#ISBN: 3790814636
TITLE: Computational Intelligence Processing in Medical Diagnosis
AUTHOR: Schmitt et.al.
TOC:

Chapter 1.
An introduction to computational intelligence in medical diagnosis
H.-N Teodorescu and L.C Jain
1 What is computational intelligence? 1
2 Why CI in medicine and especially in medical diagnosis? 3
3 CI in medical diagnosis 4
4 Data mining and knowledge discovery 6
5 Qualitative reasoning methods 6
6 Issues related to CI management in medicine 6
7 The prospects of CI in medicine 7
Acknowledgments  9
References 10
Chapter 2.
Computational intelligence techniques in medical decision making: the data mining perspective
V. Maojo, I. Sanandres, H. Billhardt, and J. Crespo
1 Background - artificial intelligence in medicine 13
2 Data mining 16
2.1 Knowledge discovery in databases 16
2.2 Methods 18
2.2.1 Statistics and Pattern recognition 21
2.2.2 Machine learning 23
2.2.3 Artificial neural networks (ANNs) 27
2.3 Data mining tools 30
3 Applications in medicine 31
4 Limitations of data mining in medicine 34
5 Conclusions 36
References 37
Chapter 3.
Internet-based decision support for evidence-based medicine
J. Simpson, J.K.C Kingston, and N. Molony
1 Introduction 45
2 The protocol assistant - feasibility assessment 47
2.1 Feasibility: organizational issues 48
2.2 Feasibility: technical issues 50
2.3 Feasibility: project & personne1 issues 52
3 Representing clinical protocols 54
3.1 Knowledge acquisition and modeling using PROforma5 4
3.2 "Running" a clinical protocol using JESS 56
3.3 Representing and reasoning with clinical uncertainty 58
4 Design and implementation of the protocol assistant 60
4.1 System design 60
4.2 User interface design 63
4.3 Implementation 64
5 Evaluation and future work 65
Acknowledgments 68
References 68
Chapter 4.
Integrating kerne1 methods into a knowledge-based approach to evidence-based medicine
K. Morik, T. Joachims, M. Imhoff, P. Brockhausen, and S. Rping
1 Introduction 71
2 Data acquisition and data set 76
2.1 Data acquisition 76
2.2 Data set 78
2.3 Statistical preprocessing 79
3 Data-driven acquisition of state-action rules 80
3.1 Support vector machine 80
3.2 Learning the directions of interventions 81
3.3 Learning when to intervene 82
3.4 SVM rules in evidence based medicine 84
3.5 More learning tasks 85
4 Medical knowledge base 87
4.1 Knowledge acquisition and representation 87
4.2 Validating action-effect rules 91
4.3 Integrating learned decision functions with the knowledge base 92
5 Using the knowledge base of effects to validate interventions 93
5.1 Validating learned decision rules 93
5.2 Validating proposed interventions 94
6 Comparison with related work 95
7 Conclusions 96
Acknowledgements 97
References 97
Chapter 5.
Case-based reasoning prognosis for temporal courses
R. Schmidt and L. Gierl
1 Introduction 101
2 Methods 102
2.1 Case-based reasoning 102
2.2 Prognostic model 104
2.2.1 State abstraction 105
2.2.2 Temporal abstraction 105
2.2.3 CBR retrieval 106
3 Applications 106
3.1 Kidney function courses 107
3.1.1 Objectives 107
3.1.2 Methods 108
3.1.3 Learuing a tree of prototypes 115
3.1.4 Evaluation 118
3.2 Prognosis of the spread of diseases 120
3.2.1 Searching for similar courses 122
3.2.2 Adaptation 122
4 Generalization of our prognostic method 123
5 Summary 125
References 125
Chapter 6.
Pattern recognition in intensive care online monitoring
R. Fried, U. Gather, and M. Imhoff
1 Introduction 129
2 Curve fitting 131
3 Median filtering 137
4 Statistical time series analysis 139
5 Intervention analysis 143
6 Statistical process control 149
7 Online pattern recognition based on statistical time series analysis 156
7.1 Dynamic linear models 156
7.2 ARMA modeling 158
7.3 Trend detection 160
8 Conclusion 163
Acknowledgements 165
References 165
Chapter 7.
Artificial neural network models for timely assessment of trauma complication risk
R.P Marble and J.C. Healy
1 Artificial neural network models 173
1.1 Background 173
1.2 Neural networks and statistical analysis 176
1.3 Neural networks in medicine 178
2 A neural network model for predicting the incidence of coagulopathy in victims of blunt injury trauma 180
2.1 Model description 181
2.2 Results 182
2.3 Remarks 184
3 Prospects for refining and utilizing neural models in trauma care settings 186
3.1 Sensitivity analysis, pruning, and rule extraction 186
3.2 Trauma Systems development 188
References 189
Chapter 8.
