Optimizing Rare Disease Patient Matching with Large Language Models

Berger, ArminArminBergerBashir, Ali HamzaAli HamzaBashirBerghaus, DavidDavidBerghausMowmita, Afsan NaziaAfsan NaziaMowmitaGrigull, LorenzLorenzGrigullFendrich, LaraLaraFendrichHögl, HenrietteHenrietteHöglErnst, GundulaGundulaErnstSchmidt, RalfRalfSchmidtBascom, DavidDavidBascomLagones, Tom AnglimTom AnglimLagonesDeußer, TobiasTobiasDeußerBell Felix de Oliveira, ThiagoThiagoBell Felix de OliveiraLübbering, MaxMaxLübberingSifa, RafetRafetSifa2025-04-142025-04-252025-04-142024https://publica.fraunhofer.de/handle/publica/48659910.1109/BigData62323.2024.109101132-s2.0-105000217140We present RepLLaMA, a neural ranking model for optimizing patient matching in rare disease communities. Using data from Unrare.me consisting of over two thousand profiles and over ten thousand ratings, our bi-encoder architecture maps profiles to 4096-dimensional vectors, enabling efficient similarity computations. The system processes unstructured symptom descriptions and structured responses, incorporating expert-guided LLM enhancements. Results show Top-10 Recall of 49.36% (±2.03), surpassing baselines while maintaining generalization. The implementation provides a scalable solution for rare disease patient matching, addressing computational complexity challenges.enfalseLarge Language ModelsRare DiseasesRecommender SystemsText EmbeddingsText MatchingOptimizing Rare Disease Patient Matching with Large Language Modelsconference paper