Case-based reasoning (CBR)-driven medical diagnostic systems demand a critical mass of up-to-date diagnostic-quality cases that depict the problem-solving methodology of medical experts. In practical terms, procurement of CBR-compliant cases is quite challenging, as this requires medical experts to map their experiential knowledge to an unfamiliar computational formalism. In this paper, we propose a novel medical knowledge acquisition approach that leverages routinely generated electronic medical records (EMRs) as an alternate source for CBR-compliant cases. We present a methodology to autonomously transform XML-based EMR to specialized CBR-compliant cases for CBR-driven medical diagnostic systems. Our multi-stage methodology features: (a) collection of heterogeneous EMR from Internet-accessible EMR repositories via intelligent agents, (b) automated transformation of both the structure and content of generic EMR to specialized CBR-compliant cases, and (c) inductive estimation of the weight of each case-defining attribute. The computational implementation of our methodology is presented as case acquisition and transcription info-structure (CATI).