Outcome Prediction after Moderate and
Severe Head Injury Using an Artificial
Neural Network
Min-Huei Hsua,b, Yu-Chuan Lic, Wen-Ta Chiud, Ju-Chuan Yene a
Department of Neurosurgery, eDepartment of Ophthalmology, Taipei City Hospital, Zhongxiao Branch
b
Graduate Institute of Medical Sciences, cGraduate Institute of Medical Informatics, Taipei Medical University
d
Taipei Municipal Wanfang Hospital (Managed by Taipei Medical University)
Abstract
Many studies have constructed predictive models for outcome after traumatic brain injury. Most of these attempts focused on dichotomous result, such as alive vs dead or good outcome vs poor outcome. If we want to predict more specific levels of outcome, we need more sophisticated models. We conducted this study to determine if artificial neural network modeling would predict outcome in five levels of Glasgow Outcome Scale (death, persistent vegetative state, severe disability, moderate disability, and good recovery) after moderate to severe head injury. The database was collected from a nation-wide epidemiological study of traumatic brain injury in Taiwan from July 1, 1995 to June 30, 1998. There were total 18583 records in this database and each record had thirty-two parameters. After pruning the records with minor cases (GCS 13) and missing data in the 132 variables, the number of cases decreased from 18583 to 4460. A step-wise logistic regression was applied to the remaining data set and 10 variables were selected as being statically significant in predicting outcome. These 10 variables were used as the input neurons for constructing neural network. Overall, 75.8% of predictions of this model were correct, 14.6% were pessimistic, and 9.6% optimistic. This neural network model demonstrated a significant difference of performance between different levels of Glasgow Outcome Scale. The prediction performance of dead or good recovery is best and the prediction of vegetative state is worst. An artificial neural network may provide a useful "second opinion" to assist neurosurgeon to predict outcome after traumatic brain injury.
Keywords:
Head injury; Artificial neural network; Outcome prediction
1. Introduction
Considerable effort has been devoted to improving our ability to predict outcome after traumatic brain injury (TBI). More reliable prediction of outcome would be helpful for clinicians as an important aid to decision making about management and for communication with relatives and other healthcare professionals.
Mathematical and statistical methods have been used to develop models for outcome prediction. The most commonly used methods include Bayes' theorem [1], logistic regression and neural networks. [2]
A artificial neural network is a computerized construct consisting of input neurons (which process input data) connected to hidden neurons (to mathematically manipulate values they receive from all the input neurons) connected to output neurons (to output a prediction). Artificial neural networks have been successfully used for pattern recognition and outcome prediction in several clinical settings. The advantage of a neural network is the ability of the model to capture nonlinearities and complex interactions between factors related to the outcome of interest. Neural networks differ from other decision support systems in that the learning occurs by example through training and not by programming or pre-defined rules.
We conducted this study to determine if artificial neural network modeling would predict outcome using five levels of Glasgow Outcome Scale [3] (death, persistent vegetative state, severe disability, moderate disability, and good recovery) after moderate and severe head injury (initial Glasgow Coma Scale-Score of 3-12 )
2. Materials and methods
This study was conducted using data collected from a nation-wide epidemiological study of traumatic brain injury in Taiwan from July 1, 1995 to June 30, 1998. One hundred and sixteen large to medium-sized teaching hospitals with qualified neurosurgeon participated in this study. There were total 18583 records in this database. The causes of head trauma were traffic crashes (14354 cases, of whom more than 65% were motorcycle crash victims), falls (2534 cases), and others (1695 cases). The mean age of the victims was 36.5 +/- 15.3 (SD), range 1 to 85 years old. In 7.8% of the cases, the victims were older than 60. The male to female ratio was about 3:1.
One hundred and thirty-two parameters including age, gender, causes of head trauma, GCS scores at the emergency department, CT findings and craniotomy for intracranial hematoma were recorded for each patient. The outcome was estimated by the Glasgow Outcome Scale (GOS), and was assessed as longer as 12 months after injury if possible.
