A two-week clinical trial screening chore now takes seconds at one Korean hospital. The software comes from Seoul St. Mary’s Blood Cancer Center, where it automatically matches myeloma patients to studies whose criteria the system has loaded.
Deployment at the Multiple Myeloma Center was announced September 2. Records are first merged into a per-patient database covering treatment counts, drug refractoriness and stem cell transplant history. A large language model converts each trial’s inclusion and exclusion rules into a comparable format and assigns patients to subgroups. Latest lab values, from kidney and liver function to blood counts, are then matched against those rules.
The refractoriness math is handled automatically. Responses across proteasome inhibitors, immunomodulators and anti-CD38 antibodies are tallied to establish double and triple refractoriness. Records missing fresh lab values are parked as pending rather than rejected, and patients who have died are removed from consideration. Each patient ends up labeled eligible, pending or ineligible, with supporting evidence shown.
Professor Park Sung-soo led the hematology team that built the software in about two months without outside funding. A patent application and program copyright are on file. Core logic has passed more than 800 validation tests, criteria for 18 trials are registered, and several thousand patients can be screened in tens of seconds.
Few blood cancers see new therapies arrive as quickly as multiple myeloma. Even so, getting patients into those studies has relied on slow manual matching, according to professor Min Chang-ki. The group now plans a clinical accuracy validation study and an expansion to other blood cancers.
