Performance of a Neural Network Trained to Make Third-molar Treatment-planning Decisions

The authors developed and tested 12 neural networks of different architectures to make lower-third-molar treatment-planning decisions, using a software-based neural network (Neudesk 1.2, Neural Computer Sciences, Southampton, UK). Network train ing was undertaken using clinical histories from 119 patients (with 238 lower third molars) referred for treatment planning (79 females and 40 males, mean age 25 years) together with output data consisting of actual treatments planned by a senior oral surgeon. Both the input clinical data and the consultant decisions were treated on a tooth-wise basis and were coded to numerical values. Binary data (e.g., present/ab sent) were coded to 1 and 0, while quantitative data (e.g., age) were scaled to fall between 0 and 1. A network based on the optimal architecture was trained and then interrogated with test data derived from a further 174 patients (119 females and 55 males, mean age 26 years) with 348 lower third molars. Network decisions were di chotomized with a threshold of 0.8. With no knowledge of the network decisions, the senior oral surgeon indicated his preferred treatments. The teeth were then assigned to "gold-standard" categories of indications present or absent based on National In stitutes of Health consensus criteria. Against this, the network achieved a sensitivity of 0.78, which was slightly inferior to that of the oral surgeon (0.88), although this difference was not significant, and a specificity of 0.98, compared with 0.99 for the oral surgeon (p = NS). Agreement between the oral surgeon and network decisions was very high (kappa = 0.850). This study demonstrates that it is possible to train a neural network to provide reliable decision support for lower-third-molar treatment planning. Key words: Decision support; neural network; third molars. (Med Decis Making 1996; 16:153-160)

[1]  J Shepherd,et al.  Trends in oral surgery practice , 1987, British Dental Journal.

[2]  C E Floyd,et al.  Artificial Neural Networks for Single Photon Emission Computed Tomography: A Study of Cold Lesion Detection and Localization , 1993, Investigative radiology.

[3]  P. Wilding,et al.  Application of neural networks to the interpretation of laboratory data in cancer diagnosis. , 1992, Clinical chemistry.

[4]  G. Pfurtscheller,et al.  Prediction of the side of hand movements from single-trial multi-channel EEG data using neural networks. , 1992, Electroencephalography and clinical neurophysiology.

[5]  NIH consensus development conference for removal of third molars. , 1980, Journal of oral surgery.

[6]  Bob Kemp.  Cerebral information processing estimated by unpredictability of the EEG , 1992, Clinical Neurology and Neurosurgery.

[7]  G. W. Gross,et al.  Neural networks in radiologic diagnosis. II. Interpretation of neonatal chest radiographs. , 1990, Investigative radiology.

[8]  J H Carter.  Myocardial infarction prediction by artificial neural networks. , 1992, Annals of internal medicine.

[9]  J V Tu,et al.  Use of a neural network as a predictive instrument for length of stay in the intensive care unit following cardiac surgery. , 1992, Proceedings. Symposium on Computer Applications in Medical Care.

[10]  Y. Wu,et al.  Artificial neural networks in mammography: application to decision making in the diagnosis of breast cancer. , 1993, Radiology.

[11]  G W Humphreys,et al.  Lesioning a connectionist model of visual search: selective effects on distractor grouping. , 1992, Canadian journal of psychology.

[12]  R M Davies,et al.  Use of panoramic radiographs in general dental practice in England. , 1986, Community dentistry and oral epidemiology.

[13]  D L Hudson,et al.  Neural network approach to detection of metastatic melanoma from chromatographic analysis of urine. , 1991, Proceedings. Symposium on Computer Applications in Medical Care.

[14]  B A Huberman,et al.  An associationist model of the paranoid process: application of phase transitions in spreading activation networks. , 1992, Psychiatry.

[15]  J R Carrie,et al.  Evaluation of a Neural Network Model of Amnesia in Diffuse Cerebral Atrophy , 1993, British Journal of Psychiatry.

[16]  A Manduca,et al.  Improvement in specificity of ultrasonography for diagnosis of breast tumors by means of artificial intelligence. , 1992, Medical physics.

[17]  James L. McClelland,et al.  Parallel distributed processing: explorations in the microstructure of cognition, vol. 1: foundations , 1986 .

[18]  Y. Kosugi,et al.  Neural Network Mapping for Nonlinear Stereotactic Normalization of Brain MR Images , 1993, Journal of computer assisted tomography.

[19]  J. S. Prater,et al.  Segmenting ultrasound images of the prostate using neural networks. , 1992, Ultrasonic imaging.

[20]  H C Kwan,et al.  Neural Networks and Parkinson's Disease , 1993, Canadian Journal of Neurological Sciences / Journal Canadien des Sciences Neurologiques.

[21]  P W Macfarlane,et al.  Neural networks for classification of ECG ST-T segments. , 1992, Journal of electrocardiology.

[22]  A Murray,et al.  Development of a neural network screening aid for diagnosing lower limb peripheral vascular disease from photoelectric plethysmography pulse waveforms. , 1993, Physiological measurement.

[23]  P. Wilding,et al.  The application of backpropagation neural networks to problems in pathology and laboratory medicine. , 1992, Archives of pathology & laboratory medicine.

[24]  J. Kruschke,et al.  ALCOVE: an exemplar-based connectionist model of category learning. , 1992, Psychological review.

[25]  P D Neilson,et al.  The problem of redundancy in movement control: The adaptive model theory approach , 1993, Psychological research.

[26]  J. Cohen,et al.  Context, cortex, and dopamine: a connectionist approach to behavior and biology in schizophrenia. , 1992, Psychological review.

[27]  C F Kugelberg,et al.  Third molar surgery. , 1992, Current opinion in dentistry.

[28]  J Shepherd,et al.  Comparison of clinical treatment decisions with US National Institutes of Health consensus indications for lower third molar removal , 1993, British Dental Journal.

[29]  D A Rosenbaum,et al.  A model for reaching control. , 1993, Acta psychologica.

[30]  A P Zelicoff.  Myocardial infarction prediction by artificial neural networks. , 1992, Annals of internal medicine.

[31]  L. Mango.  Computer-assisted cervical cancer screening using neural networks. , 1994, Cancer Letters.

[32]  M L Astion,et al.  Application of neural networks to the classification of giant cell arteritis. , 1994, Arthritis and rheumatism.

[33]  E. Somoza,et al.  A Neural-network Approach to Predicting Admission Decisions in a Psychiatric Emergency Room , 1993, Medical decision making : an international journal of the Society for Medical Decision Making.

[34]  J. S. Ostrem,et al.  Application of Neural Nets to Ultrasound Tissue Characterization , 1991 .

[35]  P Nickolls,et al.  Implantable Cardioverter Defibrillator Electrogram Recognition with a Multilayer Perceptron , 1993, Pacing and clinical electrophysiology : PACE.

[36]  P W Jamieson.  A computational model of levodopa pharmacodynamics in Parkinson's disease. , 1991, Clinical neuropharmacology.

[37]  J. Kippenhan,et al.  Evaluation of a neural-network classifier for PET scans of normal and Alzheimer's disease subjects. , 1992, Journal of nuclear medicine : official publication, Society of Nuclear Medicine.

[38]  P W Macfarlane,et al.  Classification of electrocardiographic ST-T segments--human expert vs artificial neural network. , 1993, European heart journal.

[39]  S L Swihart,et al.  Classification of chromatic visual evoked potentials with the aid of a neural net. , 1992, Computers in biology and medicine.

[40]  J P Macher,et al.  Neural network model: application to automatic analysis of human sleep. , 1993, Computers and biomedical research, an international journal.