After pruning the records with mild cases (GCS 13) and missing data in the 132 variables , the number of cases decreased from 18583 to 4460. In the second step, a step-wise logistic regression was applied to the remaining data set and 10 variables (Table 1) were selected as being statically significant (p<0.05) in predicting of the dependent variable (Glasgow Outcome Scale).
From the 4460 cases, 75% were randomly selected as the training group (n=3345) in the development of the neural network models. The validation group (n=1115) was used to test the performance of this model. Generalized regression neural network software was used (NeuralShell Classifier Version 2; Ward Systems Group; Frederick, MD).
The accuracy, sensitivity and specificity are used to describe the performance of the predictive model.
Table 1 Variables as input neurons in the ANN model Age
Number of nonreactive pupils Score of motor resonse Score of verbal response Score of eye opening
Use of helmet in motorcycle crash Intracerebral hematoma on CT Subdural hematoma on CT
Craniotomy for intracranial hematoma Alcohol-related traffic accident
Table 2 Actual and predicted outcome for 1115 patients
Actual "1" Actual "2" Actual "3" Actual "4" Actual "5" Total Predicted as "1" 134 10 17 12 23 196 Predicted as "2" 2 12 3 5 1 23 Predicted as "3" 0 0 66 7 9 82 Predicted as "4" 20 2 14 132 19 187 Predicted as "5" 18 1 3 31 574 627 Total 174 25 103 187 626 1115
1=death, 2=vegetative, 3= severe disability, 4= moderate disability, 5=good recovery Table 3 The sensitivity and specificity of prediction
1 2 3 4 5
Sensitivity 77.01% 48.00% 64.08% 70.59% 91.69% Specificity 93.41% 98.99% 98.42% 94.07% 89.16%
1=death, 2=vegetative, 3= severe disability, 4= moderate disability, 5=good recovery
3. Results
The prediction results from the ANN are shown in Table 2. As can be seen from the diagonal cells in the contingency table, Overall, 75.8% of predictions were correct, 14.6% were pessimistic (outcome better than predicted), and 9.6% optimistic (outcome worse than predicted). For patients with good recovery, 91.6 % of predictions were correct. For patients with moderate disability, 70.5% of predictions were correct. For patients with severe disability, 64.0% of predictions were correct. For death, 77.0% of prediction were correct but for vegetative state only 48.0% of prediction were correct. The sensitivity and specificity for each level are shown in Table 3.
4. Discussion
Approaches to developing prognostic models vary from using traditional probabilistic techniques, originating from the field of statistics, to more qualitative and model-based techniques, originating from the field of artificial intelligence (AI).
Until recently, attempts to predict outcome have focused on dichotomous result, such as alive vs dead or good outcome vs poor function. [4,5] The use of single variables, such as GCS [6], image finding [7], intracranial pressure [8] or cerebral blood flow [9], has allowed for a reasonable degree of accuracy in predicting those outcomes. With increased interest in predicting more specific levels of function, however, more sophisticated models are required. Such models require inclusion of multiple variables and better algorithm.
The neural network model developed in this study provided acceptable performance of overall outcome prediction. However it demonstrated a significant difference of performance between different levels of prediction. The prediction performance of dead or good recovery is best and the prediction of vegetative state is worst. This may be due to the small case number of the vegetative group in this study.
In most outcome prediction studies of TBI patients, death and vegetative state are combined as a single level (poor outcome). However, these two states have significant difference for clinicians and patient's relatives. Some people even think survival in a persistent vegetative state is worse than death. Any model could predict persistent vegetative state in early stage would be very helpful for clinicians in assisting treatment limiting decisions. But the predictive power of our model is still not good enough for that purpose.
We excluded the mild cases in our study, because most these patients will have a good recovery. Including mild cases could let prediction models have better performance without clinical significance.
