毒性病理學是臨床前藥物安全性評價的重要組成部分。隨著計算機科學的發(fā)展和全切片數(shù)字掃描切片技術的發(fā)展，人工智能（AI）已經(jīng)廣泛應用于藥物安全性評價領域，也包括病理學的各個方面，如診斷病理學、獸醫(yī)診斷學、病理學研究、監(jiān)管毒理學和病理學初級閱片及同行評議等。AI病理學平臺的應用逐漸影響精準醫(yī)學和個性化醫(yī)學的未來發(fā)展。然而，與其他重要的科學技術進步一樣，AI在毒性病理學領域的實施和應用也面臨巨大的挑戰(zhàn)。通過簡要綜述數(shù)字病理學的發(fā)展、算法類型、AI在毒性病理學中的應用概況、AI的監(jiān)管環(huán)境等，以期為AI在毒性病理學中的研究發(fā)展及廣泛應用提供新的思路。

Toxicologic pathology is an important part of preclinical safety evaluation of drugs. With the development of computer science and whole slide image scan technology, artificial intelligence (AI) has widely expanded into the field of safety evaluation of drugs, including all aspects of pathology, such as diagnostics pathology, veterinary diagnostics, pathological research, regulatory toxicology and pathology primary reads and peer review. The application of AI pathology platforms is now poised to truly impact the future of precision and personalized medicine. However, as with other critical technology advances, there are great challenges in the AI implementation and application of the toxicologic pathology. The development of digital pathology, the types of algorithms, the application of AI in toxicological pathology, and the regulatory environment of AI are briefly reviewed, so as to provide new ideas for the research, development and wide application of AI in toxicologic pathology.

中國食品藥品檢定研究院學科帶頭人課題項目（2021X2）