人工智能在前列腺癌病理诊断及分子分型中的研究进展

Artificial intelligence in pathological diagnosis and molecular typing of prostate cancer:research progress

人工智能(AI)在前列腺癌(PCa)病理诊断、影像学诊断、预后预测、分子分型等方面具有重要意义和远大前景。本文主要关注AI分析病理切片在PCa病理诊断及分子分型中的应用进展,简要介绍了AI在穿刺病理诊断和Gleason分级、切除术后病理的诊断和分级、基于病理切片预测PCa患者预后中的应用。在穿刺病理诊断和Gleason分级方面,AI已经和普通病理医师表现不相上下;在切除术后病理的诊断和分级方面,AI可以对肿瘤进行精准分级与评分;在PCa患者预后预测方面,AI可以直接从病理组织切片中提取相关预后信息,预测PCa患者的术后情况。此外,AI还可以预测PCa患者的基因突变,通过分析病理切片得出基因突变的概率。

Artificial intelligence(AI)has important significance and great promise in the pathological diagnosis,imaging diagnosis,prognosis prediction,and molecular subtyping of prostate cancer(PCa).This review focuses on the progress of AI for the diagnosis and molecular classification of PCa,and briefly introduces the application of AI in the pathological diagnosis of needle biopsy and Gleason grading,pathological diagnosis and grading after prostatectomy,and prognosis prediction of PCa patients based on pathological sections.For the pathological diagnosis of needle biopsy and Gleason grading,AI has already comparable to general pathologists;for the pathological diagnosis and grading after prostatectomy,AI can accurately grade and classify tumors;and for the prognosis prediction of PCa patients,AI can directly extract relevant prognostic information from pathological tissue sections for prognosis prediction.In addition,AI can also predict gene mutations in PCa patients and suggest the probability of gene mutation by analyzing the pathological sections.