Res and Elmer Press Inc

Journal website https://jocmr.elmerjournals.com

Original Article

Volume 17, Number 7, July 2025, pages 398-407

Improving a Clinical Prediction Model for Computed Tomography Head Scan Use in Non-Traumatic Seizures: The SeizCT Optimized Model

↓ Figure 5. Predicted probability of positive CT brain findings plotted against the linear predictor. CT: computed tomography.

↓ **Table 1.** Baseline Characteristics of Patients With Positive and Negative Findings on Computed Tomography (CT) Brain Scans
Baseline characteristics	Missing (n = 55), n (%)	Positive CT (n = 118), n (%)	Negative CT (n = 507), n (%)	P value
SD: standard deviation; SBP: systolic blood pressure; DBP: diastolic blood pressure; GCS: Glasgow Coma Scale; IQR: interquartile range.
Female	0 (0)	40 (33.9)	117 (23.1)	0.018
Age (years), mean ± SD	0 (0)	60.4 ± 15.1	53.5 ± 16.9	< 0.001
SBP (mm Hg), mean ± SD	0 (0)	145.5 ± 33.07	139.4 ± 26.8	0.035
DBP (mm Hg), mean ± SD	0 (0)	85.7 ± 20.6	84.4 ± 17.8	0.481
GCS, median (IQR)	0 (0)	11 (7, 15)	15 (11, 15)	< 0.001
Eye, median (IQR)	0 (0)	4 (2, 4)	4 (4, 4)	< 0.001
Verbal, median (IQR)	0 (0)	2 (1, 5)	5 (2, 5)	< 0.001
Motor, median (IQR)	0 (0)	5 (4, 6)	6 (5, 6)	< 0.001
Pupil (mm), median (IQR)	39 (6.2)	3 (2, 3)	3 (3, 3)	0.395
GCS change from baseline, median (IQR)	0 (0)	-4 (-8, 0)	0 (-3, 0)	< 0.001
Focal neurological deficit	0 (0)	65 (55.1)	88 (17.4)	< 0.001
Previous CT
No previous CT	0 (0)	68 (57.6)	243 (47.9)	0.019
Normal finding	0 (0)	11 (9.3)	99 (19.5)
Abnormal finding	0 (0)	39 (33.1)	165 (32.5)
Prior stroke more than 3 months	0 (0)	16 (13.6)	95 (18.7)	0.228
Current cancer	0 (0)	17 (14.4)	14 (2.8)	< 0.001
Epilepsy	0 (0)	16(13.6)	135 (26.6)	0.003
End-stage renal disease	0 (0)	9 (7.6)	35 (6.9)	0.841
Chronic liver disease	0 (0)	3 (2.5)	13 (2.6)	1.000
Anticoagulant use	0 (0)	7 (5.9)	26 (5.1)	0.654
Hypertension	0 (0)	43 (36.4)	164 (32.4)	0.447
Diabetes mellitus	0 (0)	19 (16.1)	87 (17.2)	0.892
Smoking	7 (1.1)	21 (18.1)	76(15.1)	0.479
Alcohol consumption	6 (1.0)	39 (33.6)	198 (39.4)	0.290
Substance use	3 (0.5)	2 (1.7)	20 (4.0)	0.403
Alcohol withdrawal symptoms	0 (0)	2 (1.7)	92 (18.2)	< 0.001

↓ Table 1. Baseline Characteristics of Patients With Positive and Negative Findings on Computed Tomography (CT) Brain Scans

Baseline characteristics

Missing (n = 55), n (%)

Positive CT (n = 118), n (%)

Negative CT (n = 507), n (%)

P value

SD: standard deviation; SBP: systolic blood pressure; DBP: diastolic blood pressure; GCS: Glasgow Coma Scale; IQR: interquartile range.

Female

0 (0)

40 (33.9)

117 (23.1)

0.018

Age (years), mean ± SD

0 (0)

60.4 ± 15.1

53.5 ± 16.9

< 0.001

SBP (mm Hg), mean ± SD

0 (0)

145.5 ± 33.07

139.4 ± 26.8

0.035

DBP (mm Hg), mean ± SD

0 (0)

85.7 ± 20.6

84.4 ± 17.8

0.481

GCS, median (IQR)

0 (0)

11 (7, 15)

15 (11, 15)

< 0.001

Eye, median (IQR)

0 (0)

4 (2, 4)

4 (4, 4)

< 0.001

Verbal, median (IQR)

0 (0)

2 (1, 5)

5 (2, 5)

< 0.001

Motor, median (IQR)

0 (0)

5 (4, 6)

6 (5, 6)

< 0.001

Pupil (mm), median (IQR)

39 (6.2)

3 (2, 3)

3 (3, 3)

0.395

GCS change from baseline, median (IQR)

0 (0)

-4 (-8, 0)

0 (-3, 0)

< 0.001

Focal neurological deficit

0 (0)

65 (55.1)

88 (17.4)

< 0.001

Previous CT

No previous CT

0 (0)

68 (57.6)

243 (47.9)

0.019

Normal finding

0 (0)

11 (9.3)

99 (19.5)

Abnormal finding

0 (0)

39 (33.1)

165 (32.5)

Prior stroke more than 3 months

0 (0)

16 (13.6)

95 (18.7)

0.228

Current cancer

0 (0)

17 (14.4)

14 (2.8)

< 0.001

Epilepsy

0 (0)

16(13.6)

135 (26.6)

0.003

End-stage renal disease

0 (0)

9 (7.6)

35 (6.9)

0.841

Chronic liver disease

0 (0)

