Weakly-supervised learning for lung carcinoma classification using deep learning
Abstract Lung cancer is one of the major causes of cancer-related deaths in many countries around the world, and its histopathological diagnosis is crucial for deciding on optimum treatment strategies. Recently, Artificial Intelligence (AI) deep learning models have been widely shown to be useful in...
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2020-06-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-020-66333-x |
| _version_ | 1818722484482998272 |
|---|---|
| author | Fahdi Kanavati Gouji Toyokawa Seiya Momosaki Michael Rambeau Yuka Kozuma Fumihiro Shoji Koji Yamazaki Sadanori Takeo Osamu Iizuka Masayuki Tsuneki |
| author_facet | Fahdi Kanavati Gouji Toyokawa Seiya Momosaki Michael Rambeau Yuka Kozuma Fumihiro Shoji Koji Yamazaki Sadanori Takeo Osamu Iizuka Masayuki Tsuneki |
| author_sort | Fahdi Kanavati |
| collection | DOAJ |
| description | Abstract Lung cancer is one of the major causes of cancer-related deaths in many countries around the world, and its histopathological diagnosis is crucial for deciding on optimum treatment strategies. Recently, Artificial Intelligence (AI) deep learning models have been widely shown to be useful in various medical fields, particularly image and pathological diagnoses; however, AI models for the pathological diagnosis of pulmonary lesions that have been validated on large-scale test sets are yet to be seen. We trained a Convolution Neural Network (CNN) based on the EfficientNet-B3 architecture, using transfer learning and weakly-supervised learning, to predict carcinoma in Whole Slide Images (WSIs) using a training dataset of 3,554 WSIs. We obtained highly promising results for differentiating between lung carcinoma and non-neoplastic with high Receiver Operator Curve (ROC) area under the curves (AUCs) on four independent test sets (ROC AUCs of 0.975, 0.974, 0.988, and 0.981, respectively). Development and validation of algorithms such as ours are important initial steps in the development of software suites that could be adopted in routine pathological practices and potentially help reduce the burden on pathologists. |
| first_indexed | 2024-12-17T20:55:22Z |
| format | Article |
| id | doaj.art-ca955f6401a44b0690a1db64ed281447 |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-12-17T20:55:22Z |
| publishDate | 2020-06-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-ca955f6401a44b0690a1db64ed2814472022-12-21T21:32:54ZengNature PortfolioScientific Reports2045-23222020-06-0110111110.1038/s41598-020-66333-xWeakly-supervised learning for lung carcinoma classification using deep learningFahdi Kanavati0Gouji Toyokawa1Seiya Momosaki2Michael Rambeau3Yuka Kozuma4Fumihiro Shoji5Koji Yamazaki6Sadanori Takeo7Osamu Iizuka8Masayuki Tsuneki9Medmain Research, Medmain Inc.Department of Thoracic Surgery, Clinical Research Institute, National Hospital Organization, Kyushu Medical CenterDepartment of Pathology, Clinical Research Institute, National Hospital Organization, Kyushu Medical CenterMedmain Inc.Department of Thoracic Surgery, Clinical Research Institute, National Hospital Organization, Kyushu Medical CenterDepartment of Thoracic Surgery, Clinical Research Institute, National Hospital Organization, Kyushu Medical CenterDepartment of Thoracic Surgery, Clinical Research Institute, National Hospital Organization, Kyushu Medical CenterDepartment of Thoracic Surgery, Clinical Research Institute, National Hospital Organization, Kyushu Medical CenterMedmain Inc.Medmain Research, Medmain Inc.Abstract Lung cancer is one of the major causes of cancer-related deaths in many countries around the world, and its histopathological diagnosis is crucial for deciding on optimum treatment strategies. Recently, Artificial Intelligence (AI) deep learning models have been widely shown to be useful in various medical fields, particularly image and pathological diagnoses; however, AI models for the pathological diagnosis of pulmonary lesions that have been validated on large-scale test sets are yet to be seen. We trained a Convolution Neural Network (CNN) based on the EfficientNet-B3 architecture, using transfer learning and weakly-supervised learning, to predict carcinoma in Whole Slide Images (WSIs) using a training dataset of 3,554 WSIs. We obtained highly promising results for differentiating between lung carcinoma and non-neoplastic with high Receiver Operator Curve (ROC) area under the curves (AUCs) on four independent test sets (ROC AUCs of 0.975, 0.974, 0.988, and 0.981, respectively). Development and validation of algorithms such as ours are important initial steps in the development of software suites that could be adopted in routine pathological practices and potentially help reduce the burden on pathologists.https://doi.org/10.1038/s41598-020-66333-x |
| spellingShingle | Fahdi Kanavati Gouji Toyokawa Seiya Momosaki Michael Rambeau Yuka Kozuma Fumihiro Shoji Koji Yamazaki Sadanori Takeo Osamu Iizuka Masayuki Tsuneki Weakly-supervised learning for lung carcinoma classification using deep learning Scientific Reports |
| title | Weakly-supervised learning for lung carcinoma classification using deep learning |
| title_full | Weakly-supervised learning for lung carcinoma classification using deep learning |
| title_fullStr | Weakly-supervised learning for lung carcinoma classification using deep learning |
| title_full_unstemmed | Weakly-supervised learning for lung carcinoma classification using deep learning |
| title_short | Weakly-supervised learning for lung carcinoma classification using deep learning |
| title_sort | weakly supervised learning for lung carcinoma classification using deep learning |
| url | https://doi.org/10.1038/s41598-020-66333-x |
| work_keys_str_mv | AT fahdikanavati weaklysupervisedlearningforlungcarcinomaclassificationusingdeeplearning AT goujitoyokawa weaklysupervisedlearningforlungcarcinomaclassificationusingdeeplearning AT seiyamomosaki weaklysupervisedlearningforlungcarcinomaclassificationusingdeeplearning AT michaelrambeau weaklysupervisedlearningforlungcarcinomaclassificationusingdeeplearning AT yukakozuma weaklysupervisedlearningforlungcarcinomaclassificationusingdeeplearning AT fumihiroshoji weaklysupervisedlearningforlungcarcinomaclassificationusingdeeplearning AT kojiyamazaki weaklysupervisedlearningforlungcarcinomaclassificationusingdeeplearning AT sadanoritakeo weaklysupervisedlearningforlungcarcinomaclassificationusingdeeplearning AT osamuiizuka weaklysupervisedlearningforlungcarcinomaclassificationusingdeeplearning AT masayukitsuneki weaklysupervisedlearningforlungcarcinomaclassificationusingdeeplearning |