false
Catalog
2023 North America Conference on Lung Cancer (NACL ...
PP01.19 (Poster) Comparison of AI models in Lung C ...
PP01.19 (Poster) Comparison of AI models in Lung Cancer Diagnosis and Staging
Back to course
Pdf Summary
This research study aims to compare the effectiveness of different artificial intelligence (AI) models in diagnosing and staging lung cancer using CT scans. Lung cancer is a significant health issue, with a high mortality rate if not diagnosed early. Non-invasive cancer staging innovations, such as low-dose CT scans, have improved the ability to screen for lung cancer in heavy smokers.<br /><br />The study used various AI models, including CNN, SVM, RNN, LSTM, and DNN, to analyze and classify CT images of lung cancer. The CNN, Random Tree, and SVM models performed exceptionally well, achieving around 95% accuracy on the testing dataset. However, the RNN, LSTM, DNN, L.V.Q., and Naive Bayes models had lower accuracies.<br /><br />The results showed that the CNN models, especially Alexnet, performed the best in diagnosing and staging lung cancer, with an accuracy of 98.06%. The study also found that these AI models could accurately determine the type of lung cancer, with Alexnet achieving an accuracy of 99.635%.<br /><br />To further optimize the AI models, techniques such as transfer learning, kernel optimization, and hyperparameter tuning were employed. The augmented data was split into subtypes of lung cancer and enhanced and normalized to improve accuracy and robustness.<br /><br />The study concluded that AI models, particularly the CNN models, can assist doctors in diagnosing and staging lung cancer from CT scans. The Alexnet architecture showed the most promise in accurately diagnosing and staging lung cancer. The research also suggested the potential use of AI in detecting other types of lung cancer and expanding applications to other cancers.<br /><br />Overall, this study contributes to the development of AI models for lung cancer diagnosis and staging and provides a baseline for future research in this field.
Asset Subtitle
Aidan Gao
Keywords
artificial intelligence
AI models
lung cancer diagnosis
lung cancer staging
CT scans
CNN
SVM
RNN
LSTM
DNN
×
Please select your language
1
English