Study on the Construction and Application of a Large Model Corpus for Pancreatic Cancer Based on Real Clinical Cases

Combining retrospective observational research with artificial intelligence data engineering. Collecting historical electronic medical record data, and through data governance and annotation by artificial medical experts, constructing a structured knowledge base.

Inclusion Criteria: 1. Real inpatient cases of pancreatic cancer diagnosis and treatment; 2. Prioritize the inclusion of cases with MDT discussion records, difficult and complex cases, and fatal cases; 3. Possess complete electronic medical information (including but not limited to: outpatient medical records, inpatient medical records including admission records, progress notes, surgery/procedure records, nursing records, examination and test reports, doctor's orders, discharge/death records, etc.); - Exclusion Criteria: 1. Cases that only underwent a brief transition in emergency/outpatient departments and did not form a closed loop of diagnosis and treatment; 2. Cases with incomplete diagnosis and treatment information and missing key indicators.