Successful Identification of Nasopharyngeal Carcinoma in Nasopharyngeal Biopsies Using Deep Learning
Autor: | Tong-Hong Wang, Shang-Hung Chang, Yu-Jen Liu, Chi-Ju Yeh, Kuang-Hua Chen, Chao-Yuan Yeh, Wei-Hsiang Yu, Shir-Hwa Ueng, Chang-Fu Kuo, Tai-Di Chen, Wen-Yu Chuang, Yi-Yin Hsieh, Yi Hsia, Cheng-Kun Yang, Chuen Hsueh |
---|---|
Jazyk: | angličtina |
Rok vydání: | 2020 |
Předmět: |
0301 basic medicine
Cancer Research medicine.medical_specialty cancer identification convolutional neural network Biology lcsh:RC254-282 Article 03 medical and health sciences 0302 clinical medicine gradient-weighted class activation mapping otorhinolaryngologic diseases medicine Carcinoma Stage (cooking) Training set Receiver operating characteristic business.industry Deep learning nasopharyngeal carcinoma Digital pathology deep learning medicine.disease artificial intelligence lcsh:Neoplasms. Tumors. Oncology. Including cancer and carcinogens Identification (information) 030104 developmental biology Oncology Nasopharyngeal carcinoma 030220 oncology & carcinogenesis Radiology Artificial intelligence business digital pathology |
Zdroj: | Cancers, Vol 12, Iss 2, p 507 (2020) Cancers Volume 12 Issue 2 |
ISSN: | 2072-6694 |
Popis: | Pathologic diagnosis of nasopharyngeal carcinoma (NPC) can be challenging since most cases are nonkeratinizing carcinoma with little differentiation and many admixed lymphocytes. Our aim was to evaluate the possibility to identify NPC in nasopharyngeal biopsies using deep learning. A total of 726 nasopharyngeal biopsies were included. Among them, 100 cases were randomly selected as the testing set, 20 cases as the validation set, and all other 606 cases as the training set. All three datasets had equal numbers of NPC cases and benign cases. Manual annotation was performed. Cropped square image patches of 256 × 256 pixels were used for patch-level training, validation, and testing. The final patch-level algorithm effectively identified NPC patches, with an area under the receiver operator characteristic curve (AUC) of 0.9900. Using gradient-weighted class activation mapping, we demonstrated that the identification of NPC patches was based on morphologic features of tumor cells. At the second stage, whole-slide images were sequentially cropped into patches, inferred with the patch-level algorithm, and reconstructed into images with a smaller size for training, validation, and testing. Finally, the AUC was 0.9848 for slide-level identification of NPC. Our result shows for the first time that deep learning algorithms can identify NPC. |
Databáze: | OpenAIRE |
Externí odkaz: | |
Nepřihlášeným uživatelům se plný text nezobrazuje | K zobrazení výsledku je třeba se přihlásit. |