A comparative study of large language model-based zero-shot inference and task-specific supervised classification of breast cancer pathology reports.

Objective: Although supervised machine learning is popular for information extraction from clinical notes, creating large annotated datasets requires extensive domain expertise and is time-consuming. Meanwhile, large language models (LLMs) have demonstrated promising transfer learning capability. In this study, we explored whether recent LLMs could reduce the need for large-scale data annotations.

A comparative study of zero-shot inference with large language models and supervised modeling in breast cancer pathology classification.

Although supervised machine learning is popular for information extraction from clinical notes, creating large, annotated datasets requires extensive domain expertise and is time-consuming. Meanwhile, large language models (LLMs) have demonstrated promising transfer learning capability. In this study, we explored whether recent LLMs can reduce the need for large-scale data annotations.

Cinchona Alkaloid-Inspired Urea-Containing Autophagy Inhibitor Shows Single-Agent Anticancer Efficacy.

Autophagy is upregulated in response to metabolic stress, a hypoxic tumor microenvironment, and therapeutic stress in various cancers and mediates tumor progression and resistance to cancer therapy. Herein, we identified a cinchona alkaloid derivative containing urea (), which exhibited potential cytotoxicity and inhibited autophagy in hepatocellular carcinoma (HCC) cells. We showed that not only induced apoptosis but also blocked autophagy in HCC cells, as indicated by the increased expression of LC3-II and p62, inhibition of autophagosome-lysosome fusion, and suppression of the Akt/mTOR/S6k pathway in the HCC cells.