肿瘤数字病理图像中关键特征信息的提取与分析  

Extraction and analysis of key feature information in pathology images

作　　者：熊杰 方红 伍棋 甘仲霖 XIONG Jie;FANG Hong;WU Qi;GAN Zhonglin(School of Public Health,Southwest Medical University,Luzhou 646000,China;Department of Pathology,Luzhou People's Hospital of Sichuan Province,Luzhou 646000,China;Department of Pathology,Affiliated Hospital of Southwest Medical University,Luzhou 646000,China)

机构地区：[1]西南医科大学公共卫生学院,四川泸州646000 [2]四川省泸州市人民医院病理科,四川泸州646000 [3]西南医科大学附属医院病理科,四川泸州646000

出　　处：《现代医学》2024年第12期1942-1949,共8页Modern Medical Journal

摘　　要：目的:探讨在数字病理图像中,提取细胞核特征与图像全部特征相比,是否更具有代表性;并分析两者在肿瘤的早期诊断和预后预测中的可行性和有效性,为肿瘤的诊断和预防提供新的策略。方法:收集2011年3月至2015年3月西南医科大学附属医院诊断的374例乳腺癌病例样本。提取其病理图像特征,并研究其与肿瘤病理类型、组织学分级、复发、转移及生存状态等预后因素的关系。结果:分割细胞核提取的特征在分形维数稳定性及抗干扰能力方面优于阈值分割法,差异有统计学意义(P<0.05)。在图像像素1024×1024、放大倍数63×时,分割效果最好。分割细胞核提取的特征在肿瘤的WHO分期(AUC=0.550,F 1=0.690)和转移识别(AUC=0.600,F 1=0.890)中的辨识度更高,而在复发(AUC=0.610,F 1=0.930)和生存状态(AUC=0.710,F 1=0.790)方面,两种方法的辨识度相当。结论:提取细胞核特征较传统的阈值分割方法,具有更高代表性和抗干扰能力,在乳腺癌诊断和预后预测中表现出更高识别度。该法可作为肿瘤早期诊断和预后预测的可靠指标,具有重要的临床应用价值。Objective:To explore whether extracting nuclear features from digital pathology images is more representative than extracting whole-image features and evalnates the effectiveness of both methods in the early diagnosis and prognosis prediction of tumors,providing new strategies for tumor diagnosis and prevention.Methods:A total of 374 breast cancer cases diagnosed between March 2011 and March 2015 at the Affiliated Hospital of Southwest Medical University were collected.Pathological image features were extracted and analyzed in relation to prognostic factors such as tumor pathology type,histological grade,recurrence,metastasis,and survival status.Results:Nuclear features extraction demonstrated better fractal dimension stability and greater robustness to noise than threshold-based segmentation,with significant differences(P<0.05).The best segmentation performance was achieved with an image size of 1024×1024 and a magnification of 63×.Nuclear features also demonstrated higher accuracy in tumor WHO staging(AUC=0.550,F 1=0.690)and metastasis identification(AUC=0.600,F 1=0.890),while both methods performed similarly in tumor recurrence(AUC=0.610,F 1=0.930)and survival status prediction(AUC=0.710,F 1=0.790).Conclusion:Nuclear feature extraction demonstrates greater representativeness of relevant features and stronger noise resistance compared to traditional threshold segmentation methods.It exhibits higher diagnostic accuracy in breast cancer diagnosis and prognosis prediction,establishing it as a reliable indicator with significant clinical application value.

关 键 词：数字病理图像 分形维数 特征提取 机器学习 

分 类 号：R730[医药卫生—肿瘤]