Diagnostic Pathology | |
The value of deep neural networks in the pathological classification of thyroid tumors | |
Research | |
Chengwen Deng1  Dan Li2  Qingqing Huang3  Dongyan Han4  Ming Feng5  | |
[1] Department of Nuclear Medicine, Chaohu Hospital of Anhui Medical University, 238000, Heifei, China;Department of Nuclear Medicine, Sun Yat‐sen Memorial Hospital, Sun Yat‐sen University, 510289, Guangzhou, China;Shanghai Key Laboratory of Molecular Imaging, Shanghai University of Medicine and Health Sciences, 201318, Shanghai, China;Shanghai Tenth People’s Hospital Tongji University, 200072, Shanghai, China;Tongji University, 200082, Shanghai, China; | |
关键词: Deep neural network; Thyroid tumor; Pathology; Diagnostics; Artificial intelligence; | |
DOI : 10.1186/s13000-023-01380-2 | |
received in 2023-06-21, accepted in 2023-08-08, 发布年份 2023 | |
来源: Springer | |
【 摘 要 】
BackgroundTo explore the distinguishing diagnostic value and clinical application potential of deep neural networks (DNN) for pathological images of thyroid tumors.MethodsA total of 799 pathological thyroid images of 559 patients with thyroid tumors were retrospectively analyzed. The pathological types included papillary thyroid carcinoma (PTC), medullary thyroid carcinoma (MTC), follicular thyroid carcinoma (FTC), adenomatous goiter, adenoma, and normal thyroid gland. The dataset was divided into a training set and a test set. Resnet50, Resnext50, EfficientNet, and Densenet121 were trained using the training set data and tested with the test set data to determine the diagnostic efficiency of different pathology types and to further analyze the causes of misdiagnosis.ResultsThe recall, precision, negative predictive value (NPV), accuracy, specificity, and F1 scores of the four models ranged from 33.33% to 100.00%. The area under curve (AUC) ranged from 0.822 to 0.994, and the Kappa coefficient ranged from 0.7508 to 0.7713. However, the performance of diagnosing FTC, adenoma, and adenomatous goiter was slightly inferior to other types of pathological tissues.ConclusionThe DNN model achieved satisfactory results in the task of classifying thyroid tumors by learning thyroid pathology images. These results indicate the potential of the DNN model for the efficient diagnosis of thyroid tumor histopathology.
【 授权许可】
CC BY
© BioMed Central Ltd., part of Springer Nature 2023
【 预 览 】
Files | Size | Format | View |
---|---|---|---|
RO202309152966570ZK.pdf | 1894KB | download | |
42490_2023_74_Article_IEq62.gif | 1KB | Image | download |
13690_2023_1154_Article_IEq16.gif | 1KB | Image | download |
Fig. 1 | 279KB | Image | download |
MediaObjects/12888_2023_5086_MOESM1_ESM.docx | 19KB | Other | download |
40517_2023_261_Article_IEq34.gif | 1KB | Image | download |
40517_2023_261_Article_IEq36.gif | 1KB | Image | download |
Fig. 1 | 105KB | Image | download |
Fig. 2 | 63KB | Image | download |
40517_2023_261_Article_IEq45.gif | 1KB | Image | download |
Fig. 1 | 1005KB | Image | download |
MediaObjects/12974_2023_2861_MOESM2_ESM.tif | 30484KB | Other | download |
Fig. 4 | 226KB | Image | download |
MediaObjects/12888_2023_5086_MOESM3_ESM.docx | 575KB | Other | download |
MediaObjects/42004_2023_986_MOESM6_ESM.docx | 12133KB | Other | download |
12888_2023_5113_Article_IEq2.gif | 1KB | Image | download |
12888_2023_5113_Article_IEq3.gif | 1KB | Image | download |
12888_2023_5113_Article_IEq4.gif | 1KB | Image | download |
12888_2023_5113_Article_IEq5.gif | 1KB | Image | download |
【 图 表 】
12888_2023_5113_Article_IEq5.gif
12888_2023_5113_Article_IEq4.gif
12888_2023_5113_Article_IEq3.gif
12888_2023_5113_Article_IEq2.gif
Fig. 4
Fig. 1
40517_2023_261_Article_IEq45.gif
Fig. 2
Fig. 1
40517_2023_261_Article_IEq36.gif
40517_2023_261_Article_IEq34.gif
Fig. 1
13690_2023_1154_Article_IEq16.gif
42490_2023_74_Article_IEq62.gif
【 参考文献 】
- [1]
- [2]
- [3]
- [4]
- [5]
- [6]
- [7]
- [8]
- [9]
- [10]
- [11]
- [12]
- [13]
- [14]
- [15]
- [16]
- [17]
- [18]
- [19]
- [20]