摘要
目的构建基于卷积神经网络的甲状腺液基细胞学病理辅助诊断模型。方法收集700例甲状腺液基细胞学涂片,扫描成数字图像,经良、恶性标注后按比例分成训练集和测试集,噪声滤除后提取有效区域分别在10×和40×分割成512×512的小图像块,构建分类模型对训练集深度学习并对测试集自动判读,经数据增强和参数迭代优化,统计辅助诊断模型的准确率、灵敏度、特异度、阳性预测值和阴性预测值等评价指标。结果训练集560例含4 926个细胞团的11 164个图像块,测试集140例含977个细胞团的1 402个图像块,选取YOLO网络构建细胞团检测模型,用ResNet50作为分类模型,经过40轮迭代训练,10×倍率下准确率为90.01%,灵敏度89.31%,特异度92.51%,阳性预测值97.70%,阴性预测值70.82%,曲线下面积高达0.97,平均判读时间不足1 s。40×时虽极为灵敏(98.72%)但特异性较差,提示10×放大倍率下辅助诊断模型更为可靠。结论该辅助诊断模型与病理医师水平基本相当,且诊断效率远远超出。可大大提高阅片一致性和效率,降低漏诊率。未来可继续扩大样本量获取更多病变形态,提高准确率,达到临床应用水平。
Objective To develop a convolutional neural network based model for assisting pathological diagnoses on thyroid liquid-based cytology specimens.Methods Seven-hundred thyroid TCT slides were collected,scanned for whole slide imaging(WSI),and divided into training and test sets after labeling the correct diagnosis(benign versus malignant).The extracted regions of interest after noise filtering were cropped into pieces of 512×512 patch on 10×and 40×magnifications,respectively.A classification model was constructed using deeply learning algorithms,and applied to the training set,then automatically tuned in the test set.After data enhancement and parameters optimization,accuracy,sensitivity,specificity,positive predictive value and negative predictive value of the model were calculated.Results The training set with 560 WSI contained 4926 cell clusters(11164 patches),while the test set with 140 WSI contained 977 cell clusters(1402 patches).YOLO network was selected to establish a detection model,and ResNet50 was used as a classification model.With 40 epochs training,results from 10×magnifications showed an accuracy of 90.01%,sensitivity of 89.31%,specificity of 92.51%,positive predictive value of 97.70%and negative predictive value of 70.82%.The area under curve was 0.97.The average diagnostic time was less than 1 second.Although the model for data of 40×magnifications was very sensitive(98.72%),but its specificity was poor,suggesting that the model was more reliable at 10×magnification.Conclusions The performance of a deep-learning based model is equivalent to pathologists′diagnostic performance,but its efficiency is far beyond.The model can greatly improve consistency and efficiency,and reduce the missed diagnosis rate.In the future,larger studies should have more morphology diversity,improve model′s accuracy and eventually develop a model for direct clinical use.
作者
叶美华
陈万远
蔡博君
金朝汇
何向蕾
Ye Meihua;Chen Wanyuan;Cai Bojun;Jin Chaohui;He Xianglei(Department of Pathology,Hangzhou Medical College Zhejiang Provincial People′s Hospital,Hangzhou 310014,China;Zhejiang Tonghuashun Intelligent Technology Co.,Ltd,Hangzhou 311100,China)
出处
《中华病理学杂志》
CAS
CSCD
北大核心
2021年第4期358-362,共5页
Chinese Journal of Pathology
基金
浙江省自然科学基金(浙江省基础公益研究计划项目)(LGF21H160031)。
关键词
人工智能
甲状腺疾病
细胞学技术
Artificial intelligence
Thyroid diseases
Cytological techniques
作者简介
通信作者:何向蕾,Email:xianglh992004@163.com。