Many other authors have shown age [10,11], GCS score, pupillary responsiveness [12,13] and findings of computed tomography (CT) to be significant predictors of outcome after traumatic brain injury. In our study, use of helmet in motorcycle crash was a significant outcome predictor. Before implementation of the motorcycle helmet use law, motorcycle collisions accounted for 74% of the traffic accidents in Taiwan, and most of the motorcycle riders were not helmet users. The motorcycle-related deaths have reached 48 percent of all motor vehicle-related deaths. [14] After implementation the helmet law in Taiwan on June 1, 1997, the mortality and morbidity from motorcycle-related head injuries decreased effectively in Taiwan. [15]
Although the ANN is a valuable method for outcome prediction, some of its nature should be noted before it can be widely applied. One is the `black box' nature of the ANN, which means that the logical procedure of how networks determine a prediction cannot be observed. Hart and Wyatt
believe that this "black box" aspect is a major obstacle to the acceptance of neural nets as part of medical decision support systems. [16]
Accurate prediction of outcome in the individual patient remains difficult to achieve for both clinicians and computer program. Our research indicates that an artificial neural network may provide a useful "second opinion" to assist neurosurgeon to predict outcome after traumatic brain injury.
5. Reference
[1] Barlow P, Teasdale GM, Jennett B, et al. Computer assisted prediction of outcome of severely head-injured patients. Journal of Microcomputer Applications 1984;7:271-7.
[2] Lang EW. Pitts LH. Damron SL. Rutledge R. Outcome after severe head injury: an analysis of prediction based upon comparison of neural network versus logistic regression analysis. Neurological Research. 19(3):274-80, 1997 Jun.
[3] Jennett B, Bond M: Assessment of outcome after severe brain damage. A practical scale. Lancet 1:480, 1975
[4] Hsu MH, Li YC. Predicting cranial computed tomography results of head injury patients using an artificial neural network. AMIA Annu Symp Proc. 2003; 868.
[5] Jennett B. Teasdale G. Braakman R. Minderhoud J. Knill-Jones R. Predicting outcome in individual patients after severe head injury. Lancet. 1(7968):1031-4, 1976 May 15
[6] van Dongen KJ, Braakman R, Gelpke GJ: The prognostic value of computerized tomography in comatose head-injured patients. J Neurosurg 59:951, 1983
[7] Benzer A, Mitterschiffthaler G, Marosi M, et al. Prediction of non-survival after trauma. Lancet 1991;338:977-8.
[8] Czosnyka M. Guazzo E. Whitehouse M. Smielewski P. Czosnyka Z. Kirkpatrick P. Piechnik S. Pickard JD. Significance of intracranial pressure waveform analysis after head injury. Acta Neurochirurgica. 138(5):531-41, 1996.
[9] Kelly DF. Martin NA. Kordestani R. Counelis G. Hovda DA. Bergsneider M. McBride DQ. Shalmon E. Herman D. Becker DP. Cerebral blood flow as a predictor of outcome following traumatic brain injury. Journal of Neurosurgery. 86(4):633-41, 1997 Apr.
[10] Luerssen TG, Klauber MR, Marshall LF: Age and outcome from head injury: A longitudinal prospective study of adult and pediatric head injury. J Neurosurg 68:409, 1988
[11] Vollmer DG, Torner JC, Jane JA, et al: Age and outcome following traumatic coma: Why do older patients fare worse? J Neurosurg 75:S37, 1991
[12] Price DJ, Knill-Jones R: The prediction of outcome of patients admitted following head injury in coma with bilateral fixed pupils. Acta Neurochir Suppl 28:179, 1979
[13] Jennett B, Teasdale G, Braakman R, et al: Predicting outcome in individual patients after severe head injury. Lancet 1:1031, 1976
[14] Chiu WT. The motorcycle helmet law in Taiwan. JAMA. 274(12):941-2, 1995 Sep 27.
[15] Chiu WT. Kuo CY. Hung CC. Chen M. The effect of the Taiwan motorcycle helmet use law on head injuries. American Journal of Public Health. 90(5):793-6, 2000 May.
[16] Hart A, Wyatt J. Evaluating black-boxes as medical decision aids: issues arising from a study of neural networks. Med Inform 1990; 15: 229-36.
Address for correspondence
Min-Huei Hsu, M.D.
Address: 3F, No9, Lane 176, Sec1, Ta-An Road, Taipei, Taiwan E-mail address: [email protected]