Artificial neural networks in medical diagnosis
Y. Fukuoka
1 Introduction 197
2 Foundations of artificial neural networks 198
2.1 Artificial neuron 198
2.2 Network architectures 198
2.3 Learning algorithms 199
2.3.1 Back-propagation 200
2.3.2 Self-organizing map 203
3 Applications to biomedicine 205
3.1 Pattern classification with BP 206
3.1.1 Clinical data 207
3.1.2 Bioelectric signals 208
3.1.3 Image analysis 210
3.2 Pattern classification with SOM 211
3.3 Data compression with BP 213
3.4 System modeling with BP 214
3.5 More detailed reviews 216
3 5.1 Chronic stress evaluation using ANNs 216
3.5.2 Gene expression data analysis with SOM 2 19
4 Conclusion 221
References 222
Chapter 9.
The application of neural networks in the classification of the electrocardiogram
C.D Nugent, J.A. Lopez, N.D. Black, and J.A.C. Webb
1 Introduction to the classification of the electrocardiogram 229
1.1 Diagnostic utilities of the ECG 230
1.2 Introduction to computerized classification 231
2 Fundamentals of the 12-lead ECG 233
2.1 The 12-lead ECG and associated nomenclature 234
3 Computerized classification of the 12-lead ECG 237
3.1 Classification 239
4 Neural networks in 12-lead ECG classification 241
4.1 The artificial neuron 241
4.2 The MLP and ECG classification 244
5 Summary 251
References 254
Chapter 10.
Neural network predictions of significant coronary artery stenosis in women
B.A. Mobley, W.E. Moore, E. Schechter, J.E. Eichner, and P.A. McKee
1 Introduction 262
1.1 Systems enabling the avoidance of unnecessary angiography 262
1.2 Women and angiography 263
1.3 Other clinical predictions by neural network 264
2 Methods 265
2.1 Development of the data set from the SCA&I database 265
2.2 Artificial neural network 270
2.3 Patient files 272
2.4 Logistic regression 274
2.5 ROC analysis 275
3 Results 276
3.1 Neural network training and cross validation 276
3.2 Network application to the cutoff determination file 276
3.3 Network application to the test file 278
3.4 Relative weights of the neural network 281
3.5 Logistic regression 282
4 Discussion 285
4.1 Patients and data 285
4.2 Patient files 286
4.3 Cutoff determination file 287
4.4 Predictive systems 287
4.5 Network weights 287
5 Conclusions 288
Acknowledgments 288
References 289
Chapter 11.
A modular neural network system for the analysis of nuclei 

in histopathological sections
C.S. Pattichis, F. Schnorrenberg, C.N. Schizas, M.S. Pattichis, and K. Kyriacou
1 Introduction 292
1.1 The need of quantitative analysis in diagnostic histopathology 292
1.2 A brief overview of the use of artificial neural network (ANN) systems in diagnostic histopathology 293
1.3 Quantitative analysis in immunocytochemistry 293
2 Material 294
3 Modular neural network system 295
3.1 Detection of nuclei: the receptive field-squashing function (RFS) module 295
3.1.1 Step 1: convert color image to optical density image 297
3.1.2 Step 2: compute the receptive field filter 297
3.1.3 Step 3: apply iteratively the receptive field and the squashing function 297
3.1.4 Step 4: threshold bimodal histogram 298
3.1.5 Step 5: revise the list of detected nuclei 298
3.2 Detection of nuclei: the feedforward neural network (FNN) module 299
3.2.1 Step 1: color image to ptical density image conversion 300
3.2.2 Step 2: histogram stretching and thresholding 300
3.2.3 Step 3: SV expansion and feedforward neural network identification of image blocks 301
3.2.4 Step 4: calculation of the exact nuclei locations 301
3.3 Combination of detection modules 301
3.4 Nuclei classification and diagnostic index calculation 302
3.4.1 Step 1: extract features for each nucleus 303
3.4.2 Step 2: classify each nucleus 304
3.4.3 Step 3: compute diagnostic index 304
3.5 System Validation 304
4 Results 306
4.1 Detection example 306
4.2 ROC analysis 308
4.3 Classification and diagnostic index computation module 310
5 Discussion 312
6 Future work 314
References 315
Appendix A: Semi-quantitative diagnostic index 321
Chapter 12.
