This paper presents a zero-shot large language model (LLM)-based system for matching patients to clinical trials, addressing the significant challenge of patient recruitment in drug development. The system evaluates whether a patient meets a set of trial inclusion criteria, specified as free text, using unstructured clinical notes. The authors investigate different prompting strategies and design a two-stage retrieval pipeline to reduce token processing while maintaining high performance. The system achieves state-of-the-art performance on the n2c2 2018 cohort selection challenge, the largest public benchmark for clinical trial patient matching. It also demonstrates improved data and cost efficiency, reducing the time and cost of patient matching by an order of magnitude compared to current methods. Additionally, the system can generate coherent explanations for its decisions, with clinicians agreeing with 97% of correct and 75% of incorrect decisions. The paper discusses the limitations and future directions, highlighting the potential of LLMs in accelerating clinical trial operations.This paper presents a zero-shot large language model (LLM)-based system for matching patients to clinical trials, addressing the significant challenge of patient recruitment in drug development. The system evaluates whether a patient meets a set of trial inclusion criteria, specified as free text, using unstructured clinical notes. The authors investigate different prompting strategies and design a two-stage retrieval pipeline to reduce token processing while maintaining high performance. The system achieves state-of-the-art performance on the n2c2 2018 cohort selection challenge, the largest public benchmark for clinical trial patient matching. It also demonstrates improved data and cost efficiency, reducing the time and cost of patient matching by an order of magnitude compared to current methods. Additionally, the system can generate coherent explanations for its decisions, with clinicians agreeing with 97% of correct and 75% of incorrect decisions. The paper discusses the limitations and future directions, highlighting the potential of LLMs in accelerating clinical trial operations.