US-based clinical AI platform Mendel has unveiled positive results for its Neuro-Symbolic AI system from the latest research in Automatic Cohort Retrieval (ACR).
The study compared the performance of Retrieval Augmented Generation (RAG) and LLM-based solutions and Mendel’s neuro-symbolic systems on a 1.4K patient data set.
In the study, Mendel’s clinical AI outperformed GPT-4 in the automated identification of patient cohorts from unstructured and structured EMRs in several benchmarks.
The company’s unique clinical AI approach combines large language models (LLMs) with its in-house hypergraph reasoning engine.
The research results indicate that its clinical AI can power significant advancements in ACR, a fundamental task for clinical research and patient care, said the clinical AI company.
Mendel cofounder and chief science officer Wael Salloum said: “Our latest research at Mendel marks a significant milestone in the field of AI in general, and healthcare in particular.
“We are the leader in clinical reasoning by coupling LLMs with our hypergraph reasoning, enhancing both the effectiveness and efficiency of patient cohort retrieval.
“This work is critical in paving the way for more robust and scalable clinical reasoning. This breakthrough underscores our commitment to advance the AI field to transform clinical research and improve patient outcomes.”
According to the company, identifying patient cohorts is crucial for clinical trials, retrospective studies, and other healthcare applications.
Traditional methods involve automated queries of structured data, together with manual curation, which are time-consuming and often yield low-quality results.
Mendel said that its AI offerings leverage a unique approach that combines advanced clinical LLM trained to understand structured and unstructured text.
They also use an in-house reasoning engine infused with medical knowledge, which is reviewed by medical professionals, to apply a clinician’s mind to complex medical situations.
The capability to apply clinical reasoning to ACR has been shown to offer significant improvements over existing RAG and LLM techniques, said Mendel.