Septic shock diagnosis by neural networks and rule based systems
R. Brause, F. Hamker, and J. Paetz
1 Introduction 323
2 The data 325
2.1 The data context 326
2.2 Data problems and preprocessing 326
2.3 Selecting feature variables 328
2.4 Basic statistical analysis 329
3 The neural network approach to diagnosis 331
3.1 The network 331
3.1.1 The network architecture 332
3.1.2 Treatment of missing values 333
3.2 Training and diagnosis 334
3.2.1 The training and test performance 334
3.2.2 The problem of medical data partition 335
3.3 Selection and validation of a neural network 336
3.4 Results for septic shock diagnosis 338
4 The neuro-fuzzy approach to rule generation 339
4.1 The rule extraction network 340
4.2 Application to septic shock patient data 343
5 Conclusions and discussion 345
Acknowledgments 348
References 348
Appendix A: The network adaptation and growing 352
Adaptation of the layers 352
Growing of the representation layer 353
Appendix B: The main rule building algorithm 354
Appendix C: The rule shrinking procedure 355
Chapter 13.
Monitoring depth of anesthesia
W. Huang, X.-S. Zhang, and R.J. Roy
1 Introduction 357
2 Computational intelligente (CI) for DOA 359
2.1 Fuzzy logic assessment 360
2.1.1 Fuzzy inference process 361
2.1.2 Why not fuzzy? 363
2.2 Artificial neural networks 363
2.3 Neuro-fuzzy modeling 366
3 ANN-based CI model for MLAEP 368
3.1 MLAEP-derived parameter extraction 369
3.1.1 Wavelet transformation 370
3.2 System design based on ANN for MLAEP 372
3.3 ANN System: experiment results 373
4 Neuro-fuzzy based CI model for EEG 375
4.1 EEG-derived parameter extraction 376
4.1.1 Complexity analysis 376
4.1.2 Regularity analysis 376
4.1.3 Spectral entropy analysis 377
4.2 ANFIS - "derived" fuzzy knowledge model 377
4.3 System design based on ANFIS for EEG 379
4.4 ANFIS system: experiment results 380
5 Discussions 382
5.1 ANN versus ANFIS 382
5.2 EEG versus MLAEP 383
5.3 Performance issues 384
Acknowledgments 385
References 386
Chapter 14.
Combining evolutionary and fuzzy techniques in medical diagnosis
C.A. Pea-Reyes and M. Sipper
1 Introduction 391
2 Background 392
2.1 Fuzzy modeling 392
2.2 Evolutionary computation 396
2.3 Evolutionary fuzzy modeling 400
3 Fuzzy Systems for breast cancer diagnosis 403
3.1 The WBCD Problem 403
3.2 Fuzzy-system setup 405
4 A fuzzy-genetic approach 407
4.1 The evolutionary setup 407
4.2 Results 408
5 A fuzzy coevolutionary approach: fuzzy CoCo 411
5.1 Cooperative coevolution 411
5.2 The coevolutionary algorithm 413
5.3 The evolutionary setup 416
5.4 Results 419
6 Concluding remarks 421
References 422
Chapter 15.
Genetic algorithms for feature selection in computer-aided diagnosis
B. Sahiner, H.P. Chan, and N. Petrick
1 Introduction 427
2 Genetic algorithms 429
2.1 Encoding 430
2.2 Initial population 431
2.3 Fitness function 431
2.4 Genetic operators 431
2.5 Working parameters 431
3 Feature selection and GAS 432
4 Applications in CAD 434
4.1 Classification of malignant and benign microcalcifications 436
4.1.1 Feature extraction 437
4.1.2 Data set 437
4.1.3 Morphological feature space 438


4.1.4 Texture feature space 440
4.1.5 GA implementation 442
4.1.6 Classification 444
4.1.7 Results 445
4.1.8 Discussion 452
4.2 Classification of mass and normal breast tissue 453
4.2.1 Data set 453
4.2.2 Morphological features 455
4.2.3 Texture features 455
4.2.4 Classification 455
4.2.5 GA implementation 456
4.2.6 Results 457
4.2.7 Discussion 461
4.3 Classification of malignant and benign masses 463
4.3.1 Data set 464
4.3.2 Image transformation 464
4.3.3 Texture features 466
4.3.4 Classification 467
4.3.5 GA implementation 467
4.3.6 Results 469
4.3.7 Discussion 474
5 Conclusion 475
References 476
Index 485
List of contributors 489
END