3 (2.5)

13 (2.6)

1.000

Anticoagulant use

0 (0)

7 (5.9)

26 (5.1)

0.654

Hypertension

0 (0)

43 (36.4)

164 (32.4)

0.447

Diabetes mellitus

0 (0)

19 (16.1)

87 (17.2)

0.892

Smoking

7 (1.1)

21 (18.1)

76(15.1)

0.479

Alcohol consumption

6 (1.0)

39 (33.6)

198 (39.4)

0.290

Substance use

3 (0.5)

2 (1.7)

20 (4.0)

0.403

Alcohol withdrawal symptoms

0 (0)

2 (1.7)

92 (18.2)

< 0.001

↓ **Table 2.** Predictors of Positive CT Head Scan Findings From Multivariable Logistic Regression
Predictors	Positive findings on CT head scan
CT: computed tomography; mOR: multivariable odds ratio; CI: confidence interval; GCS: Glasgow Coma Scale.
GCS change from baseline (points)	1.14	1.07, 1.21	< 0.001
Focal neurological deficit	3.83	2.40, 6.11	< 0.001
Current cancer	4.11	1.79, 9.43	< 0.001
Alcohol withdrawal symptoms	0.12	0.03, 0.49	0.003
Epilepsy	0.36	0.19, 0.68	< 0.001
Prior stroke more than 3 months	0.49	0.26, 0.93	0.030

↓ Table 2. Predictors of Positive CT Head Scan Findings From Multivariable Logistic Regression

Predictors

Positive findings on CT head scan

mOR

95% CI

P value

CT: computed tomography; mOR: multivariable odds ratio; CI: confidence interval; GCS: Glasgow Coma Scale.

GCS change from baseline (points)

1.14

1.07, 1.21

< 0.001

Focal neurological deficit

3.83

2.40, 6.11

< 0.001

Current cancer

4.11

1.79, 9.43

< 0.001

Alcohol withdrawal symptoms

0.12

0.03, 0.49

0.003

Epilepsy

0.36

0.19, 0.68

< 0.001

Prior stroke more than 3 months

0.49

0.26, 0.93

0.030

↓ **Table 3.** Internal Validation of the Predictive Model Using the Bootstrap Resampling Method
Parameters	Apparent performance	Bootstrap performance
AuROC: area under the receiver operating characteristic curve; E:O ratio: expected-to-observed outcomes ratio; CITL: calibration-in-the-large.
AuROC	0.8221 (0.7813, 0.8629)	0.8160 (0.7700, 0.8550)
Slope	1.0000 (0.7930, 1.2070)	0.9570 (0.7440, 1.1960)
E:O ratio	1.0000	1.0160
CITL	-0.0000 (-0.2260, 0.2260)	-0.0020 (-0.2300, 0.2370)
Bootstrap shrinkage	-	0.9570

↓ Table 3. Internal Validation of the Predictive Model Using the Bootstrap Resampling Method

Parameters

Apparent performance

Bootstrap performance

AuROC: area under the receiver operating characteristic curve; E:O ratio: expected-to-observed outcomes ratio; CITL: calibration-in-the-large.

AuROC

0.8221 (0.7813, 0.8629)

0.8160 (0.7700, 0.8550)

Slope

1.0000 (0.7930, 1.2070)

0.9570 (0.7440, 1.1960)

E:O ratio

1.0000

1.0160

CITL

-0.0000 (-0.2260, 0.2260)

-0.0020 (-0.2300, 0.2370)

Bootstrap shrinkage

0.9570

↓ **Table 4.** Confusion Matrices and Diagnostic Performance of the SeizCT Optimized Model at Different Probability Thresholds
Cut point	CT +ve (n = 118)	CT -ve (n = 507)	Total	Sensitivity	Specificity	PPV	NPV
CT: computed tomography; CT +ve: positive finding on CT brain; CT -ve: negative finding on CT brain; PPV: positive predictive value; NPV: negative predictive value.
Probability > 18.9%	94 (15.0%)	116 (18.6%)	210	79.7%	77.1%	44.8%	94.2%
Probability ≤ 18.9%	24 (3.8%)	391 (62.6%)	415
Probability > 22.57%	88 (14.1%)	90 (14.4%)	178	74.6%	82.2%	49.4%	93.3%
Probability ≤ 22.57%	30 (4.8%)	417 (66.7%)	447
Probability > 25%	84 (13.4%)	83 (13.3%)	167	71.2%	83.6%	50.3%	92.6%
Probability ≤ 25%	34 (5.4%)	424 (67.8%)	458

↓ Table 4. Confusion Matrices and Diagnostic Performance of the SeizCT Optimized Model at Different Probability Thresholds

Cut point

CT +ve (n = 118)

CT -ve (n = 507)

Total

Sensitivity

Specificity

PPV

NPV

CT: computed tomography; CT +ve: positive finding on CT brain; CT -ve: negative finding on CT brain; PPV: positive predictive value; NPV: negative predictive value.

Probability > 18.9%

94 (15.0%)

116 (18.6%)

210

79.7%

77.1%

44.8%

94.2%

Probability ≤ 18.9%

24 (3.8%)

391 (62.6%)

415

Probability > 22.57%

88 (14.1%)

90 (14.4%)

178

74.6%

82.2%

49.4%

93.3%

Probability ≤ 22.57%

30 (4.8%)

417 (66.7%)

447

Probability > 25%

84 (13.4%)

83 (13.3%)

167

71.2%

83.6%

50.3%

92.6%

Probability ≤ 25%

34 (5.4%)

424 (67.8%)

458