In this paper we present a self-integrating knowledge-based expert sys
tem for brain tumor diagnosis. The system we propose comprises knowled
ge building, knowledge inference and knowledge refinement. During know
ledge building, an automatic knowledge-integration process, based on D
arwin's theory of natural selection, integrates knowledge derived from
knowledge-acquisition tools and machine-learning methods to construct
an initial knowledge base, thus eliminating a major bottleneck in dev
eloping a brain tumor diagnostic system. During the knowledge inferenc
e process, art inference engine exploits rules in the knowledge base t
o help diagnosticians determine brain tumor etiologies according to co
mputer tomography pictures. And, a simple knowledge refinement method
is proposed to modify the existing knowledge base during inference, wh
ich dramatically improves the accuracy of the derived rules. The perfo
rmance of the brain tumor diagnostic system has been evaluated on actu
al brain tumor cases. Copyright (C) 1996 Elsevier Science Ltd