磁共振图像利用深度学习进行腮腺肿瘤病理亚型识别的可行性分析  

The feasibility study of using deep learning to identify the pathological subtypes of parotid gland tumors based on MR images

作　　者：王艳[1] 田慧[1] 洪悦[1] 李辉[1] WANG Yan;TIAN Hui;HONG Yue;LI Hui(Department of Radiology and Medical Imaging,People's Hospital of Xinjiang Uygur Autonomous Region,Urumqi 830001,China)

机构地区：[1]新疆维吾尔自治区人民医院放射影像中心,新疆乌鲁木齐830001

出　　处：《医学影像学杂志》2024年第7期9-12,共4页Journal of Medical Imaging

基　　金：新疆维吾尔自治区自然科学基金项目(编号:2019D01C114)。

摘　　要：目的 探讨基于MRI T_(2)WI、DWI、增强T_(1)WI的多模态图像,利用深度学习自动识别腮腺肿瘤病理类型,通过与病理结果对照评判机器学习的效能。方法 选取经病理证实的39例腮腺肿瘤患者,其中多形性腺瘤9例,Warthin瘤13例,恶性肿瘤6例,其他非肿瘤病变11例。MRI经标准化处理后通过2D U-NET网络对肿瘤类型进行鉴别,对比单通道T_(2)图像、多通道T_(2)、DWI和增强T_(1)图像输入对肿瘤亚型的分类效果。其中29例的434层图像为训练集,10例的156层图像为测试集。采用准确率、敏感度、特异度、F1分数、准确度等指标评估病理分类效能。结果 采用T_(2)单通道输入训练时,对Warthin瘤的鉴别具有最高的F1分数59.2%、敏感度为51.1%和准确度为70.3%,多形性腺瘤次之。通过常规T_(2)、DWI和增强T_(1)图像对Warthin瘤鉴别的敏感度、准确度、F1分数分别为79.5%、70.0%、74.5%,而判断多形性腺瘤的敏感度、特异度分别为46.2%、81.5%。结论 通过腮腺肿瘤MRI,利用深度学习能有效识别Warthin瘤,多模态图像结合能提高Warthin瘤、多形性腺瘤识别的敏感度。Objective To automatically identify pathological types of parotid gland tumors using deep learning methods based on multiple parameters of MR T_(2),DWI and enhanced T_(1) images,and to evaluate the efficacy of deep learning by compar‐ing with pathological results.Methods MRI images of 39 cases of parotid gland tumors confirmed by pathology,including 9 cases of pleomorphic adenoma,13 cases of Warthin tumor,6 cases of malignant tumor,and 11 cases of other non-tumor lesions,were selected in this work.After MR images were standardized,2D U-NET was used to identify tumor types,and the classifica‐tion efficiency through the input of single-channel of T_(2) image,and multi-channel of T_(2),DWI and enhanced T_(1) image were com‐pared respectively.The data of 29 cases were used for training sets and the data of 10 cases for test sets.Accuracy,sensitivity,specificity,F1 score and precision were calculated to evaluate the efficacy of tumor classification.Results When training with T_(2) single-channel input,Warthin tumor had the highest F1 score(59.2%),sensitivity(51.1%)and precision(70.3%),followed by pleomorphic adenoma.The sensitivity,accuracy and F1 scores of multi-channel input for Warthin tumor identification were 79.5%,70.0%and 74.5%,respectively,the sensitivity and specificity of polymorphic adenoma were 46.2%and 81.5%,respec‐tively.Conclusion By analyzing the imaging characteristics of parotid gland tumors,the deep learning method can effectively identify Warthin tumor,and the combination of multi-parametric MR images can improve the sensitivity of Warthin tumor and pleomorphic adenoma recognition.

关 键 词：头颈部 腮腺肿瘤 深度学习 磁共振成像 

分 类 号：R739.87[医药卫生—肿瘤] R445.2[医药卫生—临床医学]